Introduction

Diabetes mellitus (DM) is a metabolic disorder affecting the heart and kidneys, resulting from a prolonged positive energy balance and characterized by hyperglycemia; it has emerged a global health concern. In 2021, it was reported that 529 million people around the world had diabetes. By 2050, that number is expected rise to 1.31 billion people1. Type 2 diabetes mellitus (T2DM) can led to various vascular complications, with diabetic kidney disease (DKD) being one of the most common and severe; DKD is the primary cause of kidney damage and the necessity for end-stage renal replacement therapy globally2. The mechanisms underlying the etiopathogenesis of DKD are exceedingly complex and remain incompletely understood.. Research has identified several signaling pathways and mediators involved in the advancement of DKD, notably advanced glycation end products (AGEs) and reactive oxygen species (ROS)3,4.

Glycation, often known as the Maillard reaction, is a non-enzymatic processwherein free sugars bind to proteins, predominantly at lysine, arginine, and free amino groups. This transpires autonomously in the blood, where sugars or their breakdown products irreversibly attach to proteins, resulting in the formation of AGEs. Glycosylation may result in protein degradation, diminished stability, and impaired integration5.

Recent investigations have demonstrated that several proteins, including serum hemoglobin, serum albumin, and erythrocyte apoproteins, are modified by AGEs in patients with T2DM6. Advanced glycation end modified-transferrin (AGE-Tf) has been shown to have a strong correlation with the onset and progression of diabetes mellitus and its complications7,8,9,10. It has been shown that Lys206 and Lys534 sites of apo-transferrin (free iron)8and Lys103, Lys312, and Lys382 sites of holo-transferrin (bound iron)9were the most sensitive to glycation reactions. These modifications may affect Tf function, and molecular dynamics simulations also indicate that glycation of lysine residue may result in additonal loss of iron binding capacity due to stereochemical effects11.

The degree of serum transferrin (Tf) glycation in patients with T2DM was elevated compared to people without T2DM12. Meanwhile, this glycated modification of Tf may compromise its structural integrity related to iron binding, resulting in the formation of non-transferrin-bound-iron (NTBI)13and the initiation of oxidative stress reactions14, which have been observed in patients with T2DM10. Furthermore, glycated Tf is found not only in serum but also in the kidney tissues of rats with diabetes15.

Ferroptosis is an iron-dependent form of cell death distinct from apoptosis, characterized by iron overload, excessive generation of ROS, and lipid peroxidation16. The primary mechanism is the dysfunction of the glutathione (GSH)-dependent antioxidant defense system. Inhibition of the cystine/glutamate antiporter (system xc-, consisting of solute carrier family 7 member 11 (SLC7A11) and solute carrier family 3 member 2 (SLC3A2)) diminishes cellular cystine absorption, resulting in glutathione (GSH) depletion17. GSH depletion impairs the action of glutathione peroxidase 4 (GPX4), which utilizes GSH as a substrate to convert lipid hydroperoxides into benign lipid alcohols18. In the absence of this detoxification pathway, reactive iron facilitates the peroxidation of polyunsaturated fatty acids (PUFAs) through Fenton reactions. Enzymes such as acyl-coenzyme A (CoA) synthetase long-chain family member 4 (ACSL4) esterify PUFAs and incorporate them into membranes, thereby increasing their susceptibility to peroxidation19. Nuclear factor erythroid 2-related factor 2 (NRF2), as a principal regulator of the cellular antioxidant response, significantly contributes to the inhibition of ferroptosis by upregulating genes associated with iron metabolism (e.g., ferritin heavy chain 1 (FTH1)), GSH synthesis (e.g., glutamate-cysteine ligase modifier subunit (GCLM), glutamate-cysteine ligase catalytic subunit (GCLC)), and reactive oxygen species detoxification20. AGEs frequently facilitate the production of ROS21 and, consequently, represent an important molecular mechanism implicated in the development of oxidative stress and ferroptosis reactions16.This inflicts additional harm on proteins and other essential macromolecules, serving as a pathogenic mechanisms of DKD21. The impact of oxidative stress reactions induced by AGEs on DKD has been well investigated22. Kim et al. established a correlation between ferroptosis and DKD, noting that the expression levels of ferroptosis-related genes, such as SLC7A11 and GPX4, which are essential for inhibiting lipid peroxidation, were diminished levels in kidneys of patients with DKD compared to individuals without DKD, indicating a reduction in protection against ferroptotic cell death23.

Both glomerular and tubular epithelial cells have been demonstrated to be essential in the early and progressive stages of DKD. A prior work indicated that Tf glycation can induce oxidative stress and apoptosis in human tubular epithelial cells6. The relationship between the glomerulus, which plays a more critical role in DKD, and Tf glycation is not well understood. A recent study showed that high fructose can cause ferroptosis in podocytes24, while ginkgolide B may mitigate oxidative stress and ferroptosis in podocytes of mice with DKD25.

Although the association between AGE-Tf and DKD has been studied, the specific molecular mechanism remains unclear. Considering the established connections among AGEs, oxidative stress, and iron metabolism, we postulated that AGE-Tf could induce ferroptosis in podocytes by perturbing the GPX4/SLC7A11/ACSL4/NRF2 axis through the AGE-RAGE signaling pathway or by causing an imbalance in iron homeostasis due to the functional impairment of AGE-Tf. To further explore the mechanisms of DKD onset and progression, the present study therefore explored the effects of AGE-Tf produced by incubating Tf with different concentrations of glucose on human podocytes in vitro.

