Diosgenin presents a novel role in promoting diabetic wound healing: A mechanism involving Sirt6/Nrf2-mediated inhibition of ferroptosis

Abstract

Delayed wound healing in patients with diabetes is frequently caused by the dysfunction of human umbilical vein endothelial cells (HUVECs). Under high-glucose and high-fat (HGHF) conditions, the microenvironment induces uncontrolled lipid peroxidation and subsequent iron-dependent ferroptosis. Diosgenin (DG), a steroidal saponin extracted from herbs, including wild yam, exhibits diverse pharmacological activities. The present study aimed to investigate the role and underlying mechanism of DG in HGHF-induced ferroptosis and diabetic wound (DW) healing. A DW model was established in mice by intraperitoneal injection of streptozotocin (STZ). HUVECs were cultured under HGHF conditions to induce the diabetic microenvironment. C11-BODIPY, FerroOrange probes, and hematoxylin-eosin (H&E)/Masson staining were used to determine lipid peroxidation, iron ion levels, and histological changes, respectively. Results demonstrated that DG alleviated HGHF-induced ferroptosis, as evidenced by reduced malondialdehyde (MDA), lipid peroxidation, and mitochondrial reactive oxygen species (ROS) levels, and increased glutathione (GSH) levels. Additionally, DG upregulated the expression of the anti-ferroptotic protein glutathione peroxidase 4 (GPX4), promoted angiogenesis, and accelerated wound healing in diabetic mice. Network pharmacology analysis identified the sirtuin 6/nuclear factor erythroid 2-related factor 2 (Sirt6/Nrf2) pathway as a potential target of DG, a finding validated by mechanistic studies. Silencing Sirt6 in HUVECs negated the protective effect of DG against ferroptosis. In conclusion, DG mitigates ferroptosis and promotes DW healing, at least partially through activating the Sirt6/Nrf2 pathway.

Graphical abstract

Highlights

• •HGHF induce ferroptosis in HUVECs in vitro and promote ferroptosis in diabetic wound tissues in vivo.
• •Diosgenin inhibits HGHF-induced HUVECs ferroptosis and restores HUVECs functional impairment in vitro.
• •Diosgenin exerts anti-ferroptotic effects by activating the Sirt6/Nrf2 signaling pathway.
• •Diosgenin promotes diabetic wound healing in vivo in a Sirt6-dependent manner.

1. Introduction

Diabetic wounds (DW), which exhibit impaired healing, significantly increase the risk of disability and mortality among affected individuals and remain a substantial clinical burden globally [1]. The management strategies, including glycemic control, wound therapy, and surgical debridement, have improved significantly; however, these lesions continue to impose significant socioeconomic burdens. The delayed healing process involves excessive production of reactive oxygen species (ROS) and pro-inflammatory cytokines [2]. Notably, high glucose and fatty acid levels trigger excessive ROS production in the endothelium, resulting in endothelial dysfunction. Studies suggest that glucolipid metabolism significantly contributes to diabetic pathology [3].

Multiple studies implicate that dysregulated glucolipid metabolism can result in ferroptosis [4]. Ferroptosis is a regulated cell death process characterized by iron-dependent lipid peroxidation [5]. Its hallmarks, including iron excess, high ROS levels, and subsequent lipid peroxidation, distinguish it from other forms of cell death [6]. Notably, the small-molecule compound ferrostatin-1 (Fer-1) can effectively inhibit ferroptosis by eliminating lipid peroxides [7]. Ferroptosis has been associated with multiple diseases, including cancer, metabolic disorders, neurodegenerative conditions, ischemia-reperfusion injury, and DW healing [8,9]. In the latter context, mitochondrial dysfunction and ROS overproduction are known to impede recovery, suggesting that ferroptosis can contribute to its pathophysiology [10]. However, the role of ferroptosis in human umbilical vein endothelial cells (HUVECs) within diabetic microenvironments remains unclear. Therefore, therapeutic targeting of ferroptosis holds significant potential for improving the treatment of DW and other diabetic complications.

Sirtuin 6 (Sirt6) is a fundamental cellular protein responsible for repair, maintenance, and energy efficiency, and it addresses aging and disease through DNA repair, telomere protection, metabolic regulation, and inflammation suppression [11,12]. Its collaboration with nuclear factor erythroid 2-related factor 2 (Nrf2) is highly significant [13,14]. Studies indicate that activation of the Sirt6/Nrf2 signaling pathway effectively mitigates oxidative stress and inhibits ferroptosis [15]. Ma et al. [16] exemplified this mechanism by demonstrating that pathway activation upregulates the anti-ferroptosis protein glutathione peroxidase 4 (GPX4), thereby mitigating high-glucose-induced ferroptosis, a finding paralleled in osteoarthritis by Shi et al. [17].

Diosgenin (DG), a naturally occurring steroidal saponin derived from traditional Chinese herbs, including wild yam and Tribulus terrestris, has been recognized as a major bioactive constituent of these plants [18]. Previous studies have demonstrated that DG exhibits diverse pharmacological properties, including anti-inflammatory, antioxidant, and anti-apoptotic activities [19,20]. Notably, it has been reported that DG modulates the Sirt6/Nrf2 pathway to exert protective effects in diabetes-associated conditions, including non-alcoholic fatty liver disease and diabetic nephropathy [21,22]. However, despite these advances, the specific role of DG in DW healing remains largely unknown. In particular, the effect of DG on ferroptosis, a key pathological process involved in diabetic vascular complications within the high-glucose and high-fat (HGHF) microenvironment of impaired wounds, has not yet been investigated. This knowledge gap highlights the need to investigate DG's function beyond previously studied contexts and to elucidate its potential mechanism for targeting ferroptosis during wound healing. Therefore, the present study investigated the effects of DG on HUVECs under HGHF conditions in a diabetic mouse wound model. Our findings revealed that DG effectively suppressed ferroptosis by reducing oxidative stress, improving mitochondrial function, and activating the Sirt6/Nrf2 signaling pathway. These results not only expand the known functions of DG to the context of DW repair but also, for the first time, establish its anti-ferroptotic effect through Sirt6/Nrf2 as a novel mechanism underlying its therapeutic potential. Therefore, DG represents a promising candidate for DW treatment.

2. Materials and methods

2.1. Ethical approval

This study adhered to the National Institutes of Health's Guide for the Care and Use of Laboratory Animals. The Institutional Animal Care and Use Committee at Wenzhou Medical University approved the protocol (Approval No.: wydw2023-0040). Every measure was implemented to reduce animal suffering during surgery, treatment, and postoperative care.

