CORO1A, A Novel Upregulated Protein in Renal Tubular Epithelial Cells of Diabetic Nephropathy, Participates in Regulating Vesicular Trafficking and the Ferroptosis Pathway

ABSTRACT

Aim

Diabetic nephropathy (DN) is characterised by progressive renal tubular epithelial cell injury, yet underlying molecular mechanisms remain incompletely understood.

Methods

Differentially expressed genes (DEGs) in DN were identified through integrated GEO dataset analysis (GSE30529 and GSE175759). Protein–protein interaction networks identified hub genes. GO and KEGG enrichment analyses determined molecular mechanisms. Nephroseq v5 explored correlations between hub genes and clinical features. CORO1A expression was validated by immunohistochemistry and western blot in DN patients, diabetic db/db mice and high glucose‐exposed HK‐2 cells. Vps34 inhibition and CORO1A knockdown explored regulatory mechanisms in vesicular trafficking and ferroptosis.

Results

Bioinformatics identified CORO1A as a phagosome‐related core gene significantly upregulated in DN renal tubulointerstitium and correlated with eGFR, proteinuria, serum creatinine, blood pressure and age. External validation and experiments confirmed CORO1A upregulation in DN patients, diabetic mice and high glucose‐exposed cells. Vps34 inhibition impaired lysosomal degradation flux and increased tubular injury markers, ferroptosis, ferritinophagy and CORO1A expression. CORO1A knockdown attenuated ferroptosis and lipid peroxidation, while rescue experiments with Ferrostatin‐1 and Erastin confirmed the functional link between CORO1A, vesicular trafficking and ferroptosis.

Conclusion

CORO1A is elevated in diabetic renal tubular epithelial cells and regulates vesicular trafficking and ferroptosis. This mechanistic axis suggests CORO1A as a potential therapeutic target for mitigating tubular injury and diabetic nephropathy progression.

Summary at a Glance

CORO1A is significantly upregulated in diabetic nephropathy and correlates with renal function decline. This study reveals that CORO1A regulates vesicular trafficking and ferroptosis in renal tubular epithelial cells, representing a novel therapeutic target for diabetic kidney disease.

CORO1A, a phagosome‐related actin‐binding protein, is significantly upregulated in renal tubular epithelial cells of diabetic nephropathy (DN) and correlates with eGFR decline, proteinuria, serum creatinine, blood pressure and age. Bioinformatics analysis combined with in vivo and in vitro validation revealed that Vps34 inhibition disrupts lysosomal degradation flux, enhances ferritinophagy and upregulates CORO1A expression. CORO1A knockdown attenuates ferroptosis and lipid peroxidation. These findings position CORO1A as a novel mechanistic link between Vps34‐mediated vesicular trafficking and ferroptosis, suggesting its potential as a therapeutic target for DN progression.

1. Introduction

Diabetic nephropathy (DN) has been recognised as the leading cause of end‐stage renal disease (ESRD), characterised by pathological changes including glomerular basement membrane thickening and mesangial dilation [1]. While glomerular lesions have historically been emphasised, accumulating evidence shows that tubular epithelial cell dysfunction directly accelerates interstitial fibrosis and renal function decline in DN patients [2].

Recent studies have revealed that Vps34 mediated vesicular transportations, that is, autophagy–lysosome and endosome–lysosome trafficking, were critical processes for maintaining the structural and functional homeostasis of renal tubular epithelial cells by organising the transport, sorting and degradation of proteins, receptors and turnover of damaged proteins and organelles [3]. Vps34, also known as, Class III phosphoinositide 3‐kinase, is a master regulator of PI3P production on autophagic membranes and endosomes; conditional deletion of Vps34 in proximal tubule cells disrupts apical membrane protein recycling, impairs endosome–lysosome fusion and accelerates tubular atrophy in mouse models of kidney injury [4]. Pharmacological inhibitors of Vps34 block autophagosome formation and endosomal maturation, underscoring Vps34's essential role in vesicular trafficking and lysosomal degradation in renal tubular epithelial cells [5, 6]. In diabetic nephropathy, advanced glycation end products and hyperglycemia impair autophagic flux, leading to lysosomal dysfunction and accumulation of autophagic vacuoles in proximal tubular epithelial cells [6, 7]. However, the role of VPS34 and its mediated membranovesicular transport system in renal tubular epithelial cell injury in diabetic nephropathy has rarely been reported.

Ferroptosis, an autophagy regulated cell death characterised by lipid peroxidation, has emerged as a critical mediator of tubular epithelial injury in diabetic nephropathy [8]. Disrupted ferritinophagy, the selective autophagic degradation of ferritin, leads to labile iron release, promotes reactive oxygen species generation and sensitises renal tubular epithelial cells to ferroptotic death, thereby promoting inflammation, fibrosis and functional decline in DN patients [9]. Experimental models demonstrated that modulation of key ferroptosis regulators (e.g., glutathione peroxidase 4 [GPX4], Xc‐cystine/glutamate antiporter [XCT] and phosphoseryl‐tRNA kinase [PSTK]) or restoration of ferritinophagy attenuates tubular cell death, inflammation and fibrosis [10, 11, 12]. Despite that, the molecular link between vesicular trafficking and ferroptosis in renal tubular epithelial cells remains incompletely defined.