Materials and methods

Incubation of glycated transferrin

Four solutions were prepared by dissolving 0 mg, 10 mg, 20 mg, and 60 mg of D-glucose, respectively, in 10 mL of sodium phosphate buffer (0.1 M, pH 7.4), followed by the addition of 50 mg of apo-transferrinto each solution (all details of reagents and instruments were shown in Table S1). Each solution included a Tf concentration of 5 g/L and a glucose concentrations of 0 mM, 5.6 mM, 11.1 mM, and 33.3 mM, respectively. The solution was further filtered and sterilized with 0.22 μm diameter needle filters prior to encapsulation and incubation at 37 °C for 2 and 4 weeks. The glycated Tf was obtained by eliminating surplus glucose in sodium phosphate buffer using 10 kDa ultrafiltration tubes membranes. The protein concentrations were assessed using a bicinchoninic acid (BCA) protein quantification kit, and proteins were preserved at − 80 °C. This experiment was performed with four independent biological replicates.

Detection of products of glycated Tf at different stages in different glucose concentrations in vitro

Detection of fructosamine, an early glycated product

The nitroblue tetrazolium reduction assay was used to assess the fructosamine levels26, one of the amadori products representing the initial stages of AGEs formation27. According to the instructions of the fructosamine kit, 10μL of glycated Tf (3 g/L) was incubated with 0.2 mL of chromogenic agent at 37℃ for 15 min. Then, a stop solution was added, and the sample absorbance was recorded at 530 nm using the microplate reader. A calibration curve was generated using fructosamine standards. The fructosamine levels in each group were normalized to fructosamine levels per gram of Tf. This experiment was performed with four independent biological duplicates.

Detection of dicarbonyl compounds, an intermediate glycated product

The levels of dicarbonyl compounds, the products of the second stage of AGEs formation, were determined using the Girard-T assay28. Forty microliters of glycated Tf (3 g/L) were incubated with 20 μL of Girard-T stock solution (500 mM) and 0.34 mL of sodium formate (500 mM, pH 2.9) at room temperature for 1 h. The absorbance at 294 nm was determined using a microplate reader. A calibration curve was constructed using glyoxal as a standard. The glyoxal levels in each group were normalized to glyoxal levels per gram of Tf. This experiment was performed with four independent biological replicates.

Detection of AGEs, a final glycated product

The fluorescence emission spectra of glycated Tf samples (1 g/L) in the wavelength range of 400–600 nm were analyzed using a microplate reader with an excitation wavelength of 370 nm. The fluorescence intensity of the incubation solution at an emission wavelength of 440 nm was used to analyze the abundance of AGE-Tf in the non-enzymatic glycation end products29.The AGE-Tf levels in each group were normalized to AGE-Tf levels per gram of Tf. This experiment was performed with four independent biological replicates.

Determination of TIBC of glycated Tf in vitro under various glucose states

According to the total iron binding capacity (TIBC) kit instructions, 50 μL of glycated Tf (3 g/L) to be measured was mixed with 50 μL of 10 mg/L iron standard solution and left to stand for 5 min at 25 ℃, then iron adsorbent was incorporated and left to stand for 5 min at 25 ℃. 3000 g centrifugation was performed for 10 min, then 50 μL of supernatant was combined with 250 μL of iron chromogenic agent, and 10,000 g centrifugation was performed for 10 min, then 180 μL of supernatant was taken, and absorbance was measured at 520 nm. A calibration curve was prepared using iron as the standard. This experiment was performed with four independent biological replicates.

Human podocytes culture

Human podocytes (CLHT-033–0152, Future Biotech, Beijing, China) were cultured in the 1640 medium containing 10 percent fetal bovine serum in an incubator (maintained at 37℃ with 5% CO2). When the podocyte growth density reached 80% or more, the podocytes were digested and resuspended and used for subsequent experiments.

Setting up of experimental groups

Experimental grouping

To explore the effects of AGE-Tf on podocytes, the experiment comprised the following groups: a negative control (NC) group, 0mMGlu-AGE-Tf group (AGE-Tf produced by incubating Tf with 0 mM glucose), 5.6mMGlu-AGE-Tf group (AGE-Tf produced by incubating Tf with 5.6 mM glucose), 11.1mMGlu-AGE-Tf group (AGE-Tf produced by incubating Tf with 11.1 mM glucose), and 33.3mMGlu-AGE-Tf group (AGE-Tf produced by incubating Tf with 33.3 mM glucose).

Rescue experimental grouping

To explore the mechanisms involved in the effects of AGE-Tf on podocytes, the rescue experimental comprised the following groups: NC, 5.6mMGlu-AGE-Tf group, 5.6mMGlu-AGE-Tf + DFO (deferoxamine mesylate), 5.6mMGlu-AGE-Tf + Fer-1 (Ferrostatin-1), 5.6mMGlu-AGE-Tf + RAP (receptor-for-advanced-glycation-end-products (RAGE) antagonist peptide), 33.3mMGlu-AGE-Tf + DFO, 33.3mMGlu-AGE-Tf + Fer-1, and 33.3mMGlu-AGE-Tf + RAP groups. The intervention concentrations of DFO, Fer-1 and RAP were 100 μM, 2 μM and 10 μM, respectively. DFO, Fer-1 pretreated podocytes for 24 h before AGE-Tf intervention, while RAP treated podocytes concurrently with AGE-Tf.