2.2. Reagents and antibodies

DG (HY-N0177, 99.92% purity), erastin (ferroptosis inducer), Fer-1 (selective ferroptosis inhibitor), and OSS_128167 (an inhibitor of Sirt6) were obtained from MedChemExpress (Shanghai, China). Primary antibodies obtained from Abcam (Cambridge, UK) included acyl-coenzyme a synthetase long-chain family member 4 (ACSL4) (ab155282, 1:1000), solute carrier family 7 member 11 (SLC7A11) (ab307601, 1:1000), GPX4 (ab252833, 1:1000), mitofusin 1 (MFN1) (ab221661, 1:1000), mitofusin 2 (MFN2) (ab205236, 1:1000), Sirt6 (ab191385, 1:1000), kelch-like ech-associated protein 1 (Keap1) (ab119403, 1:1000), and β-actin (ab6276, 1:3000). Additional antibodies obtained from Proteintech Group (Chicago, IL, USA) included Nrf2 (16396-1-AP, 1:1000), optic atrophy protein 1 (OPA1) (27733-1-AP, 1:1000), dynamin-related protein 1 (DRP1) (12957-1-AP, 1:1000), cluster of differentiation 31 (CD31) (11265-1-AP, 1:1000), α-smooth muscle actin (α-SMA) (14395-1-AP, 1:1000), vascular endothelial growth factor A (VEGF-A) (19003-1-AP, 1:1000), and anti-Lamin B (12987-1-AP, 1:2000). 4′,6-Diamidino-2-phenylindole (DAPI) fluorescent dye was obtained from Beyotime (Shanghai, China). Other reagents and their sources were as follows: the 5-Ethynyl-2′-deoxyuridine (EdU) assay kit (Elabscience, Wuhan, China), C11-BODIPY (Invitrogen, CA, USA), FerroOrange (Tongren, Beijing, China), phosphate-buffered saline (PBS), fetal bovine serum (FBS), 10% trypsin, and dulbecco's modified eagle's dedium (DMEM) (Gibco, New York, NY, USA). All additional chemicals were obtained from Sigma-Aldrich (St. Louis, MO, USA).

2.3. Cell culture

HUVECs obtained from the American Type Culture Collection (Manassas, VA, USA) were cultured under standard conditions. Cells were cultured in DMEM with 10% heat-inactivated FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin at 37 °C in a 5% CO₂ humidified incubator. Cells were cultured in T-75 flasks (Falcon, Corning, NY, USA) at a density of 2500–3000 cells/cm² for passaging. The culture medium was replaced every 48 h, and confluence was determined daily by light microscopy, typically achieved after 6–7 days of culture.

2.4. Cell viability assay

HUVECs' viability was determined using a Cell Counting Kit-8 (CCK-8) assay (Elabscience, Wuhan, China) following the manufacturer's instructions. Following treatment, the culture medium was replaced with a working solution containing 10% CCK-8 reagent and incubated for 1–2 h. After incubating for 1 h at 37 °C, the absorbance was measured at 450 nm using a Thermo Fisher Scientific microplate reader (Waltham, MA, USA).

2.5. EdU staining assay

HUVECs' proliferation was assessed using an EdU Cell Proliferation Kit (Beyotime, Shanghai, China) following the manufacturer's instructions. Cells were labeled with EdU for 4 h, fixed with 4% paraformaldehyde (PFA) for 30 min, and stained with Hoechst 33342 (Beyotime, Shanghai, China). Fluorescent images were captured for analysis.

2.6. C11-BODIPY and FerroOrange staining

The levels of intracellular lipid peroxidation and Fe²⁺ were detected using C11-BODIPY (Invitrogen, Carlsbad, CA, USA) and FerroOrange probes (Tongren, Beijing, China), respectively. Treated HUVECs were stained with 1 μM of the respective probe for 30 min. Cells were washed with PBS and incubated with FerroOrange in the dark. Confocal imaging was carried out on a Nikon microscope (Nikon, Tokyo, Japan).

2.7. Malondialdehyde (MDA) and glutathione (GSH) activities assay

HUVECs were washed twice with PBS to remove residual culture medium and extracellular impurities before undergoing lysis. The cells were lysed on ice for 30 min for efficient cell disruption while minimizing protein degradation and enzymatic activity loss. Following lysis, the concentrations of MDA, a marker of lipid peroxidation, and the activity of GSH, a key intracellular antioxidant, were quantified using Sigma-Aldrich assay kits (Sigma Aldrich, St. Louis, MO, USA), following the manufacturer's protocols to ensure reproducibility and accuracy.

2.8. Western blot (WB)

WB analysis was performed according to standard procedures. HUVECs were washed with ice-cold PBS and lysed using radio immunoprecipitation assay (RIPA) buffer (Beyotime, Shanghai, China) containing protease and phosphatase inhibitors (MilliporeSigma, Burlington, MA, USA). The cells were centrifuged at 12,000 g for 10 min at 4 °C, and the supernatants were collected. Proteins (20 μg per lane) were then separated using 10% sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) and transferred to polyvinylidene fluoride (PVDF) membranes (MilliporeSigma, Burlington, MA, USA). Membranes were blocked using 5% skim milk (Beyotime, Shanghai, China) and incubated with primary antibodies overnight at 4 °C. Following washing, the membranes were incubated at room temperature for 1 h with horseradish peroxidase (HRP)-conjugated secondary antibodies (Beyotime, Shanghai, China) at a 1:5000 dilution. Protein bands were detected using enhanced chemiluminescence (ECL) (Beyotime, Shanghai, China) and recorded with a Bio-Rad ChemiDoc XRS + system (Bio-Rad, Foster City, CA, USA). Image Lab 6.0 software (Bio-Rad, Foster City, CA, USA) was used to quantify band intensity, utilizing β-actin as the reference.

2.9. Transmission electron microscopy (TEM)

TEM was used to examine the cells' microstructure. The cells were fixed overnight at 4 °C and then treated with 1% osmium tetroxide for 30 min, dried using a series of ethanol baths, and embedded. Uranyl acetate and lead citrate were added to ultrathin slices and observed using a Hitachi field-emission TEM (Hitachi High-Technologies Corp, Tokyo, Japan).

2.10. Evaluation of ROS and mitochondrial function

ROS, mitochondrial ROS, and mitochondrial membrane potential (MMP) were evaluated using 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA) (green) (Thermo Fisher Scientific, Waltham, MA, USA), MitoSox Red (Thermo Fisher Scientific, Waltham, MA, USA), and JC-1 (Thermo Fisher Scientific, Waltham, MA, USA) probes, respectively. Cells were incubated in the dark at 37 °C for 30 min with either 5 μM DCFH-DA, 2.5 μM MitoSox, or 5 μM JC-1. Fluorescence imaging was performed using a Nikon microscope (Nikon, Tokyo, Japan) following PBS washes. In JC-1 assays, red fluorescence indicates increased MMP (polarized mitochondria), whereas green fluorescence denotes decreased MMP (depolarized mitochondria).

2.11. In vitro tube formation assay

A tube formation assay using matrix coagulation was performed to determine the in vitro angiogenic potential of HUVECs. μ-Slide plates were coated with extracellular matrix (ECM) gel solution (Sigma Aldrich, St. Louis, MO, USA) and solidified at 37 °C for 1 h. Following trypsinization, pretreated HUVECs were seeded onto the polymerized gel at a density of 1× 10⁴ cells per well. The plates were incubated at 37 °C for 6 h to allow tube formation. Phase-contrast microscopy captured representative images of tube structures. Angiogenic activity was quantified in ImageJ by measuring total tube length and the number of branching points per field.

2.12. Cell adhesion assay

HUVECs adhesion was assessed on fibronectin-coated (5 μg/mL) 6-well plates (Sigma Aldrich, St. Louis, MO, USA). Cells (1× 10⁴ cells/mL) were seeded per well and allowed to adhere for 30 min following treatment with DG, HGHF, and Fer-1. After PBS washes to eliminate non-adherent cells, the adherent cells were fixed with 4% PFA, stained with DAPI, and quantified by counting cells in three random fields per well using fluorescence microscopy (Nikon, Tokyo, Japan).

2.13. Cell migration assay

HUVECs migration was assessed using Transwell inserts with 8-μm pores (Costar) (Cambridge, MA, USA). Cells (1 × 10⁴) were seeded in the upper chamber with serum-free medium, while the lower chamber contained medium supplemented with 1% FBS. Cells on the upper surface were swabbed off after 12 h of incubation. Cells that migrated to the lower surface were fixed with 4% PFA, stained with crystal violet, and counted in three random fields per membrane.