Collectively, our bioinformatics analyses identified CORO1A as a core hub gene upregulated in diabetic nephropathy and associated with clinical indicators: serum creatinine, eGFR, proteinuria, blood pressure and age. Experimental exploration demonstrated that CORO1A participated in Vps34‐dependent vesicular trafficking and ferritinophagy. This integrated approach revealed CORO1A as a novel mechanistic link driving tubular injury in diabetic nephropathy and highlighted its potential as a therapeutic target.

2. Materials and Methods

2.1. Microarray Data Analysis

We downloaded the datasets including GSE30529 and GSE175759 from GEO (http://www.ncbi.nlm.nih.gov/geo), an international public functional genomics data repository of next‐generation sequence, chips and microarrays. The DEGs in DN and normal renal tissues were firstly selected by the GEO2R online tool (https://www.ncbi.nlm.nih.gov/geo/geo2r/) with ∣LogFC ∣ > 1 and adjusted p < 0.05. Subsequently, statistical analysis was carried out for each dataset; the intersecting DEGs among the two datasets were identified using the Venn diagram webtool (bioinformatics.psb.ugent.be/webtools/Venn/).

2.2. Protein–Protein Interaction (PPI) Network Construction, Topological Analysis and Hub Gene Identification

A PPI network of overlapping DEGs was established based on the STRING platform (https://string‐db.org/) to retrieve nearly all functional interactions among the expressed proteins. Protein interaction information derived from the STRING database was imported into the Cytoscape software where the interaction information was integrated and visualised. Then, the most significant modules in the PPI network were selected by molecular complex detection (MCODE); the parameters of clustering and scoring were MCODE score ≥ 2, degree cutoff = 2, node score cutoff = 0:2, max depth = 100 and k‐score = 2. At the same time, topological analysis was performed by using the CytoNCA plugin in Cytoscape software and four topological features (degree, betweenness centrality, average shortest path length and closeness centrality) were analysed. Subsequently, selecting nodes with degrees higher than the average number, taking the intersection of the topological analysis and clustering analysis result as the hub genes. If the LogFC < 0, the expression of the hub gene was deemed to be downregulated, while LogFC > 0, gene expression was deemed to be upregulated. Additionally, a network of the hub genes and their coexpression network was performed by using the Cytoscape software.

2.3. GO and KEGG Pathway Enrichment Analysis

GO is a comprehensive and widely used database for the classification of gene functions, consisting of biological process (BP), molecular function (MF) and cell component (CC). KEGG (http://www.kegg.jp/) is an encyclopaedia of genomes, which links genomic information with higher‐order functional information to capture significantly enriched biological pathways. GO annotation analysis and KEGG pathway enrichment analysis of DEGs in this study was performed using the Database for Annotation, Visualization and Integrated Discovery (DAVID) tools (https://david.ncifcrf.gov/). Adjusted p < 0.05 was deemed as a screening threshold.

2.4. The Association Between Hub Genes and Clinical Features of DN

Correlation analysis and subgroup analysis between hub genes and clinical features were carried out via the Nephroseq v5 online tool (http://v5.nephroseq.org) to confirm the potential functions of hub genes participated in DN.

2.5. Human Tissue Samples and Animals

Human tissue samples were collected from renal biopsy samples of DN patients or samples of kidneys donated by deceased donors in the Department of Pathology, School of Basic Medical Sciences, Shanghai Medical College, Fudan University. db/db mice, a model for type 2 diabetes, were purchased from Selleck Laboratory Animal Company, Shanghai, China. Only male mice were used in this study. All human and animal experiments were approved by the Animal Research Ethics Committee of the School of Basic Medical Sciences of Fudan University.

2.6. Cell Culture

HK‐2 cells (human renal proximal tubular epithelial cells) were obtained from American Type Culture Collection (Manassas, VA). For high glucose treatment, cells were starved in DMEM/‐F12 (Thermo Fisher Scientific, Waltham, MA, USA) without foetal bovine serum (FBS) (Thermo Fisher Scientific) overnight followed by incubation for 48 h with 30 mM glucose in DMEM/F12, equal concentration of mannitol as control. For Vps34 inhibition experiments, HK‐2 cells were first treated with high glucose (30 mM) for 48 h, co‐treated with Vps34‐IN1 (1 μM) or DMSO vehicle control. For rescue experiments, cells with Vps34 inhibition were co‐treated with Ferrostatin‐1 (Fer‐1, 1 μM) or DMSO vehicle control. For ferroptosis induction experiments, HK‐2 cells transfected with si‐CORO1A or si‐NT were co‐treated with Erastin (1 μM) or DMSO.