Intervention concentration and duration of AGE-Tf determined by pre-experimental measurement of cell viability

The pre-testing concentrations of Tf were 0.1, 0.25, 0.5, and 1 g/L, and the duration of the intervention was 24, 48, and 72 h, respectively. The concentration and duration of the AGE-Tf intervention of the final experiment were 1 g/L and 72 h, respectively. This experiment was performed with three independent biological replicates.

Conduction of cell viability assay

Human podocytes were cultured in 96-well plates with 2 × 10^3 cells per well, and different interventions were performed in accordance with the different experimental groups. After reaching the point of intervention, 10 μL of CCK-8 reagent was added. After shaking and mixing, the cells were incubated in a light-proof box for 2 h, and then the absorbance was determined at 450 nm. This experiment was performed with three independent biological replicates.

Detection of cell apoptosis

In accordance with the guildelines of the apoptosis detection kit, Human podocytes were collected and cultured in 6-well plates at a density of 1 × 105 cells per well, with various interventions applied in accordance with the different experimental groups. Following a 72-h intervention, the cell subsequently supernatant was collected, and podocytes were treated with ethylenediaminetetraaceticacid (EDTA)-free trypsin and centrifuged. The cells were washed thrice with pre-cooled phosphate-buffered saline, resuspended with 500µL of 1 × annexin V conjugate (approximately 1 × 106 cells/mL), and treated with 5µL of annexin V-fluorescein isothiocyanate (FITC) and 5µL of propidium iodide (PI) at a concentration of 50 µg/mL. This sample was subsequently mixed gently and incubated at 25 ℃ for 15 min, shielded from light, after which it was promptly examined using a flow cytometer. This experiment was performed with three independent biological replicates.

Determination of indices of oxidative stress and ferroptosis

Human podocytes were collected and cultured in 6-well plates with 1 × 10^5 cells per well, and different interventions were performed in accordance with the different experimental groups. After 72 h of intervention, the podocytes were lysed by radioimmunoprecipitation assay (RIPA) lysis buffer containing phosphatase inhibitors and protease inhibitors. The protein samples were obtained by centrifuging the cells, and the supernatant was gathered. Then, the protein levels were examined. The total-antioxidant-capacity (T-AOC) kit, superoxide-dismutase (SOD) kit, glutathione (GSH) kit, lipid-peroxide (LPO) kit, and malonic-dialdehyde (MDA) kit were used to detect the levels of the respective indices. This experiment was performed with three independent biological replicates.

Determination of the levels of ROS

Human podocytes were collected and cultured in 6-well plates with 1 × 10^5 cells per well, and different interventions were performed in accordance with the different experimental groups. After 72 h of intervention, 1 mL of 2’,7’-dichlorofluorescein diacetate (DCFH-DA) (10 μM) was added according to the ROS kit instructions, and the cells were incubated in a cell culture incubator at 37℃ for 30 min. The adherent podocytes were then digested with EDTA-free trypsin, washed three times with serum-free medium, resuspended in 500 μl of phosphate-buffered saline, and analyzed by flow cytometry. This experiment was performed with three independent biological replicates.

Detection of intracellular Fe2+ levels

After 72 h of intervention, the detection of intracellular Fe2+ was carried out according to the kit instructions, the fluorescence intensity of Fe2+was observed by fluorescence microscopy. This experiment was performed with three independent biological replicates.

Extraction of RNA and conduction of quantitative real-time polymerase chain reaction

Total RNA was extracted with a kit, and cDNA was synthesized with a reverse transcription reagent kit. QRT-PCR was performed using a qRT-PCR system after mixing the cDNA with specific primers using a qRT-PCR kit. The relative expression levels of mRNA were normalized by the housekeeping gene β-actin. The specific primer sequences are shown in Table S2. This experiment was performed with three independent biological replicates.

Extraction of protein and conduction of western blotting

The cellular samples were lysed using RlPA lysis buffer containing protease and phosphatase inhibitors. The protein samples were obtained using centrifugation of the cells, followed by the collection of the supernatant, and the protein concentration was examined. A Loading buffer was subsequently included into the designated volume of proteins, and the vials were immersed in boiling water for 10 min. The proteins were isolated via sodium dodecyl-sulfate polyacrylamide gel electrophoresis and subsequently transferred onto polyvinylidene difluoride membranes. They were subsequently treated to blocking with 5% nonfat milk. The primary antibodies were as follows: GPX4 monoclonal antibody, acyl-CoA-synthetase-long-chain-family-member 4 (ACSL4) monoclonal antibody, nuclear-factor-erythroid-2-related-factor 2 (NRF2) recombinant antibody, SLC7A11/xCT polyclonal antibody, anti-ferritin heavy chain 1(FTH1) antibody, transferrin receptor (TfR1) antibody, and RAGE antibody. All primary antibodies were incubated at 4℃ overnight, after which the corresponding secondary antibody was diluted at a ratio of 1:5,000 and incubated at 25℃ for 1.5 h. Image J was utilized for analyzing the expression levels of target proteins. This experiment was performed with three independent biological replicates.

Statistical analysis

The experimental data were analyzed using IBM SPSS Statistics (version 26.0; Armonk, NY, USA). Results were expressed as mean ± standard deviation (x̄ ± s) from three or four (n = 3 or n = 4) independent biological replicates. The differences were examined using t-tests or one-way ANOVAs. The P values for intra-group multiple comparisons were adjusted using Tukey’s post hoc test. A P value of less than 0.05 was defined as a statistical difference.

Results

Comparison of products of glycated Tf at different stages in different glucose concentrations in vitro

After 2 or 4 weeks of incubation of Tf with glucose, the fructosamine levels increased significantly in the 11.1mMGlu-AGE-Tf and 33.3mMGlu-AGE-Tf groups when compared to the 0mMGlu-AGE-Tf group (P < 0.05 for all groups). The fructosamine levels in 11mMGlu-AG-Tf and 33.3 mM-AGE-Tf groups after 4 weeks of incubation were significantly increased compared with those of 11mMGlu-AGE-Tf and 33.3 mM-AGE-Tf groups after 2 weeks of incubation (P < 0.05 for all groups). (Table 1).

Table 1 Comparison of the degree of Tf glycation and TIBC of AGE-Tf with different glucose concentrations in vitro.
Full size table

After incubation with Tf for 2 weeks, the glyoxal levels increased significantly in the 11.1mMGlu-AGE-Tf and 33.3mMGlu-AGE-Tf groups when compared to the 0mMGlu-AGE-Tf group (P < 0.05 for all groups). Following incubation with Tf for 4 weeks, the glyoxal levels increased significantly in the 33.3mMGlu-AGE-Tf group when compared to the 0mMGlu-AGE-Tf group(P < 0.05 for all groups) (Table 1).

After 2 or 4 weeks of incubation of Tf with glucose, the fluorescence intensity of AGE-Tf increased significantly in the 11.1mMGlu-AGE-Tf and 33.3mMGlu-AGE-Tf groups when compared to the 0mMGlu-AGE-Tf group (P < 0.05 for all groups). The fluorescence intensities of AGE-Tf in 11mMGlu-AG-Tf and 33.3 mM-AGE-Tf groups after 4 weeks of incubation were significantly increased compared with that in 11mMGlu-AGE-Tf and 33.3 mM-AGE-Tf groups after 2 weeks of incubation. (Table 1).

Comparison of the TIBC of AGE-Tf with different glucose concentrations in vitro

After incubation with Tf for 2 or 4 weeks, the levels of TIBC of AGE-Tf significantly decreased in both the 11.1mMGlu-AGE-Tf and the 33.3mMGlu-AGE-Tf groups compared to the 0mMGlu-AGE-Tf group(all P < 0.05). The levels of TIBC of AGE-Tf in both the 11.1mMGlu-AGE-Tf and the 33.3mMGlu-AGE-Tf groups significantly decreased after four weeks of Tf incubation compared to two weeks (all P < 0.05) (Table 1).

The effect of AGE-Tf on human podocyte viability

The effect of AGE-Tf on podocytes viability was explored by pre-experimentation, and the optimal AGE-Tf intervention concentration and intervention time were screened, the results of the study showed that AGE-Tf led to a decrease in podocyte viability as the concentration of glucose in incubated Tf increased, and the higher the concentration of AGE-Tf, as shown in Table S3.

After 72 h of intervention, the cell viability decreased significantly in the 33.3mMGlu-AGE-Tf group compared to the NC, 0mMGlu-AGE-Tf, and 5.6mMGlu-AGE-Tf groups (P < 0.05 for all groups). The cell viability decreased significantly in the 11.1mMGlu-AGE-Tf group compared to the NC and 0mMGlu-AGE-Tf groups but increased compared to the 33.3mMGlu-AGE-Tf group (P < 0.05 for all groups) (Fig. 1A).

Fig. 1
figure 1

The effect of AGE-Tf on human podocyte viability. (A), (B), and (C): Comparison of podocyte viability among different groups. AGE-Tf Advanced glycated end-products modified transferrin, NC Negative control, 0mMGlu-AGE-Tf Glucose concentration is 0 mM when preparing AGE-Tf, 5.6mMGlu-AGE-Tf Glucose concentration is 5.6 mM when preparing AGE-Tf, 11.1mMGlu-AGE-Tf Glucose concentration is 11.1 mM when preparing AGE-Tf, 33.3mMGlu-AGE-Tf Glucose concentration is 33.3 mM when preparing AGE-Tf, DFO deferoxamine mesylate, RAP receptor-for-advanced-glycation–end-productsantagonist peptide, Fer-1 Ferrostatin-1. Data are presented as mean ± SD (n = 3 for all panels) from independent biological replicates; *p < 0.05;**p < 0.01; ***p < 0.001; ****p < 0.0001.

Full size image

The results of the rescue experiment showed that after 72 h of intervention, there was no significant difference in cell viability among 5.6mMGlu-AGE-Tf, 5.6mMGlu-AGE-Tf + DFO, 5.6mMGlu-AGE-Tf + Fer-1, and 5.6mMGlu-AGE-Tf + RAP groups (Fig. 1B). The cell viability was significantly higher in the 33.3mMGlu-AGE-Tf + Fer-1 and 33.3mMGlu-AGE-Tf + RAP groups compared to the 33.3mMGlu-AGE-Tf group (P < 0.05 for all groups). No significant differences were observed in the cell viability between 33.3 and 33.3mMGlu-AGE-Tf + DFO groups (Fig. 1C).

The effect of AGE-Tf on apoptosis of human podocytes

The apoptosis rate of human podocytes treated by AGE-Tf for 72 h was determined using a flow cytometer. Compared to the NC group, the apoptosis rates were significantly increased in the 11.1mMGlu-AGE-Tf and 33.3mMGlu-AGE-Tf groups (P < 0.05 for all groups). Compared to the 0mMGlu-AGE-Tf, 5.6mMGlu-AGE-Tf, and 11.1mMGlu-AGE-Tf groups, the apoptosis rate was significantly increased in the 33.3mMGlu-AGE-Tf group (P < 0.05 for all groups) (Fig. 2A).