2.14. Scratch assay analysis

A wound healing assay was used to determine the migratory capacity of HUVECs. Confluent cell monolayers in 6-well plates were scratched with sterile pipette tips to create uniform wounds. The cells were then maintained in serum-free DMEM. Wound closure was evaluated by capturing images of the same fields at 6 and 24 h post-scratching, following an incubation at 37 °C with 5% CO₂, using an Olympus bright-field microscope (Olympus, Tokyo, Japan). The extent of cell migration was quantified by measuring the change in wound width at three predetermined locations per scratch using ImageJ software (NIH ImageJ, Bethesda, MD, USA).

2.15. Network pharmacology analysis

To identify potential pharmacological targets, we used traditional Chinese medicine systems pharmacology database and analysis platform (TCMSP), similarity ensemble approach (SEA), and supervised prediction of drug targets (SuperPRED) databases. Additionally, targets associated with DW were retrieved from online mendelian inheritance in man (OMIM) and GeneCards databases. The associations between DG, ferroptosis, and DW were visualized using Venn diagrams. Subsequently, the overlapping targets identified from these intersections were selected as common targets for DG in the treatment of DW. A protein-protein interaction (PPI) network was constructed using the search tool for the retrieval of interacting genes (STRING) database. Gene Ontology (GO) functional annotation and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were performed using the database for annotation, visualization and integrated discovery (DAVID) database to identify key biological processes and signaling pathways, thereby providing a comprehensive understanding of the relevant functional domains.

2.16. Docking analysis

To prepare for molecular docking, the chemical structure of DG was drawn in ChemBioDraw (PerkinElmer, Waltham, MA, USA) and its energy was minimized using ChemBio3D. The three-dimensional (3D) structures of the receptor proteins Sirt6 and Nrf2 were obtained from the Universal Protein database. The docking conformation with the minimum potential energy was generated using PyMOL's default parameters. Prior to docking, the ligand and receptor protein data bank (PDB) files were converted to protein data bank, partial charge (Q), and atom type format (PDBQT) using AutoDock Tools (v1.5.6). The tools were used to identify active binding pockets, the grid box for conformational sampling was set, and the automated docking script was executed in AutoDock Vina (v1.2.1). The final 3D representations of the docking complexes were rendered for visualization and analysis using PyMOL (University of California, San Francisco, CA, USA).

2.17. Cellular thermal shift assay (CETSA)

HUVECs were exposed to 20 μM DG for 24 h, subsequently washed with cold PBS, and lysed in PMSF buffer containing protease inhibitors. Lysate aliquots were heated across a temperature gradient (37–64 °C, Δ3 °C) for 3 min or subjected to three liquid nitrogen freeze-thaw cycles. Following centrifugation at 20,000 g for 20 min at 4 °C, the soluble fraction was heated in loading buffer at 95 °C for 5 min and subjected to WB analysis.

2.18. Small interfering RNA (siRNA) transfection

To knock down Sirt6 expression, HUVECs were transfected with Sirt6-targeting siRNA (Invitrogen, Carlsbad, CA, USA) at 30%–50% confluence. The transfection procedure exhibited high biocompatibility, with cell viability exceeding 95% after 12 h. After replacing the transfection medium, the cells were cultured for an additional 72 h to allow sufficient protein depletion before use in downstream assays. WB confirmed the efficiency of Sirt6 knockdown.

2.19. Streptozotocin (STZ) -induced diabetes

The Laboratory Animal Ethics Committee of Wenzhou Medical University reviewed and approved all animal experimental procedures. Upon arrival, mice were acclimated for one week in a pathogen-free facility before being randomly divided into a normal control group (n = 8) or a diabetic model group (n = 40). A single intraperitoneal injection of STZ (Sigma Aldrich, St. Louis, MO, USA) at 100 mg/kg was used to induce diabetes. Blood glucose levels were measured before and after induction, and hyperglycemia was defined as a level exceeding 16.7 mmol/L. Only diabetic mice with stable blood glucose levels exceeding 16.7 mmol/L were included in subsequent experiments.

2.20. In vivo wound healing model and drug administration

Forty-eight 8-week-old specific pathogen-free (SPF) C57BL/6 male mice were randomly divided into six groups (n = 8 each): a non-diabetic control group (Control), a DW group (DW), and four diabetic treatment groups receiving either a low dose of DG (DW + DG^L) at 10 mg/kg/day, a high dose of DG (DW + DG^H) at 20 mg/kg/day, Fer-1 (DW + Fer-1) at 5 mg/kg/day, and a high dose of DG with 10 mg/kg/day of OSS_128167 (DW + DG^H + OSS_128167). The doses of DG, OSS, and Fer-1 were selected as previously reported [21,23,24]. Following the induction of anesthesia with intraperitoneal sodium pentobarbital (100 mg/kg) (Sigma Aldrich, St. Louis, MO, USA) and confirming the loss of corneal reflexes, bilateral full-thickness excisional wounds (8 mm diameter) were made along the dorsal midline. DG and OSS_128167 were administered by oral gavage, while Fer-1 was administered through intraperitoneal injection. All drug treatments were initiated 24 h post-operation and continued once daily. Wound healing progression was monitored by photographic documentation on days 0, 5, 10, and 15. Wound areas were quantified using ImageJ (NIH ImageJ, Bethesda, MD, USA), and the data were analyzed statistically.

2.21. Laser Doppler perfusion imaging

To evaluate wound perfusion, laser Doppler imaging (Millwey, Axminster, UK) was performed on anesthetized mice receiving 2% pentobarbital sodium. Measurements were taken non-invasively on postoperative day 10. Following dorsal hair removal, blood flow images of the wound areas were acquired. Quantitative analysis was performed by measuring perfusion units (PU) in three predefined regions of interest (ROIs) at each wound site.

2.22. Hematoxylin-Eosin (H&E) staining and Masson's trichrome staining

Skin tissue samples were fixed in 4% PFA overnight, then embedded in paraffin and sectioned into 4-μm slices. Sections were stained with H&E (Beyotime, Shanghai, China) for general morphology and Masson's trichrome (Beyotime, Shanghai, China) to determine collagen deposition. Stained sections were then observed and imaged under a light microscope (Olympus, Tokyo, Japan).

2.23. Immunohistochemistry (IHC)

Paraffin-embedded sections underwent antigen retrieval in heated citrate buffer (95 °C, 20 min) for IHC staining, followed by a 10-min treatment with 3% H₂O₂ to inhibit endogenous peroxidase activity. Sections were blocked with 5% bovine serum albumin (BSA) (Beyotime, Shanghai, China) for 30 min and then incubated with specific primary antibodies overnight at 4 °C. The following day, the sections were probed with HRP-conjugated secondary antibodies for 1 h at room temperature. Color development was performed using diaminobenzidine (DAB) (Beyotime, Shanghai, China) for 5 min, followed by hematoxylin counterstaining. The stained sections were dehydrated, mounted, and observed using a Carl Zeiss Axio Imager bright-field microscope (Zeiss, Oberkochen, Germany).

2.24. Statistical analysis

Continuous data are presented as mean ± standard error of the mean. All data were confirmed to be normally distributed. Each experiment was performed at least three times. Statistical analyses were performed using the IBM Statistical Package for the Social Sciences software (Armonk, NY, USA). Group differences were determined using a one-way analysis of variance followed by Tukey's test; a P < 0.05 was considered statistically significant.