2.7. RNA Interference

Small interfering RNA (siRNA) targeting CORO1A for HK‐2 cells was purchased from Genepharma Biotechnology (Shanghai), including 3 independent siRNAs. Non‐targeting siRNA (siNT) was used as a control. For transient knockdown of protein expression, siRNA was transfected into HK‐2 cells by using TransIT‐X2 Transfection Reagent (Mirus Bio, Madison, WI, USA) according to the manufacturer's protocol.

2.8. Antibodies and Reagents

We used the following antibodies: antibodies for CORO1A (Santa Cruz Biotechnology, Dallas, TX, USA), Collagen1, XCT and GPX4 (Cell Signalling Technology, Beverly, MA, USA), FTH (Abcam, Waltham, MA, USA), KIM1, LAMP2 and NCOA4 (Novus Biologicals, Centennial, CO, USA), PSTK (Thermo Fisher Scientific, Waltham, MA, USA), P62 and GAPDH (Proteintech, Rosemont, IL, USA). Reagents we used in this study were listed: Vps34‐IN‐1 (MedChemExpress, Monmouth Junction, NJ), DMSO (Sigma‐Aldrich), glucose, mannitol, phosphate buffered saline (Sangon Biotech, Shanghai, China). Ferrostatin‐1 (Fer‐1), Erastin and BODIPY 581/591 C11 were purchased from MedChem Express (Monmouth Junction, NJ).

2.9. BODIPY 581/591 C11 Lipid Peroxidation Assay

Lipid peroxidation was assessed using BODIPY 581/591 C11. Cells were incubated with BODIPY 581/591 C11 for 30 min at 37°C, washed three times with PBS and imaged using a fluorescence microscope. The ratio of oxidised (green, 510 nm) to reduced (red, 590 nm) fluorescence was quantified using ImageJ software.

2.10. Measurement of Kidney Function and Proteinuria

Serum Creatinine chemical kit (sarcosine oxidase) was purchased from Enzyme‐linked Biotechnology (Shanghai, China). Urine microalbumin kit (Immunoturbidimetry) and Urine protein test kit (CBB) were from Jiancheng Bioengineering Institute of Nanjing (Nanjing, China). The ratio (UACR and UPCR) was calculated against creatinine concentration. Experiments were performed according to the manufacturer's protocol and repeated three times for each indicator.

2.11. Immunohistochemistry

Paraffin‐embedded sections were deparaffinated and hydrated followed by antigen retrieval in citrate buffer. Then, 3% H₂O₂ was used to remove endogenous peroxidase and 5% normal sheep serum was used for blocking. The sections were incubated with primary antibody at 4°C overnight (CORO1A, 1:200) and secondary antibody for 1 h at room temperature. Last, the expression was detected by DAB (DAKO, Glostrup, Denmark) and haematoxylin (Longisland, Shanghai, China) was used as the nuclear counterstain. Tubular immunohistochemistry scoring was performed by two renal pathologists. Positive area percentage was scored as follows: 0, < 5%; 1, 5%–25%; 2, 25%–50%; 3, 50%–75%; and 4, > 75%. The mean score of the tubular positive area percentage of at least 10 tubular fields in each human slide and 20 tubular fields in each mouse slide was calculated.

2.12. Western Blot

Lysates (20–40 μg) of renal cortex or cultured HK‐2 cells harvested using cell lysis buffer and measured by BCA Protein Assay Reagent kit (Piece, IL, USA) were used for western blot as described [13]. The band intensities were quantified using ImageJ software and normalised to the expression of GAPDH (loading control) and expressed as fold‐change relative to control group.

2.13. Statistical Analysis

Comparisons between two subgroups were carried out via an unpaired Student's t‐test. Multiple group comparisons were performed using two‐way ANOVA. Pearson's correlation analysis between hub genes and clinical features in patients with DN was conducted via applying Nephroseq v5. The statistical analyses were performed by using GraphPad Prism (GraphPad Software, La Jolla, California). All tests were two‐tailed, with a p value < 0.05 considered statistically significant.

3. Results

3.1. Identification of 102 DEGs in DN

GSE30529 and GSE175759, which are two datasets downloaded from the GEO database, contain 44 normally functioning human tubular samples and 13 diabetic human tubular samples. These samples were derived from diagnostic renal biopsies, healthy human living donors and tumour nephrectomy specimens; the latter two serve as control samples. The expression data obtained from the two datasets were analysed by GEO2R. The cutoff criteria were |LogFC| > 1 and adjusted p < 0.05. We found 557 and 692 DEGs in GSE30529 and GSE175759 datasets, respectively. Then, Venn diagram software was used to identify the common DEGs in these two datasets. A total of 102 DEGs were detected, 97 of which were upregulated and 5 of which were downregulated (Table S1 and Figure 1a).