Fig. 2
figure 2

The effect of AGE-Tf on apoptosis of human podocytes. (A), (B), and (C): Comparison of apoptosis of human podocytes among different groups. AGE-Tf Advanced glycated end-products modified transferrin, NC Negative control, 0mMGlu-AGE-Tf Glucose concentration is 0 mM when preparing AGE-Tf, 5.6mMGlu-AGE-Tf Glucose concentration is 5.6 mM when preparing AGE-Tf, 11.1mMGlu-AGE-Tf Glucose concentration is 11.1 mM when preparing AGE-Tf, 33.3mMGlu-AGE-Tf Glucose concentration is 33.3 mM when preparing AGE-Tf, DFO deferoxamine mesylate, RAP receptor-for-advanced-glycation–end-productsantagonist peptide, Fer-1 Ferrostatin-1. Data are presented as mean ± SD (n = 3 for all panels) from independent biological replicates; *p < 0.05;**p < 0.01; ***p < 0.001; ****p < 0.0001.

Full size image

The results of the rescue experiment showed that there was no significant difference in apoptosis among the 5.6mMGlu-AGE-Tf, 5.6mMGlu-AGE-Tf + DFO, 5.6mMGlu-AGE-Tf + Fer-1, and 5.6mMGlu-AGE-Tf + RAP groups (Fig. 2B). Compared to the 33.3mMGlu-AGE-Tf group, the apoptosis rates were significantly decreased in the 33.3mMGlu-AGE-Tf + Fer-1 and 33.3mMGlu-AGE-Tf + RAP groups (P < 0.05 for all groups). No significant differences were observed in the apoptosis rates between 33.3mMGlu-AGE-Tf and 33.3mMGlu-AGE-Tf + DFO groups (Fig. 2C).

The effect of AGE-Tf on oxidative stress in human podocytes

The levels of T-AOC in the 33.3mMGlu-AGE-Tf group decreased significantly compared to the NC, 0mMGlu-AGE-Tf, 5.6mMGlu-AGE-Tf, and 11.1mMGlu-AGE-Tf groups (P < 0.05 for all groups), (Fig. 3A). The levels of SOD and GSH in the 33.3mMGlu-AGE-Tf group decreased significantly compared to the NC, 0mMGlu-AGE-Tf, and 5.6mMGlu-AGE-Tf groups (P < 0.05 for all groups) (Fig. 3B and C). The levels of LPO and MDA in the 33.3mMGlu-AGE-Tf group increased significantly compared to the NC, 0mMGlu-AGE-Tf, 5.6mMGlu-AGE-Tf, and 11.1mMGlu-AGE-Tf groups (P < 0.05 for all groups) (Fig. 3D and E). Compared to the 0mMGlu-AGE-Tf, 5.6mMGlu-AGE-Tf, and 11.1mMGlu-AGE-Tf groups, the levels of ROS increased significantly in the 33.3mMGlu-AGE-Tf group (P < 0.05 for all groups) (Fig. 4).

Fig. 3
figure 3

The effect of AGE-Tf on oxidative stress in human podocytes. (A), (B), (C), (D), and (E): Comparison of T-AOC, SOD, GSH, LPO, and MDA levels among different groups, respectively. AGE-Tf Advanced glycated end-products modified transferrin, NC Negative control, 0mMGlu-AGE-Tf Glucose concentration is 0 mM when preparing AGE-Tf, 5.6mMGlu-AGE-Tf Glucose concentration is 5.6 mM when preparing AGE-Tf, 11.1mMGlu-AGE-Tf Glucose concentration is 11.1 mM when preparing AGE-Tf, 33.3mMGlu-AGE-Tf Glucose concentration is 33.3 mM when preparing AGE-Tf, DFO deferoxamine mesylate, RAP receptor-for-advanced-glycation-end-productsantagonist peptide, Fer-1 Ferrostatin-1. Data are presented as mean ± SD (n = 3 for all panels) from independent biological replicates;*p < 0.05;**p < 0.01; ***p < 0.001; ****p < 0.0001.

Full size image
Fig. 4
figure 4

The effect of AGE-Tf on ROS in human podocytes. (A), (B), and (C): Comparison of ROS levels of human podocytes among different groups. AGE-Tf Advanced glycated end-products modified transferrin, NC Negative control, 0mMGlu-AGE-Tf Glucose concentration is 0 mM when preparing AGE-Tf, 5.6mMGlu-AGE-Tf Glucose concentration is 5.6 mM when preparing AGE-Tf, 11.1mMGlu-AGE-Tf Glucose concentration is 11.1 mM when preparing AGE-Tf, 33.3mMGlu-AGE-Tf Glucose concentration is 33.3 mM when preparing AGE-Tf, DFO deferoxamine mesylate, RAP receptor-for-advanced-glycation-end-productsantagonist peptide, Fer-1 Ferrostatin-1. *p < 0.05;**p < 0.01; ***p < 0.001; ****p < 0.0001.

Full size image

The results of the rescue experiment showed that the levels of T-AOC and GSH were significantly higher in 33.3mMGlu-AGE-Tf + RAP group compared to the 33.3mMGlu-AGE-Tf (P < 0.05 for all groups) (Fig. 3A and B). In addition, no significant differences were observed in the levels of SOD between 33.3mMGlu-AGE-Tf and 33.3mMGlu-AGE-Tf + RAP groups. Figure 3C, the levels of LPO and MDA in the 33.3mMGlu-AGE-Tf group were significantly higher compared to 33.3mMGlu-AGE-Tf + Fer-1 and 33.3mMGlu-AGE-Tf + RAP groups (P < 0.05 for all groups). Figure 3D and E, the ROS levels were significantly lower in the 33.3mMGlu-AGE-Tf + Fer-1 and 33.3mMGlu-AGE-Tf + RAP groups compared to the 33.3mMGlu-AGE-Tf (P < 0.05 for all groups) (Fig. 4).

Effects of AGE-Tf on mRNA levels of related genes of ferroptosis in human podocytes

The expression levels of mRNA of GPX4, NRF2, and SLC7A11 in the 33.3mMGlu-AGE-Tf group decreased significantly compared to the NC, 0mMGlu-AGE-Tf, and 5.6mMGlu-AGE-Tf groups (P < 0.05 for all groups) (Figure S1A, S1C and S1D, respectively), and the levels of mRNA of GPX4 and NRF2 in the 33.3mMGlu-AGE-Tf group decreased significantly compared to the 11.1mMGlu-AGE-Tf group. The expression levels of mRNA of ACSL4 in the 33.3mMGlu-AGE-Tf group increased significantly compared to the NC, 0mMGlu-AGE-Tf, 5.6mMGlu-AGE-Tf, and 11.1mMGlu-AGE-Tf groups (P < 0.05 for all groups) (Figure S1B).

Effects of AGE-Tf on protein levels of related genes of ferroptosis in human podocytes

The expression levels of proteins of GPX4, NRF2, and SLC7A11 in the 33.3mMGlu-AGE-Tf group decreased significantly compared to the NC, 0mMGlu-AGE-Tf, and 5.6mMGlu-AGE-Tf groups (P < 0.05 for all groups) (Fig. 5A, C, and D, respectively). The expression levels of protein of ACSL4 in the 33.3mMGlu-AGE-Tf group increased significantly compared to the NC, 0mMGlu-AGE-Tf, and 5.6mMGlu-AGE-Tf groups (P < 0.05 for all groups) (Fig. 5B).

Fig. 5
figure 5

The expression levels of protein of GPX4, ACSL4, NRF2, SLC7A11, FTH1 and TFR1. (A): the expression levels of protein of GPX4, (B): the expression levels of protein of ACSL4, (C): the expression levels of protein of NRF2, (D): the expression levels of protein of SLC7A11, (E): the expression levels of protein of FTH1, (F): the expression levels of protein of TFR1, (G): the expression levels of protein of RAGE. AGE-Tf Advanced glycated end-products modified transferrin, NC Negative control, 0mMGlu-AGE-Tf Glucose concentration is 0 mM when preparing AGE-Tf, 5.6mMGlu-AGE-Tf Glucose concentration is 5.6 mM when preparing AGE-Tf, 11.1mMGlu-AGE-Tf Glucose concentration is 11.1 mM when preparing AGE-Tf, 33.3mMGlu-AGE-Tf Glucose concentration is 33.3 mM when preparing AGE-Tf, DFO deferoxamine mesylate, RAP receptor-for-advanced-glycation-end-productsantagonist peptide, Fer-1 Ferrostatin-1. Data are presented as mean ± SD (n = 3 for all panels) from independent biological replicates; *p < 0.05;**p < 0.01; ***p < 0.001; ****p < 0.0001.

Full size image

The results of the rescue experiment showed that the expression levels of GPX4, NRF2, and SLC7A11 in the 33.3mMGlu-AGE-Tf + Fer-1 and 33.3mMGlu-AGE-Tf + RAP groups were significantly higher compared to the 33.3mMGlu-AGE-Tf group (P < 0.05 for all groups), and no significant differences were observed in the levels of GPX4, NRF2, and SLC7A11 between 33.3mMGlu-AGE-Tf and 33.3mMGlu-AGE-Tf + DFO groups (Fig. 5A, C, and D, respectively). The expression levels of protein of ACSL4 in the 33.3mMGlu-AGE-Tf + Fer-1 and 33.3mMGlu-AGE-Tf + RAP groups were significantly lower compared to the 33.3mMGlu-AGE-Tf group, and no significant differences were observed in the levels of ACSL4 between 33.3 and 33.3mMGlu-AGE-Tf + DFO groups (P < 0.05) (Fig. 5B).

The Effect of AGE-Tf on the expression levels of FTH1 and TFR1 genes in human podocytes

No significant differences were observed in the expression levels of mRNA of FTH1 (Figure S1E) and TFR1 (Figure S1F) among NC, 0mMGlu-AGE-Tf, 5.6mMGlu-AGE-Tf, 11.1mMGlu-AGE-Tf, and 33.3mMGlu-AGE-Tf groups. No significant differences were observed in the expression levels of protein of FTH1 (Fig. 5E) and TFR1 (Fig. 5F) among NC, 0mMGlu-AGE-Tf, 5.6mMGlu-AGE-Tf, 11.1mMGlu-AGE-Tf, and 33.3mMGlu-AGE-Tf groups.

The Effect of AGE-Tf on the levels of RAGE and Fe2+ in human podocytes

Compared with the NC group, RAGE levels were significantly higher in the 33.3mMGlu-AGE-Tf, whereas RAGE levels decreased in the 33.3mMGlu-AGE-Tf + RAP group compared with the 33.3mMGlu-AGE-Tf group (Fig. 5G).

Meanwhile, the results of the study showed that there was no significant difference in intracellular Fe2 + levels between NC, 5.6mMGlu-AGE-Tf, and 33.3mMGlu-AGE-Tf groups, the intracellular Fe2+ levels showed a decreasing trend in 33.3mMGlu-AGE-Tf + DFO group compared to 33.3mMGlu-AGE-Tf group, but no significant difference was observed; and there was no significant trend of change between 33.3mMGlu-AGE-Tf and 33.3mMGlu-AGE-Tf + Fer-1 and 33.3mMGlu-AGE-Tf + RAP groups (Figure S2).

Discussion

The present study incubated Tf in vitro with different concentrations of glucose (0 mMGlu, 5.6 mMGlu, 11.1 mMGlu, and 33.3 mMGlu) for durations of 2 and 4 weeks. The results showed that the levels of fructosamine, glyoxal, and AGE-Tf escalated in correlation with higher glucose concentration, with the most significant increase being observed at the glucose concentration of 33.3 mM. Meanwhile, with increasing glucose concentration, a gradually increasing glycation of Tf leads to a decrease in the TIBC of Tf.

Subsequently, podocytes were intervened in vitro using AGE-Tf. The results showed that compared to the NC group, the cell viability diminished, apoptosis escalated, ferroptosis and oxidative stress were triggered, RAGE levels rose, and there was no significant alteration in intracellular Fe2+ in 33.3mMGlu-AGE-Tf group. Rescue experiments showed that podocyte viability and apoptosis rates did not recover with the addition of DFO in 33.3mMGlu-AGE-Tf group, oxidative stress and ferroptosis were not inhibited, suggesting that the effect of AGE-Tf on podocytes may not be mediated via iron overload. Under our specific experimental conditions, the iron-chelating function of DFO was insufficient to counteract AGE-Tf-induced cytotoxicity. A plausible explanation is that the standard culture medium, although containing iron, may not constitute an iron-replete environment capable of eliciting significant iron overload upon impairment of Tf iron-binding function. Thus, the primary mechanism driving ferroptosis appears to be RAGE-mediated signaling rather than iron-dependent pathways30. In contrast, Fer-1 inhibited ferroptosis and restored podocyte viability without altering intracellular Fe2⁺ levels. Similarly, RAP reduced RAGE expression, suppressed ferroptosis, and ameliorated cellular damage independently of Fe2⁺ concentrations, indicating that AGE-Tf promotes injury through RAGE-dependent activation of ferroptosis.

A key novel finding of is that AGE-Tf-induced ferroptosis occurs independently of the iron overload mechanism we had previously hypothesized, as DFO failed to reverse the ferroptosis. This suggests an alternative mechanism whereby AGE-Tf binding to RAGE directly initiates molecular drivers of ferroptosis. RAGE activation exacerbates oxidative stress and suppresses the NRF2 antioxidant pathway31,32, a central regulator of ferroptosis that modulates GPX4 and SLC7A11 expression33. Our data show that RAP restored NRF2, GPX4, and SLC7A11 levels and inhibited lipid peroxidation (LPO and MDA) without affecting intracellular Fe2⁺, robustly supporting this concept. We propose that AGE-Tf binding to RAGE primarily drives ferroptosis by compromising antioxidant defenses and promoting lipid peroxidation, rather than through iron-mediated Fenton chemistry.

Individuals with DM have been shown to exhibit increased non-enzymatic glycation of serum Tf, which increases from 1–2% in a healthy population to 5% in adults with DM34, and glycation of serum Tf in healthy and type 1 diabetic children were 4% and 11%, respectively35. Dextran gel electrophoresis demonstrated higher glycation of serum albumin and Tf in patients with DM34. In addition, through liquid chromatography-mass spectrometry (LC–MS), Soboleva et al. identified six glycation sites (K553, K659, K258, K299, K315, and K225) in plasma Tf of patients with T2DM compared to healthy adults36. Similarly, Makan et al. detected six lysines (K103, K206, K276, K296, K534, and K640) and one arginine residue (R678) modified by glycation7.

Disturbances in glycolipid metabolism in DM lead to an acceleration of AGEs formation37, which accumulate in various renal compartments, including the glomerular basement membrane, mesangial cells, endothelial cells, and podocytes38. The binding of AGEs to the receptor for advanced glycation end products (RAGE) has been implicated the pathogenesis and progression of DKD39. In this study, stimulation of human podocytes with AGE-Tf resulted in increased apoptosis, elevated ROS, and decreased antioxidant capacity (T-AOC, SOD, and GSH). Notably, these effects were most pronounced with the 33.3 mM Glu-AGE-Tf, correlating with its highest degree of glycation, highlighting the structure–function relationship of the modified protein.

The AGE/RAGE pathway is hypothesized to stimulate ROS production, which may subsequently promote oxidative stress by regulating signaling pathways such as AKT, NRF2, and mammalian target of rapamycin (mTOR)31,32 , contributing to renal damage in DKD40. Furthermore, AGEs and their downstream signaling disrupt renal architecture and impair kidney function via matrix protein cross-linking41. AGEs-induced structural changes impair the degradation of AGE-modified proteins by metalloproteinases, resulting in the thickening of the basement membrane, which is a known histological feature of DKD42. Immunohistochemical analysis of kidney samples indicated that both podocytes and renal tubular cells express RAGE. Thus, binding to AGEs induces pathological changes in the cells (primarily by promoting ROS production). This contributes to chronic diabetes-related renal damage such as glomerulosclerosis and tubulointerstitial fibrosis43. Our data delineate a potential pathway wherein AGE-Tf binding to RAGE triggers oxidative stress, dysregulating key ferroptosis mediators. Observed NRF2 suppression may be a critical node, as oxidative stress inhibits NRF2 signaling, creating a vicious cycle that amplifies damage and culminates in ferroptosis20,44. We further confirmed that RAP treatment attenuated AGE-Tf-induced oxidative stress, supporting the involvement of RAGE signaling.

The buildup of ROS and Fe2 + within cells is a critical component that induces ferroptosis45. Kim et al. discovered a correlation between ferroptosis and DKD, noting that the expression levels of ferroptosis-related genes, such as SlC7A11 and GPX4 were diminished in kidneys of patients with DKD compared to those without DKD23. Zhao et al. identified an association between DKD and serum GPX4 and ACSL4 levels2. In addition, it has been reported that AGEs induce osteoblastic ferroptosis, leading to osteoporosis46 and increased intracellular free iron in osteoblastic cells, with the potential correlation between AGEs levels and iron overload found in previous studies47,48. The above studies showed a correlation between AGEs and ferroptosis, nevertheless, to our knowledge, the present study is the first to explore the effects of AGE-Tf on ferroptosis in human podocytes. The present study revealed that the addition of 33.3mMGlu-AGE-Tf to human podocytes resulted in an increase in the levels of ferroptosis-related indices LPO and MDA, a decrease in the expression levels of the related genes GPX4, NRF2, and SLC7A11, and an increase in the levels of ACSL4. Fer-1 treatment reversed these changes and improved cell viability, providing functional evidence for ferroptosis involvement. RAP also inhibited AGE-Tf-induced ferroptosis, further supporting a RAGE-dependent mechanism.

Notably, several previous studies reported that the TIBC of glycated Tf was decreased12,49,50. We observed the similar results in the study, and non-transferrin-bound iron (NTBI) production has been detected in patients with T2DM49. Therefore, we initially hypothesized that AGE-Tf might also promote oxidative stress and iron metabolic responses by inducing NTBI and iron overload. We combined 33.3mMGlu-AGE-Tf and DFO together to intervene in the podocytes, and the results showed that compared with the 33.3mMGlu-AGE-Tf group, the addition of DFO showed that ferroptosis and oxidative stress induced by AGE-Tf did not have a significant changes, and the intracellular Fe2+ tended to decrease, but there was no statistically significant difference. While this finding does not support iron overload as the main mechanism. In our study, treating with the iron chelator DFO did not reduce AGE-Tf–triggered ferroptosis, nor did it restore the diminished antioxidant capacity. One explanation is that glycated Tf causes ferroptosis and oxidative stress mainly through direct binding of AGE to its receptor RAGE, an effect independent of intracellular iron overload that directly promotes oxidative stress and ferroptosis. Another explanation may be that the in vitro system may not have provided sufficient bioavailable iron to podocytes: both Tf and AGE-Tf likely sequestered the available iron in the culture medium, so podocytes could not exhibit functional differences in TIBC among Tf with different glycation levels. Thus, adding DFO could not reverse the levels of iron and ferroptosis markers51,52. In fact, a single in vitro experiment like this cannot completely rule out iron-mediated effects, because the in vitro model may not reproduce the long-term disturbances of iron metabolism present in diabetic patients53. Future studies should add iron together with AGE-Tf in vitro to simulate functional iron overload and should validate findings in diabetic animal models to assess the RAGE-mediated pathway and possible iron-related pathways.

Although this study reveals novel insights, several issues require discussion. Firstly, the in vitro model employed lacks the physiological complexity of the diabetic microenvironment, including hemodynamic forces, metabolic variations, and intercellular communication54, which may modulate the effects of AGE-Tf. Secondly, this negative result with DFO suggests that the ferroptosis induced by AGE-Tf is primarily driven by RAGE-dependent signaling pathways rather than iron overload in vitro. However, this conclusion requires further validation. On the one hand, the iron concentration in our culture system may have been too low to reveal the impact of decreased TIBC of from glycation-modified Tf. The culture system may not have provided enough iron to podocytes, so there were no obvious differences in NTBI or intracellular iron between groups, which could mask any iron-mediated effects that DFO might chelate. On the other hand, the results showed, the 3-day intervention in this study, although able to activate RAGE signaling, may be too short to model the chronic iron metabolism disturbances that develop over months or years in diabetic patients. The iron overload phenotype might require a longer duration to develop. Therefore, although RAGE activation appears to be the direct mechanism for AGE-Tf–induced ferroptosis, iron overload could still play a contributing role under different conditions (e.g., high iron load or chronic hyperglycemic exposure), this needs further investigation. Subsequent work will employ iron-supplemented media and DM models to address this. Finally, the definitive proof of the AGE-Tf/RAGE/ferroptosis axis in diabetic kidney disease requires validation in animal models, such as diabetic mice treated with RAGE antagonists or ferroptosis inhibitors, which represents an essential direction for our future research.

Conclusion

High glucose-induced Tf glycation in vitro may induce oxidative stress, regulates the expression of ferroptosis-related genes, and accelerates apoptosis of human podocytes via the RAGE pathway.