3. Results

3.1. A diabetic microenvironment induces HUVECs ferroptosis in vitro and in vivo

In diabetic mice, elevated levels of free fatty acids (FFAs), particularly palmitic acid (PA), were observed. As the most abundant saturated fatty acid, PA has been demonstrated to induce damage across various cell types. Furthermore, as reported by Jin et al. [25], 5 mM glucose can serve as an exogenous stimulus to mimic a hyperglycemic microenvironment in cells. The optimal concentration of PA was determined through the CCK-8 assay, and we decided to use high glucose (25.5 mM) combined with PA (200 μM) to simulate HGHF conditions in vitro (Fig. 1A). For the subsequent HGHF group, we co-treatment with dimethyl sulfoxide (DMSO), the volume of DMSO administered was equal to that used in DG-treated groups. Fer-1 has been demonstrated to restore the viability of bone marrow mesenchymal stem cells (BMSCs) and osteoblasts impaired by elevated glucose levels [26,27]. Accordingly, our CCK-8 assay indicated that 8 μM Fer-1 significantly restored the viability of HUVECs under HGHF stimulation (Fig. 1B). To determine whether the decrease in cell viability caused by HGHF was associated with ferroptosis, cells were treated with several inhibitors: the autophagy inhibitor 3-methyladenine (3-MA), the necroptosis inhibitor Necrostatin-1 (Nec-1), the apoptosis inhibitor Z-Val-Ala-Asp(OMe)-fluoromethylketone (Z-VAD-FMK), the ROS scavenger N-acetylcysteine (N-Ace), and the ferroptosis inhibitor Fer-1. The findings indicated that N-Ace and Fer-1 were the only inhibitors to significantly inhibit HGHF-induced death in HUVECs, while the others had minimal impact (Fig. 1C). Next, EdU staining indicated that HGHF and erastin inhibited cell proliferation; however, it was significantly alleviated by Fer-1 in the presence of HGHF (Figs. 1D and E). Considering that iron excess is a hallmark of ferroptosis, we determined the expression of proteins involved in iron metabolism and ferroptosis defense. Consistent with this hallmark, HGHF and erastin treatment significantly downregulated the key anti-ferroptosis proteins GPX4 and SLC7A11, and, conversely, upregulated ACSL4 (a mediator of ferroptosis). Additionally, the ferroptosis inhibitor Fer-1 successfully reversed these changes (Figs. 1F and G). Intracellular lipid peroxides were detected using C11-BODIPY. Consistent with this, staining with the lipid peroxidation probe C11-BODIPY revealed that Fer-1 significantly reduced HGHF-induced lipid peroxidation (Figs. 1H and I). Next, we investigated whether ferroptosis occurs in DW in vivo. WB and fluorescent staining of wound tissues exhibited a significant downregulation of SLC7A11 and GPX4 and an upregulation of ACLS4 in diabetic mice (Figs. 1J and K). These findings were corroborated by reduced GPX4 fluorescence (Figs. 1L and M). Consistent with the in vitro findings in HUVECs, these results collectively indicate that the hyperglycemic environment promotes ferroptosis in DW tissues.

Fig. 1.

High-glucose and high-fat (HGHF) induces ferroptosis in human umbilical vein endothelial cells (HUVECs). HUVECs were treated with erastin, HGHF and ferrostatin-1 (Fer-1) for 24 h. (A) Cell viability of HUVECs exposed to different concentrations of palmitic acid (PA) for 24 h with or without HGHF (25.5 mM). (B) Cell viability of HUVECs exposed to different concentrations of Fer-1 for 24 h with or without HGHF. (C) Cell Counting Kit-8 (CCK-8) assay results of HUVECs pretreated with phosphate-buffered saline (PBS), 3-methyladenine (3-MA), Necrostatin-1 (Nec-1), Z-Val-Ala-Asp(OMe)-fluoromethylketone (Z-VAD-FMK), N-acetylcysteine (N-Ace), or Fer-1 followed by HGHF for 24 h. (D, E) Representative fluorescent images (D) and positive quantitative analysis (E) of HUVECs labeled with 5-ethynyl-20-deoxyuridine (EdU) (green); nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI) (blue). (F, G) Western blot (WB) analysis of solute carrier family 7 member 11 (SLC7A11), acyl-Coenzyme a synthetase long-chain family member 4 (ACSL4), and glutathione peroxidase 4 (GPX4) protein expression in differentially treated HUVECs: representative WB image (F) and quantitative analysis (G). (H, I) Representative fluorescent images (H) and quantitative analysis (I) of C11-BODIPY staining in differentially treated HUVECs. (J, K) WB analysis of SLC7A11, ACSL4, and GPX4 protein expression in wound tissues from the Control and diabetic wounds (DW) groups: representative WB image (J) and quantitative analysis (K). (L, M) Representative immunofluorescence images (L) and quantitative analysis (M) of GPX4 in wound tissues. Data are presented as the mean ± standard deviation (SD) from 5 independent experiments (n = 5). Statistical significance is denoted as ^∗P < 0.05, ^∗∗P < 0.01, ^∗∗∗P < 0.001; ns indicates no significant difference.

3.2. DG protects HUVECs against HGHF-induced ferroptosis

The chemical structure of DG is presented in Fig. 2A. The optimal non-toxic concentration was initially identified using the CCK-8 assay. HUVECs were treated with various concentrations of DG (0−80 μM) for 24 or 48 h. The results demonstrated no cytotoxicity at these concentrations. Based on these findings and subsequent experimental outcomes, 20 and 10 μM were selected as the optimal (DG^H) and low-dose (DG^L) concentrations of DG, respectively (Figs. 2B and C). Furthermore, the CCK-8 assay demonstrated that DG dose-dependently and significantly mitigated the reduction in cell viability induced by HGHF and erastin (Figs. 2D and E). WB analysis indicated that the HGHF group exhibited downregulated GPX4 and SLC7A11 expression, with upregulated ACSL4 expression (Figs. 2F and G). Ferroptosis is characterized by smaller mitochondria and denser membranes. TEM revealed mitochondria with shrinkage and torn membranes in HGHF-treated HUVECs, whereas DG treatment partially restored mitochondrial structure (Fig. 2H). Ferroptosis is commonly associated with various morphological and metabolic changes, including lipid peroxidation and iron overload. To confirm ferroptosis in HGHF-treated HUVECs, we performed C11-BODIPY and FerroOrange staining with and without DG/Fer-1 rescue. C11-BODIPY staining indicated significant lipid peroxidation stimulated by HGHF, which was mitigated by DG and Fer-1 treatment (Figs. 2I and J). FerroOrange is a Fe²⁺-specific probe. The fluorescence intensity significantly increased following HGHF stimulation; however, this change was reversed by DG and Fer-1 treatment (Figs. 2K and L). These changes were significantly mitigated by either DG or Fer-1 treatment. The data collectively suggest that DG can inhibit ferroptosis induced by HGHF.

Fig. 2.

Role of diosgenin (DG) in counteracting high-glucose and high-fat (HGHF)-induced ferroptosis in human umbilical vein endothelial cells (HUVECs). To validate the role of DG in ferroptosis, five groups were set up: control, HGHF, HGHF + low dose of DG (10 μM) (DG^L), HGHF + high dose of DG (20 μM) (DG^H), and HGHF + ferrostatin-1 (Fer-1). (A) Chemical structure of DG. (B, C) Cell viability percentages after exposure to different concentrations of DG for 24 h (B) and 48 h (C). (D, E) Cell viability percentages after pretreatment with HGHF (D) and erastin (E), combined with or without different doses of DG. (F, G) Western blot (WB) analysis of solute carrier family 7 member 11 (SLC7A11), acyl-Coenzyme a synthetase long-chain family member 4 (ACSL4), and glutathione peroxidase 4 (GPX4) protein expression in differentially treated HUVECs: representative WB image (F) and quantitative analysis (G). (H) Representative images of mitochondrial morphology in differentially treated HUVECs. (I, J) Representative fluorescent images (I) and quantitative analysis (J) of C11-BODIPY staining in differentially treated HUVECs. (K, L) Representative fluorescent images (K) and quantitative analysis (L) of FerroOrange staining in differentially treated HUVECs. Data are presented as the mean ± standard deviation (SD) from 5 independent experiments (n = 5). Statistical significance is denoted as ^∗P < 0.05, ^∗∗P < 0.01, ^∗∗∗P < 0.001; ns indicates no significant difference.