FIGURE 1.

Identification of 36 hub DEGs by PPI network analysis. (a) Overlapping DEGs of two datasets. (b) The PPI network of overlapping DEGs of two datasets. Each circular node represents a protein target, and the 3D structure in the circular nodes shows the protein spatial structure. The lines among different nodes represent the association among potential protein targets, while the width of lines was according to the action intensity. (c) PPI network visualisation and analysis. (d) 36 hub genes were visualised by the heatmap.

3.2. 36 Hub Genes Were Selected by PPI Network Analysis

The construction of the PPI network was based on Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) online analysis. A total of 102 common DEG data points were imported into the PPI network for analysis, the interaction network was based on the selected targets with an interaction score of 0.400 at medium confidence and finally, 623 edges and 85 key nodes were embodied (Figure 1b). PPI network of common DEGs was visualised by Cytoscape (Figure 1c) and the four most significant modules were recognised by the MCODE plugin of Cytoscape. Among these four modules, a total of 47 DEGs were clustered (Table S2). Furthermore, the average node degree of common DEGs is 14.659 after the topological analysis by the CytoNCA plugin; 39 genes higher than the average degree were obtained. We took the intersection of the 39 genes with higher average degree and the 47 clustered genes as the hub genes.

All the 36 hub genes are upregulated genes, including: TYROBP, ITGB2, ITGAM, CTSS, CCL5, CD48, C1QA, HLA‐DRA, IRF8, CD52, SELL, CD53, C1QB, CCR2, CD2, LAPTM5, LY86, FCGR2B, GZMA, KLRB1, CD163, HLA‐DPA1, PLEK, CX3CR1, LCK, CD1C, VSIG4, CORO1A, HLA‐DMA, CASP1, CXCL1, HCLS1, MS4A6A, GZMK, NCF2, MS4A4A, which are displayed in Table S3. In addition, the heat map of the expression changes of hub genes in DN samples compared to control samples in the two GEO datasets is exhibited in Figure 1d.

3.3. GO and KEGG Pathway Enrichment Analysis of Hub Genes in DN

GO function and KEGG pathway enrichment analysis for hub DEGs were performed using the DAVID. The enriched GO terms were divided into CC, BP and MF ontologies. Based on the threshold of adjusted p < 0.05, the results of GO analysis indicated that DEGs were mainly enriched in CCs, where the top 10 remarkably enriched CC terms are plasma membrane, membrane, extracellular region, external side of plasma membrane, cell surface, immunological synapse, MHC class II protein complex, plasma membrane raft, tertiary granule membrane and endocytic vesicle membrane. The results of the GO analysis are illustrated in Figure 2a. Furthermore, the top 5 MF terms screened, consisting of protein binding, amyloid‐beta binding, complement component C3b binding, MHC class II receptor activity and MHC class II protein complex binding. Moreover, the top BP terms were also shown, containing immune response, innate immune response, inflammatory response, adaptive immune response, antigen processing and presentation of exogenous peptide antigen via MHC class II, positive regulation of T cell activation, phagocytosis, antigen processing and presentation, cellular defence response and chemokine‐mediated signalling pathway.

FIGURE 2.

GO and KEGG pathway enrichment analysis and validation of core genes in external datasets. (a) Enriched GO terms for BP, MF, and CC of hub genes. The colour of the node is displayed in a gradient from red to blue according to the ascending order of the adjusted p value, while the size of the node is shown according to the ascending order of the number of gene counts. (b) KEGG pathway analysis of hub genes. The top 10 pathways which are screened out according to the criteria of adjusted p < 0.05. The longitudinal axis displays the name of different pathways, and the transverse axis shows the count of enriched genes. In addition, the length of the bar is shown according to the ascending order of the number of gene counts, while the colour of the bar is displayed in a gradient from blue to red according to the descending order of the p value. (c) Validation of core genes in external datasets. The expression changes of core genes in tubulointerstitial samples of diabetic nephropathy (DN) compared to control samples (CT) from GSE104954, GSE47184, GSE99340, and GSE142025. Data were shown as mean ± standard error of mean. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, ns no significance.

In the KEGG pathway enrichment analysis, a total of 32 significantly enriched pathways were selected out based on the threshold of adjusted p < 0.05 (Figure 2b). The top 10 KEGG pathways indicated that these hub genes were mainly associated with pathways including phagosome, staphylococcus aureus infection, tuberculosis, cell adhesion molecules, leishmaniasis, pertussis, rheumatoid arthritis, haematopoietic cell lineage, complement and coagulation cascades and systemic lupus erythematosus. Among them, phagosome was the most enriched pathway. These results indicated that the hub DEGs were mainly associated with the regulation of vesicular trafficking, of which phagosome‐lysosome transportation was most clustered. Nine phagosome‐related core genes were listed in Table S4 for further exploration.