3.3. DG rescues HGHF-induced mitochondrial defects in HUVECs

Given that ferroptosis is characterized by elevated ROS levels and mitochondrial oxidative stress, resulting in mitochondrial dysfunction, we next investigated the effect of DG on mitochondrial function under HGHF induction. Intracellular and mitoROS levels were determined using DCFH-DA and MitoSox probes, respectively. Fluorescence imaging indicated that DG significantly reduced ROS levels in both compartments (Fig. 3A−D). Next, we investigated the effect of DG on the MMP. HGHF treatment significantly reduced the MMP, a trend that was substantially reversed by the administration of either DG or Fer-1 (Figs. 3E and F). Mitochondrial fusion and fission are fundamental processes that directly govern mitochondrial function, including energy production, calcium buffering, and apoptosis regulation. A delicate balance between these opposing processes is crucial for maintaining mitochondrial health and cellular homeostasis. Mitochondrial morphology in the HGHF group appeared discontinuous and fragmented compared with the control group. To determine mitochondrial integrity, we performed MitoTracker Green staining, which reflects mitochondrial mass and membrane potential. As depicted in Figs. 3E and F, mitochondrial morphology and fluorescence intensity were significantly improved in DG and Fer-1 treatment groups compared with the control, indicating a significant alleviation of mitochondrial damage (Figs. 3G and H). As a key endogenous antioxidant, GSH activity was significantly reduced by HGHF exposure, accompanied by an increase in MDA, a marker of lipid peroxidation. This HGHF-induced oxidative imbalance was effectively prevented by DG and Fer-1 postconditioning (Figs. 3I and J). Mitochondrial dynamics are regulated by key proteins: OPA1, which mediates inner membrane fusion; MFN1 and MFN2, the essential MFNs for outer membrane fusion; and DRP1, which acts as the primary executor of mitochondrial fission [28]. WB analysis indicated that DG upregulated the expression of OPA1, MFN1, and MFN2, while downregulating DRP1 expression (Figs. 3K and L). These results suggest that DG can rescue the mitochondrial dysfunction induced by HGHF.

Fig. 3.

Role of diosgenin (DG) in mitochondrial preservation and functionality. Human umbilical vein endothelial cells (HUVECs) were treated with high-glucose and high-fat (HGHF) alone, HGHF combined with DG (low or high dose), or HGHF with ferrostatin-1 (Fer-1) for 24 h. (A, B) Representative images (A) and corresponding quantitative (B) of reactive oxygen species (ROS) staining. (C, D) Representative fluorescent images (C) and quantitative analysis (D) of MitoSox (red) staining in differentially treated HUVECs. (E, F) Detection of JC-1 staining (E) and quantitative analysis (F) in HUVECs by confocal fluorescence microscopy. (G, H) Representative fluorescent images (G) and quantitative analysis (H) of MitoTracker staining in differentially treated HUVECs. (I, J) Quantitative determination of glutathione (GSH) (I) and malondialdehyde (MDA) (J) activities in HUVECs using GSH and MDA assay kits. (K, L) Western blot (WB) analysis of optic atrophy protein 1 (OPA1), dynamin-related protein 1 (DRP1), mitofusin 1 (MFN1), and mitofusin 2 (MFN2) protein expression in differentially treated HUVECs: representative WB image (K) and quantitative analysis (L). Data are presented as the mean ± standard deviation (SD) from 5 independent experiments (n = 5). Statistical significance is denoted as ^∗∗∗P < 0.001; MMP: Mitochondrial Membrane Potential.

3.4. DG promoted cell function in HGHF-treated HUVECs

To investigate the therapeutic potential of DG on HUVECs' function, we performed a series of experiments under HGHF conditions. HGHF treatment significantly suppressed HUVECs migration, whereas supplementation with DG and Fer-1 substantially enhanced this process (Figs. 4A−D). Tube formation assays revealed that DG and Fer-1 restored the tube-forming ability of HUVECs impaired by HGHF (Figs. 4E and F). Additionally, considering the critical role of cell adhesion in migration, we next evaluated adhesive capacity. Cell adhesion staining demonstrated that both DG and Fer-1 promoted cell adhesion (Figs. 4G and H). Then, we identified key molecular markers of angiogenesis: CD31 and α-SMA, widely used in vascular and smooth muscle studies, and VEGF-A, a critical regulator of tube formation. WB analysis revealed that HGHF downregulated protein expression, whereas DG and Fer-1 treatments significantly upregulated their levels (Figs. 4I and J). In conclusion, DG significantly mitigates the functional impairments in HUVECs induced by HGHF exposure.

Fig. 4.

Effect of diosgenin (DG) on the function of high-glucose and high-fat (HGHF)-treated human umbilical vein endothelial cells (HUVECs). Five groups in this part to investigate the effect of DC on HUVECs: control, HGHF, HGHF + low dose of DG (10 μM) (DG^L), HGHF + high dose of DG (20 μM) (DG^H), and HGHF + ferrostatin-1 (Fer-1). (A, B) Evaluation of DG-mediated HUVECs migration (A) and quantitative analysis (B) using the transwell migration assay. (C, D) The scratches assay (C) and quantitative analysis (D) of DG-mediated HUVECs. (E, F) Evaluation of DG-mediated HUVECs angiogenesis (E) and quantitative analysis (F) using the tube formation assay. (G, H) Evaluation of the effect of DG on HUVECs adhesion ability (G) and quantitative analysis (H) using the cell-matrix adhesion assay. (I, J) Western blot (WB) (I) and quantitative analysis (J) of cluster of differentiation 31 (CD31), α-smooth muscle actin (α-SMA), and vascular endothelial growth factor A (VEGF-A) protein expression in HUVECs. Data are presented as the mean ± standard deviation (SD) from 5 independent experiments (n = 5). Statistical significance is denoted as ^∗∗∗P < 0.001.

3.5. Activation of the Sirt6/Nrf2 signaling pathway mediates the positive effects of DG

We explored DG's potential to modulate ferroptosis in DW using network pharmacology analysis. The chemical structure of DG was obtained from relevant databases, identifying 101 targets. We identified 6241 targets associated with DW disease and 551 targets with ferroptosis. We identified 14 potential regulators of ferroptosis in DW by intersecting these gene sets, including Sirt6 and MTOR (Figs. 5A and B). The PPI networks of these potential targets and the subnetwork of key targets in Cytoscape (Fig. 5B). KEGG enrichment analysis identified significantly enriched signaling pathways, including ferroptosis, type II diabetes mellitus, and fatty acid metabolism (Fig. 5C). GO enrichment analysis revealed that these targets participate in diverse biological processes associated with DW, including wound healing, ECM and mitochondrion (Fig. 5D). Furthermore, we performed molecular docking to predict the relationship between DG and the Sirt6/Nrf2 signaling pathway. Based on the comprehensive analysis of all possible models, DG associates with and engages the Sirt6/Nrf2 docking site. The ribbon model was modified to demonstrate both macro- and local-level viewpoints of these interactions. A space-filling model was used to visually illustrate this interaction. The hydrogen-binding energies between DG and Sirt6 or Nrf2 are −7.6 and −9.2 kcal/mol, respectively (Fig. 5E and Table S1). CETSA is another experimental method commonly used to identify interactions between drugs and their cellular protein targets. The study found that interaction with DG upregulated Sirt6 and Nrf2 protein expression within the 37−64 °C range, with the gray value at 37 °C used as a reference. The increased stability made the proteins less susceptible to proteolytic degradation (Figs. 5F−H). Consequently, this suggests that DG effectively modulates the activation of Sirt6 and Nrf2 signaling pathways.