3.4. External Validation of Screened Nine Core Genes

Four datasets, GSE104954, GSE47184, GSE99340 and GSE142025, were used to conduct external validation of the nine phagosome related core genes. As shown in Figure 2c, all the nine core genes, that is, HLA‐DMA, ITGAM, NCF2, ITGB2, HLA‐DRA, FCGR2B, CORO1A, CTSS and HLA‐DPA1, had significantly upregulated expression in DN samples compared to controls in GSE104954 datasets. Besides, the nine core genes also upregulated in advanced DN samples compared to normal controls in GSE142025 datasets, which contained kidney biopsy samples from patients with early DN, advanced DN and normal kidney tissue from nephrectomy samples. However, there is no significant in the expression of nine core genes difference between early DN and control group. Furthermore, three genes (ITGAM, NCF2 and ITGB2) and four genes (ITGAM, NCF2, ITGB2 and HLA‐DPA1) showed upregulation in DN tissues compared to controls of GSE47184 and GSE99340 datasets, respectively.

3.5. Diagnostic Effectiveness of Nine Core Genes

With the use of Nephroseq v5, the expression of six of the nine phagosome‐related core genes (ITGAM, NCF2, ITGB2, FCGR2B, CORO1A and CTSS) showed over‐expression in tubulointerstitial samples from DN patients compared to healthy living donors (Figure 3a). In addition, the correlation between the core genes and the clinical features of DN patients was analysed. Four genes were associated with proteinuria. ITGB2, FCGR2B and HLA‐DPA1 were over‐expressed in nephrotic cases compared to sub‐nephrotic cases. The expression of ITGB2, CORO1A and HLA‐DPA1 was positively correlated with proteinuria level (Figure 3b). Moreover, four genes (ITGAM, ITGB2, FCGR2B and CORO1A) were positively correlated with serum creatinine of the DN patients (Figure 3c). The expression of NCF2, CORO1A and HLA‐DPA1 was negatively correlated with eGFR of DN samples (Figure 3d). Besides, ITGAM and CORO1A were positively correlated with blood pressure (Figure 3e). Six genes, ITGAM, NCF2, ITGB2, FCGR2B, CORO1A and HLA‐DPA1, were negatively correlated with age of the DN patients (Figure 3f). Among these phagosome‐related core genes, only one gene, CORO1A, is correlated with all five clinical indicators: serum creatinine, eGFR, proteinuria, blood pressure and age (Figure 3g). CORO1A is highly expressed in the renal tubulointerstitium of patients with diabetic nephropathy and shows a positive correlation with serum creatinine, proteinuria and blood pressure levels, while it is negatively correlated with eGFR and age. Therefore, we focused our attention on CORO1A.

FIGURE 3.

Correlation between the core genes and the clinical features of DN patients. (a) The expression of six of the nine phagosome‐related core genes (ITGAM, NCF2, ITGB2, FCGR2B, CORO1A, and CTSS) in tubulointerstitial samples from DN patients compared to healthy living donors. The correlation between the core genes and the clinical features, including proteinuria (b), serum creatinine (c), eGFR (d), blood pressure (e), and age (f). (g) Number of core genes correlated with each clinical feature of DN patients.

3.6. Experimental Validation Confirmed the Upregulation of CORO1A in the Renal Tubular Interstitium in Diabetic Nephropathy

Next, we verified the expression changes of the CORO1A molecule and the signalling pathways it regulates in the tubular interstitium of diabetic nephropathy through in vivo and in vitro experiments. Immunohistochemical staining of kidney tissues revealed that CORO1A is primarily expressed in renal tubular epithelial cells. Compared to healthy donors, the expression of CORO1A in renal tubular epithelial cells was significantly increased in patients with diabetic nephropathy (Figure 4a). We next investigated the expression of CORO1A in db/db mice, a type 2 diabetic model that showed obesity, hyperglycemia and proteinuria at age 18 weeks compared to the wild type control mice (Figure 4b). Immunohistochemistry staining and western blot showed that the expression of CORO1A in renal tubular epithelial cells and cortex tissue of db/db mice is significantly higher compared to the control group (Figure 4c,d). Furthermore, in vitro experiments demonstrated that treating cultured human proximal tubular epithelial cells, HK‐2 cells, with high glucose significantly increased the expression of CORO1A in these cells (Figure 4e).

FIGURE 4.

CORO1A was upregulated in the renal tubular interstitium in diabetic nephropathy both in vivo and in vitro. (a) Immunohistochemistry staining of CORO1A in diabetic nephropathy (DN) and healthy donor (HD). (b) Body weight, blood glucose, urine microalbumin creatinine rate, and urine protein creatinine rate of db/db mice and wild type (WT) control mice. (c) Immunohistochemistry staining of CORO1A in db/db mice and WT mice. (d, e) Western blot examined the protein level of CORO1A in renal cortex of db/db and WT mice, and in HK‐2 cells under normal glucose (NG) and high glucose (HG) conditions. Data were shown as mean ± standard error of mean. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

3.7. CORO1A as a Phagosome‐Related Core Gene Linked Vesicular Trafficking to Ferroptosis in Diabetic Tubular Cells

Vps34, also known as phosphatidylinositol 3 kinase class III (PI3KC3), was critical for lysosomal integrity and vesicular trafficking reported by our prior work [14]. Given that phagosomes represent a part of the lysosome‐directed vesicular trafficking system, Vps34 inhibition (Vps34‐IN1) was used as a tool to investigate whether CORO1A interfaces with the vesicular trafficking system under diabetic conditions. When HK‐2 cells were exposed to high glucose with Vps34‐IN1, autophagy and lysosome degradation flux was disrupted, as evidenced by robust accumulation of p62 and LAMP2 and KIM1 and Collagen‐I expression was elevated, reflecting tubular epithelial cell injury and a profibrotic response. Concomitantly, CORO1A showed significant increase in cells with Vps34‐IN1 compared to DMSO control. Moreover, inhibition of Vps34 led to the enhancement of ferroptosis and ferritinophagy signalling, which was manifested as the downregulation of Ferritin heavy chain (FTH) and upregulation of nuclear receptor co‐activator 4 (NCOA4). Moreover, three key ferroptosis regulators (PSTK, GPX4 and XCT) showed significant downregulation in Vps34‐IN1 treated cells compared to control (Figure 5a). To determine whether CORO1A directly influences ferroptosis‐related pathways, we knocked down CORO1A in high glucose‐treated HK‐2 cells using siRNA (si‐CORO1A) (Figure 5b) and upregulation of PSTK, XCT and GPX4 was observed in si‐CORO1A cells compared to non‐targeting control under high glucose condition.

FIGURE 5.

CORO1A as a phagosome‐related core gene, regulated by Vps34, linked vesicular trafficking to ferroptosis in diabetic tubular cells in vitro. (a) Western blot examined the protein level of p62, LAMP2, KIM1, Collagen‐I, CORO1A, FTH, NCOA4, PSTK, GPX4, and XCT in HK‐2 cells with or without Vps34‐IN1 under high glucose (HG) condition. (b) qPCR and western blot examined the protein level of CORO1A, PSTK, GPX4, and XCT in HK‐2 cells with transfection of siRNA against CORO1A (siCORO1A) or the non‐target shRNA control (siNT) under high glucose (HG) condition. Data were shown as mean ± standard error of mean. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, ns no significance.

3.8. CORO1A Links Vps34 Inhibition‐Induced Ferroptosis to Vesicular Trafficking in Renal Tubular Cells

To further confirm the involvement of ferroptosis in Vps34 inhibition‐mediated CORO1A upregulation, rescue experiments were performed using Ferrostatin‐1 (Fer‐1), a specific ferroptosis inhibitor. Western blot analysis demonstrated that CORO1A expression was significantly elevated in HK‐2 cells treated with Vps34‐IN1 compared to control and this upregulation was attenuated by co‐treatment with Fer‐1 (Figure 6a). BODIPY 581/591 C11 staining revealed that Vps34‐IN1 treatment markedly increased lipid peroxidation, as indicated by enhanced green fluorescence intensity. Importantly, co‐treatment with Fer‐1 significantly reduced lipid peroxidation levels (Figure 6b). These results confirmed that Vps34 inhibition induced ferroptosis and associated lipid peroxidation in HK‐2 cells and that the upregulation of CORO1A was closely linked to the ferroptotic process.

FIGURE 6.

Ferrostatin‐1 rescue experiments confirmed ferroptosis involvement in Vps34 inhibition‐induced CORO1A upregulation. (a) Western blot examined the protein level of CORO1A in HK‐2 cells treated with vehicle control, Vps34‐IN1 or Vps34‐IN1 combined with Ferrostatin‐1 (Fer‐1). (b) Representative images of BODIPY 581/591 C11 staining showing lipid peroxidation levels in HK‐2 cells under the same treatment conditions. Bar: 10 μm. Data were shown as mean ± standard error of mean. *p < 0.05, ***p < 0.001, ****p < 0.0001.

To further elucidate the functional role of CORO1A in ferroptosis and vesicular trafficking, HK‐2 cells were treated with si‐CORO1A alone or in combination with Erastin, a ferroptosis inducer. Western blot analysis showed that si‐CORO1A significantly reduced the expression of CORO1A, LAMP2 and NCOA4, while PSTK was upregulated compared to control, suggesting that CORO1A knockdown attenuated ferritinophagy signalling and enhanced anti‐ferroptotic defence. Notably, p62 levels were elevated in the si‐CORO1A group and further increased in the si‐CORO1A + Erastin group, indicating impaired autophagic flux under combined conditions. Treatment with Erastin reversed the effects of CORO1A knockdown, restoring the expression of CORO1A, LAMP2 and NCOA4, while reducing PSTK levels (Figure 7a). BODIPY 581/591 C11 staining showed that si‐CORO1A reduced lipid peroxidation compared to control, whereas co‐treatment with Erastin significantly enhanced lipid peroxidation (Figure 7b). These findings demonstrated that CORO1A played a pro‐ferroptotic role by promoting ferritinophagy‐mediated iron release and lipid peroxidation in renal tubular epithelial cells under diabetic conditions.