Fig. 5.

Network pharmacology analysis and verification. Analyzing existing databases to investigate the impact of diosgenin (DG) on the sirtuin 6/nuclear factor erythroid 2-related factor 2 (Sirt6/Nrf2) signaling pathway. (A) Venn diagram of DG active ingredients and targets related to diabetic wounds (DWs), DG, and ferroptosis. (B) Primary clusters extracted from the protein-protein interaction (PPI) network illustrate the main functional gene clusters of potential targets. (C) Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis of core targets of DG against DWs. (D) Gene Ontology (GO) functional enrichment analysis of core targets of DG against DWs. (E) A ribbon model representing protein residues and illustrating the three-dimensional (3D) binding mode. DG docking with Sirt6 and Nrf2 showed binding affinities of −7.6 kcal/mol and −9.2 kcal/mol, respectively. A space-filling model shows DG embedded in the Sirt6 and Nrf2 binding pockets. (F−H) Human umbilical vein endothelial cells (HUVECs) were incubated with phosphate-buffered saline (PBS) or DG (20 μM) for 24 h, and cellular thermal shift assay (CETSA) (F) and quantitative analysis was performed to detect the thermal stability of Sirt6 (G) and Nrf2 (H) proteins. Data are presented as the mean ± standard deviation (SD) from 5 independent experiments (n = 5). Statistical significance is denoted as ^∗P < 0.05, ^∗∗P < 0.01, ^∗∗∗P < 0.001; ns indicates no significant difference. MAPK: mitogen-activated protein kinase; BP: biological process; CC: cellular component; MF: molecular function.

3.6. The Sirt6/Nrf2/Keap1 pathway is significantly modulated by DG treatment

The functional activity of Sirt6 is typically mediated by its phosphorylation. Under normal conditions, Keap1 associates with Nrf2, resulting in its ubiquitination and proteasomal degradation in the cytoplasm. Therefore, attenuating the Keap1-Nrf2 interaction presents a potential strategy to improve Nrf2 stability and increase its nuclear accumulation in HUVECs. Once stabilized, Nrf2 dimerizes and binds to the antioxidant response element (ARE) in DNA, thereby initiating transcription of target genes. WB analysis was performed to experimentally validate this mechanism. The inhibitory effects of HGHF on Sirt6 phosphorylation and Nrf2 nuclear translocation were effectively reversed by DG or Fer-1 treatment (Figs. 6A−E). Furthermore, we confirmed that Keap1 inhibits Nrf2 signaling by promoting its nuclear export. Using cycloheximide to inhibit protein synthesis in HUVECs, we found that DG significantly delayed Nrf2 degradation compared with the control (Fig. 6F). Co-immunoprecipitation assays further indicated that DG treatment reduced the levels of Keap1 bound to Nrf2. This effect was consistent with observations using the proteasome inhibitor MG132 (Fig. 6G). Finally, cellular immunofluorescence staining confirmed that DG promotes Nrf2 nuclear translocation (Figs. 6H and I). Furthermore, this study confirmed that DG significantly upregulates the protein expression of the Nrf2-regulated anti-ferroptosis factors GPX4 and SLC7A11, while downregulating Keap1 expression. These results indicate that DG promotes the Keap1-dissociated (stabilized) form of Nrf2, facilitating its transcriptional activity. To demonstrate the importance of the Sirt6/Nrf2/Keap1 pathway in HGHF-induced ferroptosis, Sirt6 expression was subsequently inhibited using Sirt6-siRNA. WB analysis revealed significantly downregulated Sirt6 expression (si-Sirt6 #2) subsequent to the transfection procedure (Figs. 7A and B). In HUVECs subjected to si-Sirt6 #2-mediated Sirt6 knockdown and co-treated with DG and HGHF, functional assays exhibited a significant reduction in both migratory and tube-forming abilities (Figs. 7C−H). Therefore, the expression of key angiogenic proteins (CD31, α-SMA, and VEGF) was downregulated. WB analysis further confirmed that the activatory effects of DG on the Sirt6/Nrf2 signaling pathway and its downstream anti-ferroptosis proteins were significantly suppressed by si-Sirt6 #2 (Figs. 7I−J). In summary, our data demonstrate that DG promotes the dissociation of Nrf2 from Keap1 by promoting Sirt6 phosphorylation. This causes Nrf2 to translocate into the nucleus, resulting in increased expression of the anti-ferroptosis proteins GPX4 and SLC7A11 (Figs. 7K−O).

Fig. 6.

Beneficial effects of diosgenin (DG) mediated by the sirtuin 6/nuclear factor erythroid 2-related factor 2 (Sirt6/Nrf2) pathway. After 24 h of drug treatment, proteins were extracted for detection. (A) Western blot (WB) of phosphorylated-Sirt6 (p-Sirt6), kelch-like ECH-associated protein 1 (Keap1), solute carrier family 7 member 11 (SLC7A11), and glutathione peroxidase 4 (GPX4). (B) Quantitative analysis of Keap1, SLC7A11, and GPX4. (C) Quantitative analysis of p-Sirt6. (D) WB of nuclear-Nrf2 (n-Nrf2). (E) Quantitative analysis of n-Nrf2. (F) Cycloheximide (CHX) chase assay to assess Nrf2 protein stability: HUVECs were incubated with 30 μM cycloheximide with or without DG for the indicated times. (G) Co-immunoprecipitation and WB to detect the interaction between Nrf2 and Keap1: HUVECs were treated with Z-Leu-Leu-Leu-al (MG-132) (5 μM) with or without DG, followed by cell lysis. (H, I) Nrf2 immunofluorescence images (H) and quantitative analysis (I) in differentially treated HUVECs. Data are presented as the mean ± standard deviation (SD) from 5 independent experiments (n = 5). Statistical significance is denoted as ^∗∗∗P < 0.001.

Fig. 7.

Inhibition of ferroptosis and improvement of human umbilical vein endothelial cells (HUVECs) function by diosgenin (DG) are regulated by small interfering RNA of sirtuin 6 (Sirt6-siRNA). After Sirt6-siRNA used, angiogenesis-related proteins were detected. (A, B) Western blot (WB) (A) and quantitative analysis (B) of sirtuin 6 (Sirt6) protein expression in differentially treated HUVECs. (C, D) Evaluation of HUVECs migration (C) and quantitative analysis (D) using the transwell migration assay. (E, F) The scratches assay (E) and quantitative analysis (F) of HUVECs. (G, H) Evaluation of HUVECs angiogenesis (G) and quantitative analysis (H) using the tube formation assay. (I, J) WB (I) and quantitative analysis (J) of cluster of differentiation 31 (CD31), α-smooth muscle actin (α-SMA), and vascular endothelial growth factor A (VEGF-A) protein expression in differentially treated HUVECs. (K, L) WB (K) and quantitative analysis (L) of nuclear-Nrf2 (n-Nrf2) expression in differentially treated HUVECs. (M−O) WB (M) and quantitative analysis of phosphorylated-Sirt6 (p-Sirt6) (N), kelch-like ECH-associated protein 1 (Keap1), solute carrier family 7 member 11 (SLC7A11), and glutathione peroxidase 4 (GPX4) protein expression (O) in differentially treated HUVECs. Data are presented as the mean ± standard deviation (SD) from 5 independent experiments (n = 5). Statistical significance is denoted as ^∗∗∗P < 0.001; ns indicates no significant difference. Nc: negative control.