FIGURE 7.

CORO1A knockdown attenuated ferroptosis and modulated vesicular trafficking markers in HK‐2 cells. (a) Western blot examined the protein levels of CORO1A, LAMP2, NCOA4, PSTK, and P62 in HK‐2 cells treated with siNT control, si‐CORO1A or si‐CORO1A combined with Erastin. (b) Representative images of BODIPY 581/591 C11 staining showing lipid peroxidation levels in HK‐2 cells under the same treatment conditions. Bar: 20 μm. Data were shown as mean ± standard error of mean. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

4. Discussion

By integrating GEO datasets and performing both external in silico validation and in vivo/in vitro experiments, our study identified CORO1A as a phagosome‐related core gene that was significantly upregulated in the renal tubulointerstitium of patients with diabetic nephropathy (DN) and correlated with all key clinical parameters (eGFR, proteinuria, serum creatinine, blood pressure and age). Functional experiments demonstrated that knockdown of CORO1A attenuated ferroptosis and lipid peroxidation, while ferroptosis inducers reversed the protective effects of CORO1A silencing. Rescue experiments with Ferrostatin‐1 further confirmed the involvement of ferroptosis in Vps34 inhibition‐mediated CORO1A upregulation. These findings positioned CORO1A at the nexus of vesicular trafficking, autophagy and ferroptosis in DN (Figure 8).

FIGURE 8.

A flow chart based on an integration strategy of bioinformatics analysis and experimental exploration. Bioinformatics analysis identified CORO1A as a phagosome‐related core gene that was significantly upregulated in the renal tubulointerstitium of patients with diabetic nephropathy (DN) and correlated with all key clinical parameters (eGFR, proteinuria, serum creatinine, blood pressure, and age). In vivo and in vitro experiments demonstrated that CORO1A, as a F‐actin‐binding protein that modulated actin filament organisation, participated in regulating vesicular trafficking and the ferroptosis pathway. These findings positioned CORO1A at the nexus of vesicular trafficking, cytoskeleton, and ferroptosis in DN.

The Coronin family comprises seven evolutionarily conserved proteins that regulate a wide array of biological activities, including cell motility, phagocytosis, vesicular trafficking and cytokinesis [15]. CORO1A, also known as Coronin‐1a, a 57 kDa protein, encodes a member of the coronin family, characterised as a F‐actin‐binding protein that modulates actin filament organisation (both to stabilise and to destabilise actin filaments) by bridging adjacent protomers and signalling (e.g., Arp2/3 and Rho/Rac family), thereby regulating cytoskeletal dynamics [16, 17]. A growing body of work suggested that CORO1A was a key regulator of and lysosomal trafficking processes. Fei et al.'s study reported that CORO1A promoted the recruitment and activation of Rab7 on Multivesicular bodies (MVBs) by directly binding Rab7, thereby determining their fate toward lysosomal degradation or secretion as extracellular vesicles [18]. Moreover, Suo et al.'s study established Coronin‐1 as an essential component of a novel feedback loop mediating NGF‐TrkA endosome stability, recycling and signalling as a critical mechanism governing developmental competition for survival of the mouse sympathetic neurons [19].

In contrast to its well‐characterised roles in immune and nervous system development, CORO1A's function in the kidney, particularly under diabetic conditions, remains unexplored. Given that CORO1A has been shown to regulate Rab7 dependent lysosome trafficking, which was controlled by Vps34 in our previous work [14], we treated high glucose exposed tubular cells with a selective Vps34 inhibitor and then assessed CORO1A's involvement in directing cargo to lysosomal degradation. Interestingly, beyond general autophagy, our study showed Vps34 also governed ferritinophagy, which was selective turnover of ferritin by autophagic machinery [20, 21]. Macroautophagy/autophagy is a complex degradation process with a dual role in cell death that is influenced by the cell types that are involved and the stressors they are exposed to. The concepts of autophagy‐dependent ferroptosis and autophagy‐independent ferroptosis underscore the intricate relationship between autophagy and ferroptosis [22]. Whether autophagy positively or negatively regulates ferroptosis remains inconclusive and is highly dependent on the specific context. Some views suggested that autophagy mediated an amplification loop during ferroptosis, initial lipid peroxidation triggered autophagic degradation of ferritin, releasing iron that further derives ROS generation and cell death [23, 24]. On the other hand, other studies have showed that autophagy induction cleared lipid peroxides and damaged mitochondria protecting cells from ferroptotic death under oxidative stress [25]. In our study, inhibition of Vps34 disrupted lysosome degradation flux, but enhanced the ferroptosis and ferritinophagy signalling and concomitant CORO1A upregulation, implying the negative correlation between macroautophagy and ferritinophagy in tubular epithelial cell under diabetic condition and suggesting CORO1A's involvement in directing cargo to lysosomal degradation, including autophagy and ferritinophagy.

Importantly, our rescue experiments using Ferrostatin‐1 provided direct evidence that Vps34 inhibition‐induced CORO1A upregulation was mechanistically linked to ferroptosis. Ferrostatin‐1, by scavenging lipid peroxyl radicals, not only reduced lipid peroxidation but also attenuated CORO1A expression, suggesting a feedback relationship between ferroptotic signalling and CORO1A levels. Furthermore, CORO1A knockdown combined with Erastin treatment demonstrated that CORO1A functionally modulated ferroptosis sensitivity: silencing CORO1A reduced lysosomal and ferritinophagy markers (LAMP2, NCOA4), while Erastin reversed these effects. The elevated p62 accumulation observed in both si‐CORO1A and si‐CORO1A + Erastin groups indicated that CORO1A knockdown impaired autophagic flux, which was further exacerbated by ferroptotic stress. These findings collectively support the concept that CORO1A serves as a critical mediator linking vesicular trafficking to ferroptosis in diabetic tubular epithelial cells.

PSTK, as a key kinase in selenocysteine biosynthesis, is required for the synthesis of the catalytic selenocysteine residue in GPX4, a central regulator of ferroptosis [26, 27]. Multiple studies have established the crucial role of the ferroptosis–ferritinophagy system in mediating renal tubular epithelial cell injury. Dysregulated iron metabolism and lipid peroxidation, which are hallmarks of ferroptosis, contributed directly to tubular damage in various kidney diseases, including ischemic and toxic acute kidney injury, as well as diabetic nephropathy [28, 29]. Moreover, it has been reported that PSTK could prevent cell apoptosis of renal tubular epithelial cells via the BAX/BCL2/Caspase 3 pathway in Cisplatin induced AKI [12]. Our study implied that CORO1A might act as an upstream regulator of both PSTK signalling and ferroptotic cell injury of the tubular epithelial cells. Through its role in actin cytoskeleton organisation and vesicular trafficking, CORO1A may influence the efficiency of ferritin turnover, thereby modulating PSTK and GPX4 and suppressing ferroptosis under diabetic conditions. However, the mechanistic links between CORO1A and the PSTK‐ferroptosis axis require further research with conditional knockout models or pharmacological modulation (e.g., ZCH‐2B8a, a monoclonal anti‐CORO1A antibody, presented potential to be a therapeutic agent for B‐lineage malignancies [30]).

Several limitations of this study should be acknowledged. First, the sample sizes of the GEO datasets used in the bioinformatics analysis were relatively small, which may limit the generalizability of the findings. Second, since CORO1A is co‐clustered with immune cell‐related genes such as TYROBP and ITGB2, the observed upregulation in bulk tubulointerstitial samples may partially reflect increased immune cell infiltration rather than solely tubular cell‐specific changes. Although our immunohistochemistry results demonstrated that CORO1A was primarily expressed in renal tubular epithelial cells, single‐cell RNA sequencing would be needed to definitively resolve cell‐type‐specific contributions. Third, the current study lacks in vivo functional validation of CORO1A, such as conditional knockout or overexpression models in diabetic mice, which would strengthen the causal relationship between CORO1A and tubular injury. Future studies using tubule‐specific CORO1A knockout mice are warranted to validate these findings in vivo.

5. Conclusion

The identification and validation of CORO1A as a phagosome‐related core gene linking vesicular trafficking to ferroptosis advance our understanding of DN pathogenesis. By integrating bioinformatics with experimental validation, this work for the first time highlighted CORO1A as a novel mechanistic insight and potential therapeutic target for diabetic nephropathy.

Author Contributions

S.M. conducted experiments. X.X. and H.C. performed the data collection and analysis. Z.Z. assisted with project preparation. H.W. supervised the data analysis. C.Q. performed manuscript writing and funding acquisition.

Funding

This work was supported by the National Natural Science Foundation of China (82100757) and ZHENGYl Scholar Scientific Research Project of Fudan University School of Basic Medical Sciences.

Ethics Statement

The School of Basic Medical Sciences, Shanghai Medical College, Fudan University ethical committee approved the study protocol (approval No. #2017‐C008). And this study was performed following the Declaration of Helsinki. All participants signed informed consent prior to enrollment.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Table S1: 102 differentially expressed genes (DEGs) were identified from two profile datasets.

Table S2: The most significant clusters of common DEGs.

Table S3: Hub genes of diabetic nephropathy.

Table S4: Top 10 KEGG pathways.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.