3.7. DG promotes wound healing process in DW mice

We determined DG's efficacy in DW healing by establishing a diabetic mouse model through intraperitoneal STZ injection followed by surgical induction of cutaneous wounds (Fig. 8A). To investigate the underlying mechanism, treatment groups receiving a ferroptosis inhibitor or the Sirt6 inhibitor OSS_128167 (OSS) (a recognized and effective inhibitor of Sirt6) were included in the study [29]. The overall experimental timeline, including model establishment, treatment protocols, and group allocation, is outlined in Fig. 8A. The wound-healing rate in the diabetic group was slow by gross appearance, whereas DG treatment significantly accelerated DW healing in a dose-dependent manner (Figs. 8B and C). Additionally, Fer-1 was observed to promote DW healing (Figs. 8B and C). However, Sirt6 inhibitor administration suppressed the therapeutic effect of DG, indicating that, compared with the DG treatment group, wound healing in the OSS group was poorer (Figs. 8B and C). Furthermore, quantitative analysis of the wound area supports this conclusion (Fig. 8D). On day 10 of wound healing, tissue samples were collected and stained with H&E and Masson's trichrome. The results revealed that the diabetic group exhibited no significant evidence of epithelialization, in contrast to the control group. DW treated with DG and Fer-1 demonstrated enhanced neo-epidermis formation. Additionally, epithelialization was weaker in the OSS group than in the DG treatment group (Figs. 8E−G).

Fig. 8.

Diosgenin (DG) promotes diabetic wound (DW) healing in vivo. To validate the in vivo effect of DG, mice were randomly divided into six groups: a non-diabetic control group (Control), a DW group (DW), and four diabetic treatment groups receiving either a low dose of DG (DW + DG^L), a high dose of DG (DW + DG^H), Fer-1 (DW + Fer-1), and a high dose of DG with OSS_128167 (DW + DG^H + OSS_128167). group (DW), and four diabetic treatment groups receiving either a low dose of DG (DW + DG^L), a high dose of DG (DW + DG^H), Fer-1 (DW + Fer-1), and a high dose of DG with OSS_128167 (DW + DG^H + OSS_128167). (A) Overview of the animal experiment. (B, C) Wound images during the healing process (B) and a schematic diagram of the wound healing process (C). (D) Quantitative data of wound area relative to day 0 for the 6 groups. (E−G) hematoxylin-eosin (H&E) (E) and quantitative analysis (F), Masson's trichrome staining images (G) of wound length. Results are presented as the mean ± standard deviation (SD) from quintuple tests, n ≥ 5. Significance is denoted as ^∗∗∗P < 0.001; ns indicates no significant difference. STZ: streptozocin.

3.8. DG promotes DW healing by enhancing angiogenesis and inhibiting ferroptosis

To determine whether DG significantly affects the microvascular network in the dorsal skin of mice, we performed doppler flowmetry and wound-tissue fluorescence experiments. Laser Doppler imaging was used to visualize changes in blood flow at the wound site and adjacent areas during healing, with flux expressed in PUs rather than blood flow velocity. On days 5 and 10, the blood flow in both the wound margins and wound beds of the groups treated with DG and Fer-1 was significantly higher than that in the DW group (Fig. 9A). The OSS group exhibited a significantly lower blood flow than the DG group (Figs. 9A and B). Histological analysis using immunofluorescence staining for α-SMA and CD31 demonstrated the development of new and mature blood vessels in DW. On day 10, the DG and Fer-1 treatment groups exhibited a significant increase in both newly formed and mature blood vessels compared to the DW group (Figs. 9C, 9D, and S1). Furthermore, we investigated in vivo indicators of ferroptosis. Tissue fluorescence exhibited that DG significantly downregulated the expression of the ferroptosis-associated protein ACSL4 and the lipid peroxidation marker 4-hydroxynonenal (4-HNE), and significantly decreased Fe²⁺ levels in vivo (Figs. 9E−J). Histochemical results indicated that DG promoted Sirt6 expression. Interestingly, the positive therapeutic effects of DG were reversed by OSS (Figs. 9K and L). In summary, alterations in ferroptosis-related protein expression were observed in wound tissue samples. WB analysis demonstrated that DG treatment upregulated anti-ferroptosis protein expression while downregulating ACSL4 expression (Figs. 9M and N). Additionally, we collected the mice's major organs at the end of the experiment for H&E staining. The results revealed no significant abnormalities in tissue morphology or pathological damage in the heart, liver, spleen, lungs, or kidneys of the DG-treated group compared with the control group, suggesting a favorable preliminary safety profile for oral administration of DG (Fig. S2). In conclusion, our in vivo experiments indicated that DG promotes DW healing by inhibiting ferroptosis and promoting angiogenesis through the Sirt6 signaling pathway.

Fig. 9.

Impact of diosgenin (DG) application on diabetic wound (DW) healing and reactive oxygen species (ROS) diminishment in vivo. To evaluate the in vivo effects of DG on tube formation and ROS reduction, six groups were seted up: a non-diabetic control group (Control), a DW group (DW), and four diabetic treatment groups receiving either a low dose of DG (DW + DG^L), a high dose of DG (DW + DG^H), Fer-1 (DW + Fer-1), and a high dose of DG with OSS_128167 (DW + DG^H + OSS_128167). (A, B) Laser Doppler perfusion imaging showing angiogenesis recovery (A) and quantitative analysis (B). (C, D) Representative immunofluorescence images (C) and quantitative analysis (D) of α-smooth muscle actin (α-SMA) in wound from each group. (E, F) Representative immunofluorescence images (E) and quantitative analysis (F) of acyl-Coenzyme a synthetase long-chain family member 4 (ACSL4), in wound from each group. (G, H) Representative fluorescent images (G) and quantitative analysis (H) of 4-hydroxynonenal (4-HNE) in wound from each group. (I, J) Representative fluorescent images (I) and quantitative analysis (J) of Fe²⁺ in wound from each group. (K, L) Representative immunohistochemical images (K) and quantitative analysis (L) of Sirt6 in wound from each group. (M, N) Western blot (WB) (M) and quantitative analysis (N) of solute carrier family 7 member 11 (SLC7A11), ACSL4, and glutathione peroxidase 4 (GPX4) protein expression in wound tissues from the control groups, DW groups and DW + DG^H groups. Results are presented as the mean ± standard deviation (SD) from quintuple tests, n ≥ 5. Significance is denoted as ^∗P < 0.05, ^∗∗∗P < 0.001; ns indicates no significant difference. Fer-1: ferrostatin-1; OSS_128167: selective Sirt6 inhibitor.

4. Disscussion

This study investigated the impact of ferroptosis on DW healing using combined in vitro and in vivo methods. We found that HGHF conditions impaired the survival and migration of HUVECs, while concurrently elevating ROS, lipid peroxidation, and the expression of ferroptosis-related proteins. Treatment with DG or the ferroptosis inhibitor Fer-1 effectively reversed these adverse effects. Consistent with these cellular findings, both DG and Fer-1 significantly accelerated wound closure in STZ-induced diabetic mice. The therapeutic benefits were associated with suppression of ROS and ferroptosis, along with activation of the Sirt6/Nrf2 antioxidant pathway. Together, these results suggest that DG promotes DW healing by reducing ferroptosis and enhancing angiogenesis, likely through the Sirt6/Nrf2 signaling pathway.

In DW, persistent hyperglycemia promotes a chronic inflammatory microenvironment and excessive ROS accumulation, collectively exacerbating oxidative stress and significantly impeding the healing process [30]. Jin et al. [25] and Li et al. [31] reported that high-glucose-induced ROS can induce mitochondrial dysfunction, establishing a vicious cycle of ROS generation that ultimately promotes mitochondrial-dependent cell death. Therefore, mitigating ROS levels and preserving mitochondrial integrity under hyperglycemic conditions are crucial for facilitating wound repair [32]. Iron dyshomeostasis has been extensively implicated in multiple chronic diseases. Studies suggest that iron overload is associated with impaired wound healing [33]. Notably, Khanna and co-workers [34] emphasized that ferroptosis, an iron-dependent form of regulated cell death, plays a pivotal role in initiating and sustaining inflammation in DW. Supporting this, Wang et al. [35] highlighted in Diabetes Care that oxidative stress and lipid peroxidation are fundamental pathological mechanisms underlying both ferroptosis and delayed healing in diabetic ulcers. Furthermore, previous studies have demonstrated that the ferroptosis inhibitor desferrioxamine (DFO) effectively promotes DW healing [36]. Cytologically, ferroptosis is characterized by distinctive mitochondrial alterations, including condensed mitochondrial membranes, reduced cristae, and rupture of the outer mitochondrial membrane, distinguishing it from other forms of mitochondrial dysfunction [37]. This process originates from impaired mitochondrial bioenergetics and GSH depletion, which ultimately result in iron-dependent lipid peroxidation and eventual cell death [38]. Consequently, strategies aimed at restoring mitochondrial homeostasis represent a promising therapeutic avenue to inhibit ferroptosis and promote wound recovery in diabetic settings.

Sirt6, part of the sirtuin family, exhibits histone deacetylase and adenosine diphosphate (ADP)-ribosyltransferase functions, playing a role in various biological processes, including epigenetic regulation and inflammation [39]. Furthermore, recent studies have revealed its role in glucose and lipid metabolism. Nrf2, a key transcription factor, upregulates the expression of several downstream antioxidant enzymes, including GPX4 and SLC7A11 [40,41]. Under normal physiological conditions, Keap1 binds to Nrf2, promoting its ubiquitination and subsequent proteasomal degradation. Under oxidative stress, Nrf2 dissociates from Keap1, translocates to the nucleus, and activates ARE-driven genes, thereby increasing cellular resistance to peroxidative damage [42].

DG, a saponin present in fenugreek and wild yam, is known to mitigate type 2 diabetes complications and inhibit ferroptosis by activating Sirt6 and Nrf2 pathways [21,43]. We hypothesized that DG could improve mitochondrial function and reduce HGHF-induced ferroptosis in DW by competitively binding to Keap1, thereby disrupting the Keap1-Nrf2 interaction and activating the Sirt6/Nrf2 axis. Our experimental data support this mechanism: DG promoted Sirt6 phosphorylation, which facilitated Nrf2 dissociation from Keap1, resulting in Nrf2 nuclear translocation and upregulation of key anti-ferroptosis proteins, including GPX4 and SLC7A11. To further investigate the role of Sirt6 in HGHF-injured HUVECs, we used Sirt6-specific siRNA and found that Sirt6 knockdown upregulated Keap1 expression, likely due to impaired Nrf2-Keap1 dissociation. Consequently, inhibition of this pathway downregulated GPX4 and SLC7A11 levels and reduced the beneficial effects of DG on tubulogenesis under HGHF conditions. Furthermore, in vivo evaluations confirmed that DG treatment upregulated GPX4 expression in DW, reduced ROS levels, and improved wound healing. These intricate mechanisms are summarized in Fig. 10.

Fig. 10.

Mechanism of diosgenin (DG) action on high-glucose and high-fat (HGHF)-induced human umbilical vein endothelial cells (HUVECs): DG activates the sirtuin 6/nuclear factor erythroid 2-related factor 2 (Sirt6/Nrf2) signaling pathway, reduces reactive oxygen species (ROS) and lipid peroxidation accumulation, and ameliorates mitochondrial dysfunction, thereby inhibiting HGHF-induced ferroptosis in HUVECs and ultimately promoting diabetic wound (DW) healing. Sirt6: sirtuin 6; GSH: glutathione; MDA: malondialdehyde; SLC7A11: solute carrier family 7 member 11; GPX4: glutathione peroxidase 4; CD31: cluster of differentiation 31; VEGF: vascular endothelial growth factor; α-SMA: α-smooth muscle actin.

To situate our findings within the broader context of innovative DW therapies, a comparison with strategies, including adipose-derived stem cell exosomes (ADSC-exo) and engineered hydrogels (4-octyl itaconate-loaded exosome-gel, 4OI-Exo-Gel), is required [44,45]. ADSCs-exo function as multi-factorial “biological” cocktails, and advanced gels provide sophisticated controlled release, whereas DG represents a distinct, small-molecule-based approach. Its defined chemical structure and precise mechanism of action through the Sirt6/Nrf2 pathway provide advantages in pharmacological tractability, manufacturing scalability, and formulation simplicity. This positions DG not as a substitute but as a complementary and potentially more easily translatable approach. Future studies could investigate synergies by integrating DG with such bioactive delivery systems to optimize therapeutic outcomes. Our study presents compelling findings; however, it has certain limitations. First, the current study was conducted exclusively in cellular and murine models. While our findings establish a strong mechanistic foundation in a mouse model, future studies using large-animal models are required to further determine the translational potential of DG for DW therapy. Second, although our data strongly suggest the involvement of Sirt6, the precise role of this sirtuin would be more conclusively demonstrated using endothelial-specific Sirt6 knockout or transgenic mouse models. Future studies incorporating these advanced genetic tools, alongside pharmacokinetic and toxicological assessments, are required to translate these promising findings into clinically applicable therapies for DW healing. Our findings indicate that Sirt6 is essential for DG-mediated Nrf2 activation and ferroptosis suppression; the exact molecular mechanism, whether Sirt6 directly deacetylates Keap1/Nrf2 or acts through intermediate redox sensors, remains to be determined. Future studies should investigate potential direct post-translational modifications of Keap1/Nrf2 by Sirt6.

5. Conclusion

This research suggests that DG significantly promotes DW healing, as evidenced by both in vitro and in vivo experiments. The underlying mechanism involves DG activating the Sirt6/Nrf2 signaling pathway, thereby effectively ameliorating mitochondrial dysfunction and suppressing ferroptosis in HUVECs under HGHF conditions. These findings indicate the therapeutic potential of DG and elucidate its pharmacological basis, making it a promising candidate for treating DW.

CRediT authorship contribution statement

Ji-qi Wang: Writing – original draft, Methodology. Lu Chen: Methodology, Investigation, Formal analysis. Yu-zhe Lin: Methodology, Investigation, Formal analysis. Xiu-zhi Zhang: Methodology, Investigation, Formal analysis. Yi-tian Yu: Validation, Supervision, Software. Yi-han Lin: Resources, Project administration. Li-jiang Han: Validation, Supervision, Software. Yi-yun Lv: Resources, Project administration. Nai-feng Tian: Project administration, Funding acquisition. Zhen Lin: Project administration, Funding acquisition. Wei-jun Guo: Project administration, Funding acquisition.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article: