Endothelial cell-derived exosomes inhibit high glucose-induced osteoblast ferroptosis by activating microRNA-335-3p/prostaglandin endoperoxide synthase 2

Abstract

BACKGROUND

Diabetic osteoporosis (DOP), a serious complication of type 2 diabetes mellitus (T2DM), involves ferroptosis-mediated disruption of bone metabolism. While endothelial cell-derived exosomes (EC-Exos) demonstrate inherent bone-targeting properties, their role in counteracting high glucose (HG)-induced osteoblast ferroptosis remains unexplored.

AIM

To investigate whether EC-Exos protect against HG-induced osteoblast ferroptosis through microRNA (miR)-335-3p-mediated regulation of prostaglandin endoperoxide synthase 2 (PTGS2) and evaluate clinical relevance in DOP.

METHODS

Mouse vascular endothelial cells (bEND.3) and osteoblasts (MC3T3E1) were used. Exosomes were isolated and subsequently characterized by transmission electron microscopy, nanoparticle tracking analysis, and western blotting for CD63 and CD81. miR expression profiles were compared between HG-treated osteoblasts and exosome-cocultured groups using high-throughput sequencing and quantitative reverse transcription polymerase chain reaction. Targeting of PTGS2 mRNA by miR-335-3p was validated by dual-luciferase reporter assay. Ferroptosis markers, reactive oxygen species, malondialdehyde, glutathione (GSH), PTGS2, GSH peroxidase 4, solute carrier family 7 member 11, and solute carrier family 3 member 2, were quantified following miR-335-3p inhibition. Serum samples from 30 T2DM patients and 32 DOP patients were analyzed. miR-335-3p levels were measured by quantitative reverse transcription polymerase chain reaction, and PTGS2 concentrations were determined via enzyme-linked immunosorbent assay. Diagnostic performance was assessed using receiver operating characteristic curves and logistic regression.

RESULTS

EC-Exos significantly reduced reactive oxygen species levels and malondialdehyde, while increasing GSH in HG-treated osteoblasts. miR-335-3p expression increased 3.7-fold in exosome-treated cells vs HG controls. miR-335-3p directly bound the PTGS2 3’ untranslated region. Inhibition of miR-335-3p abolished exosomal protection against ferroptosis, as demonstrated by increased PTGS2 expression and reduced levels of GSH peroxidase 4, solute carrier family 7 member 11, and solute carrier family 3 member 2. DOP patients exhibited lower serum miR-335-3p and higher PTGS2 compared with T2DM controls, showing a strong inverse correlation. miR-335-3p demonstrated diagnostic potential for DOP.

CONCLUSION

EC-Exos affect ferroptosis in osteoblasts induced by HG by activating miR-335-3p/PTGS2. Serum miR-335-3p may be a novel diagnostic biomarker.

Key Words: Type 2 diabetes; Osteoporosis; Exosomes; Ferroptosis; Oxidative stress

Core Tip: Endothelial cell-derived exosomes demonstrate inherent bone-targeting properties, their role in counteracting high glucose-induced osteoblast ferroptosis remains unexplored. This study reveals that endothelial cell-derived exosomes protect osteoblasts from high glucose-induced ferroptosis by regulating microRNA-335-3p/prostaglandin endoperoxide synthase 2. The diagnostic potential of serum microRNA-335-3p for diabetic osteoporosis was demonstrated, warranting further validation with an expanded sample size.

INTRODUCTION

Diabetic osteoporosis (DOP) is a common complication of diabetes. Every year, there are > 9 million cases of osteoporotic fractures worldwide; most of which are related to DOP[1]. Although increasing clinical evidence suggests that the diabetic microenvironment has a devastating impact on bone metabolism, the potential pathophysiological mechanisms and effective treatments for DOP need further research[2]. Ferroptosis may be a mechanism affecting the bone metabolic microenvironment[3,4]. Ferroptosis has unique biological characteristics, such as iron accumulation, increased production of lipid peroxides, and downregulation of glutathione (GSH) peroxidase 4 (GPX4)[5,6]. Most current treatments for preventing or treating fractures in diabetic patients are derived from treatments of postmenopausal women and patients with nondiabetic osteoporosis, and there are no unique treatment plans or drugs specifically for this disease. Therefore, actively seeking DOP-specific targets is of clinical importance.

Exosomes (Exos) are nanoscale extracellular vesicles (40-160 nm) released by numerous cells. They participate in multiple physiological and pathological processes by transferring biomolecules such as proteins, nucleic acids, and transcription factors to recipient cells[7]. Endothelial cell-derived Exos (EC-Exos) exhibit potent bone-targeting capability and have been reported to suppress osteoclast differentiation, thereby modulating the development of osteoporosis[8]. Therefore, the targeted delivery capacity of EC-Exos holds promise for offering novel perspectives and therapeutic strategies in osteoporosis treatment. Recent studies have indicated that EC-Exos can improve glucocorticoid-induced osteoblast ferroptosis[9]. However, the impact of EC-Exos on osteoblast ferroptosis induced by high glucose (HG) levels remains unclear. MicroRNAs (miRs) are endogenous single-stranded noncoding small RNAs (17-24 nucleotides) that can regulate gene expression to affect various biological functions[10]. Research has indicated that miRs are crucial in the biological functions of osteoblasts[11-13]. We hypothesized that EC-Exos might influence ferroptosis by regulating osteoblast miRs.

In this study, we selected the mouse vascular endothelial cell line bEND.3 and the mouse osteoblast cell line MC3T3E1 to observe the morphological characteristics of EC-Exos, determined the concentration and particle size distribution of Exos, and observed the capacity of MC3T3E1 cells to take up EC-Exos. Through high-throughput sequencing, we analyzed the differences in the expression of miRs in osteoblasts cultured in HG and osteoblasts cocultured with EC-Exos after culturing with HG (HG + EC-Exos). We predicted the target proteins of miRs through the miRDB database, which suggested that EC-Exos might regulate the synthesis of prostaglandin endoperoxide synthase 2 (PTGS2) by modulating osteoblast miR-335-3p to alleviate HG-induced osteoblast ferroptosis. We further verified whether EC-Exos affect HG-induced osteoblast ferroptosis through miR-335-3p/PTGS2 by inhibiting miR-335-3p. Finally, we conducted clinical verifications, selecting 30 patients with type 2 diabetes mellitus (T2DM) and 32 patients with DOP. A comparative analysis of the levels of miR-335-3p and PTGS2 in the serum of the two groups of patients was conducted. Our results revealed an inverse correlation between serum miR-335-3p and PTGS2 levels in the DOP and T2DM groups. The receiver operating characteristic (ROC) curve for miR-335-3p had an area under the curve of 0.875, indicating diagnostic significance.

MATERIALS AND METHODS

Cell culture

The bEND.3 mouse vascular endothelial cell line and MC3T3E1 mouse osteoblast cell line (both from Pricella, Wuhan, China) were maintained under standardized conditions of 37 °C in a humidified atmosphere supplemented with 5% CO₂. Both cell lines were cultured in Dulbecco’s modified Eagle’s medium (VivaCell, Shanghai, China), which was enriched with 10% fetal bovine serum and 1% penicillin-streptomycin solution. To ensure cells were in the logarithmic growth phase and retained high viability for subsequent experiments, trypsin digestion was employed for cell passage.

Isolation and identification of EC-Exos

The procedures for the isolation and identification of Exos adhered to the “Minimal Information for Studies of Extracellular Vesicles 2018” guidelines[14]. Endothelial cells were maintained in a medium supplemented with cell-free Exo-depleted fetal bovine serum. Exos were extracted from the cell culture medium utilizing a Total Exo Isolation Reagent (Thermo Fisher Scientific, Waltham, MA, United States). The medium containing Exos was first centrifuged at 300 × g for 10 minutes and subsequently at 1500 × g for 10 minutes at 4 °C. The supernatant obtained after these steps was then passed through a 0.22-μm filter. The Exo isolation reagent was combined with the medium and incubated overnight at 4 °C. The mixture underwent centrifugation at 10000 × g for 1 hour, the supernatant was discarded, and the pellet was resuspended in 500 μL 1 × phosphate-buffered saline (PBS). Exo morphology was examined using transmission electron microscopy, while the size distribution was assessed by nanoparticle tracking analysis.

Exo labeling and uptake experiments

In the growth phase, MC3T3E1 cells were plated in a 24-well cell culture plate at 10⁵ cells per well. A working solution of the red fluorescent dye PKH26 Exo (Solarbio, Beijing, China) was prepared under dark conditions. This was achieved by diluting the PKH26 stock solution with Diluent B in a 9:1 ratio (Diluent B to PKH26 stock solution by volume). For every 500 μg of Exo protein, 100 μL dye working solution was required. An adequate volume of the dye working solution was added to the EC-Exos, which were subsequently mixed using vortexing for 1 minute and incubated in the dark for 15 minutes. Following this, the PKH26-labeled EC-Exos were purified via ultracentrifugation to eliminate any excess dye. The PKH26-labeled EC-Exos were introduced into the culture wells and cocultured with the cells for 6 hours. After this incubation period, the culture medium was removed, and the cells were rinsed with PBS for 5 minutes, which was repeated twice. After washing with PBS, 4,6-diamidino-2-phenylindole solution was applied, and the cells were allowed to incubate at room temperature for a further 5 minutes. Post-incubation, the cells underwent a final wash with PBS for 5 minutes, which was repeated twice. The cells were examined and photographed using a fluorescence microscope, and the acquired images were saved for further analysis.

High-throughput sequencing

Total RNA was extracted from the HG and HG + EC-Exos groups utilizing TRIzol reagent, followed by comprehensive high-throughput sequencing analyses to assess mRNA and miR profiles. A threshold of |log₂-fold change| ≥ 1 was established for identifying significant alterations, accompanied by q < 0.05 (where q represents adjusted P value) for the selection of differentially expressed genes. Differentially expressed genes were analyzed through Gene Ontology and Kyoto Encyclopedia of Genes and Genomes enrichment analyses within designated comparison groups. Additionally, the identification of key miR-targeting mRNAs was performed, integrating this with ferroptosis enrichment analysis related to the miR target genes.

Quantitative reverse transcription polymerase chain reaction to validate differential miRs

The 10 miRs that exhibited significant upregulation in the HG + EC-Exos group relative to the HG group were chosen for further validation. Total RNA was isolated from samples of the HG and HG + EC-Exos groups. Following this, the RNA underwent reverse transcription to synthesize complementary DNA. Detection was performed utilizing a SYBR Green quantitative reverse transcription polymerase chain reaction (PCR) kit (Biosharp, Hefei, China). Primer details are provided in Supplementary Table 1.

Luciferase validation of the target gene

The wild-type PTGS2 luciferase reporter plasmid was constructed by cloning the 3’-untranslated region (UTR) of PTGS2, containing the predicted miR-335-3p binding site (5’-ATGAAAAATG-3’), into the pEZX-MT06 vector (Gene Copoeia, Anhui, China). Subsequently, the 3’ UTR of PTGS2 featuring a mutated miR-335-3p binding site was also cloned and inserted into the pEZX-MT06 vector, leading to the creation of the mutant PTGS2 luciferase reporter plasmid, with the mutated binding site sequence represented as 5’-TTTTTTTTTG-3’. Twenty-four hours prior to the transfection process, MC3T3 cells were plated in a 24-well plate at 105 cells per well. Following a 24-hour incubation period, once the cell confluence reached 70%-80%, transfection was performed using Lip3000 (Thermo Fisher Scientific, Waltham, MA, United States), with simultaneous transfection of 1 μg DNA and 15 pmol miR; three replicate wells were established for each condition. The cells were subsequently lysed, and reaction solutions for both Firefly and Renilla luciferase were prepared, allowing them to equilibrate at room temperature. A volume of 20 L of the cell lysate supernatant was dispensed into a 96-well plate. Firefly luciferase reaction solution was added, followed by mixing by plate shaking, and luciferase activity was quantified at a wavelength of 560 nm. The Renilla luciferase reaction solution was introduced, mixed in the same manner, and luciferase activity was assessed at a wavelength of 465 nm.

Cell transfection

The cells were transfected with the mmu-miR-335-3p inhibitor, mmu-miR-335-3p negative control (NC, General Bio, Anhui, China) (for the oligonucleotide sequences, see Supplementary Table 2). MC3T3E1 cells were seeded into six-well plates (10⁶ cells per well) and cultured at 37 °C until they reached 80%-90% confluence. MC3T3E1 cells were cultured in medium supplemented with 25 mmol/L glucose for 48 hour and transfected with the above vectors by Lipofectamine 2000 (Invitrogen, Carlsbad, CA, United Staates).

Cell grouping

MC3T3E1 cells were divided into control, HG, HG + Exos, HG + Exos + NC, and HG + Exos + inhibitor groups. The control group was cultured for 48 hours at a glucose concentration of 5.5 mmol/L; the HG group was cultured for 48 hours at a glucose concentration of 25 mmol/L; the HG + Exos group was cocultured with 20 μL Exos for 6 hours under HG conditions; the HG + Exos + NC group was transfected with NC sequences in cells cultured with a glucose concentration of 25 mmol/L for 48 hours and then cocultured with 20 μL Exos for 6 hours; and the HG + Exos + inhibitor group was transfected with mmu-miR-335-3p inhibitor sequences in cells cultured with a glucose concentration of 25 mmol/L for 48 hours and then cocultured with 20 μL Exos for 6 hours. Exosomal protein content was determined with a BCA assay kit (Thermo Fisher Scientific, Rockford, IL, Unites States), and a standardized dose of 5 μg per treatment was applied to ensure efficient cellular uptake.

Reactive oxygen species measurement

The overall concentration of intracellular reactive oxygen species (ROS) was evaluated utilizing a ROS Assay Kit (Beyotime, Nantong, China). After incubation, cells from each group were collected and diluted with 2’,7’-dichlorofluorescin diacetate (DCFH-DA) at a 1:1000 ratio to a final concentration of 10 μmol/L. The cells were then resuspended in the diluted DCFH-DA solution at a density of 10⁵ cells/mL and incubated at 37 °C for 20 minutes. To ensure optimal interaction between the probe and the cells, the mixture was agitated every 3-5 minutes. Following incubation, the cells were washed three times with serum-free cell culture medium to eliminate any unincorporated DCFH-DA. The fluorescence intensity of each group was quantified using a fluorescence microplate reader. The levels of intracellular ROS were indicated by the fluorescence intensity of the DCFH-DA probe, which exhibited an excitation wavelength of 488 nm and emission wavelength of 525 nm. Images were captured using a fluorescence microscope. Each experimental condition was assessed in triplicate wells.

Malondialdehyde measurement

Cells were harvested following the protocol outlined in the malondialdehyde (MDA) assay kit (Biyun, Nantong, China). The cells were homogenized in a buffer or lysis solution at specified ratios, thoroughly mixed, and incubated on ice for 10 minutes to facilitate lysis. The resulting homogenate was subjected to centrifugation at 15000 × g for 10 minutes, allowing for collection of the supernatants for subsequent analysis. MDA concentrations were quantified utilizing a sensitive assay kit (Biyun, Nantong, China). Samples in 96-well plates were incubated with working solution at 37 °C for designated durations. Absorbance was measured spectrophotometrically, and MDA levels were quantified according to the reagent protocol.

GSH assay

Cells from each group were digested and pelleted by centrifugation at 10000 × g for 10 minutes at 4 °C according to the protocol of the GSH assay kit (BiYuntian, Nantong, China). Subsequent to sonication, a 10-μL aliquot of the cell suspension was extracted and combined with a suitable volume of precipitant, ensuring thorough mixing. The resulting samples were subjected to centrifugation at 2000 × g for 10 minutes, after which, the supernatants were harvested. Additionally, a 20-μL solution of GSH standard was prepared, and an appropriate quantity of buffer and colorimetric developer was incorporated following the previously outlined protocol. Upon thorough mixing, the solution was allowed to incubate for 5 minutes, after which its absorbance was recorded at 405 nm. The GSH concentration was subsequently computed using the formula specified in the GSH assay kit.

Reverse transcription-quantitative PCR

Total RNA was isolated from cellular samples or Exos utilizing TRI Reagent (Biosharp, Hefei, China), followed by quantification and assessment of RNA quality via spectrophotometry. Subsequently, 2 μg extracted total RNA was used to synthesize high-quality complementary DNA with amplification conducted using the Prime Script RT Master Mix Kit (TaKaRa, Shiga, Japan). For reverse transcription-quantitative PCR, the SYBR Green RT-PCR Kit (Biosharp, Hefei, China) was utilized. The thermal cycling protocol consisted of an initial denaturation at 95 °C for 30 seconds, followed by 40 cycles of denaturation at 95 °C for 5 seconds and annealing/extension at 60 °C for 30 seconds. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and U6 served as endogenous controls for mRNA and miR, respectively. The relative expression levels of RNA were determined using the 2^-ΔΔCt method (Supplementary Table 3 for primer details).

Western blotting

Each cell group or exosomal sample was subjected to lysis utilizing radioimmunoprecipitation assay lysis buffer for the extraction of total protein. Following this, centrifugation was conducted at 12000 × g for 15 minutes at 4 °C to isolate the supernatant, which contained the total protein. After denaturation, the proteins were separated by electrophoresis and transferred onto a membrane. The membrane was blocked with 5% bovine serum albumin and subsequently incubated overnight at 4 °C with specific primary antibodies (Zhongshan Jinqiao, Beijing, China) against GAPDH, PTGS2, GPX4, solute carrier family 7 member 11 (SLC7A11), solute carrier family 3 member 2 (SLC3A2), as well as the exosomal markers CD63 and CD81, all at a dilution of 1:1000. Thereafter, the membrane was incubated with a horseradish peroxidase-conjugated goat anti-rabbit secondary antibody (Zhongshan Jinqiao, Beijing, China) diluted at 1:5000 for 1 hour at room temperature. Protein bands were visualized using a Western ECL substrate kit (Thermo Fisher Scientific, Rockford, IL, United States). Images were acquired with a Bio-Rad scanner (Bio-Rad Laboratories, Inc., Hercules, CA, United States) and quantitatively analyzed using ImageJ software (Version 1.53k, National Institutes of Health, Bethesda, MD, United States). GAPDH was used as an internal reference to evaluate the expression levels of PTGS2, GPX4, SLC7A11, SLC3A2, CD63, and CD81.

Clinical data collection

Data were collected from 32 female patients with T2DM and DOP who were treated at the Second Affiliated Hospital and the First Affiliated Hospital of Bengbu Medical University from January 2023 to December 2023 (DOP group). The inclusion criteria were as follows: (1) T2DM meeting the diagnostic criteria of the “Chinese Guidelines for the Prevention and Treatment of Type 2 Diabetes (2020 Edition)”[15]; (2) Osteoporosis diagnosed according to the “Primary Osteoporosis Diagnosis and Treatment Guidelines (2017)”[16], with a bone density T-score ≤ -2.5; (3) Aged ≥ 55 years and naturally postmenopausal; and (4) No history of fractures or antiosteoporosis treatment. The exclusion criteria were as follows: (1) Secondary osteoporosis due to other causes; (2) Other severe underlying diseases; (3) Acute or chronic infection; (4) Long-term use of anticoagulants, sex hormones, or other drugs with adverse effects on bone; (5) Incomplete data; (6) Acute complications of diabetes; and (7) Diseases affecting bone metabolism, such as Cushing’s syndrome. This study was approved by the Ethics Committee of Bengbu Medical College. Additionally, 30 patients with T2DM during the same period comprised the T2DM group. The general clinical data of both groups were collected, and serum miR-335-3p expression was determined by reverse transcription-quantitative PCR, and serum PTGS2 levels were determined by enzyme-linked immunosorbent assay (ELISA).

ELISA for determination of serum PTGS2 levels

Fasting venous blood (6 mL) was drawn from both groups and centrifuged at 2000 × g for 10 minutes at 4 °C to remove impurities, and the supernatant was collected and stored at 80 °C. 2 mL of serum was used for determination of PTGS2 levels by ELISA (Tiande Biotechnology, Wuhan, China).

Statistical analysis

Statistical analysis was performed with GraphPad Prism 6.0 software (GraphPad, San Diego, CA, United States) for analysis of three independent experiments. Continuous variables with a normal distribution are presented as the mean ± SD; non-normal variables are reported as median (interquartile range). Comparisons between two groups and among multiple groups for continuous variables were made with Student’s t test or one-way analysis of variance, followed by Tukey’s test. P < 0.05 indicated statistical significance. The efficacy of serum miR-335-3p in predicting fractures was determined by ROC curves. Multivariate logistic regression was used to analyze the factors affecting DOP.

RESULTS

EC-Exos can be taken up by osteoblasts

The Exos obtained in this investigation exhibited a cup-like morphology under electron microscopy (Figure 1). Nanoparticle tracking analysis revealed Exos with a size distribution of 74-193 nm, a peak at 82 nm, and a mean size of 114.6 ± 57.3 nm (Figure 1). Fluorescence microscopy indicated that after 6 hours coculture with PKH26-labeled Exos, significant accumulation of red fluorescence was observed within the osteoblast membrane and cytoplasm (Figure 1). These findings suggest that PKH26-labeled Exos are capable of being internalized by osteoblasts, thereby facilitating intercellular communication. The presence of characteristic exosomal markers, CD63 and CD81, was confirmed by western blotting (Figure 1).

Figure 1 Endothelial cell-exosomes identification and uptake. A: Transmission electron microscopy, orange arrows indicate exosomes; B: Nanoparticle tracking analysis for exosome size distribution; C: Western blotting for detection of exosomes surface markers; D: Exosome phagocytosis experiment. ECs: Endothelial cells; EC-Exos: Endothelial cell-exosomes.

EC-Exos inhibit ferroptosis in osteoblasts cultured in high-glucose conditions

In comparison to the control group, the HG group showed increases in ROS and MDA levels, alongside significant reductions in GSH levels (P < 0.05). Conversely, the HG + EC-Exos group demonstrated markedly lower levels of ROS and MDA, as well as significantly higher levels of GSH when compared to the HG group (P < 0.05; Figure 2A-D). The HG group showed an increase in PTGS2 mRNA expression, coupled with a significant decline in expression of SLC3A2, SLC7A11, and GPX4 compared to the control group (P < 0.05). In contrast, the HG + EC-Exos group exhibited reduced PTGS2 mRNA levels and significant upregulation of mRNA expression of SLC3A2, SLC7A11, and GPX4 relative to the HG group (P < 0.05; Figure 2E).

Figure 2 Endothelial cell-exosomes attenuate high glucose-induced oxidative stress and ferroptosis. A-C: Comparison of reactive oxygen species levels, malondialdehyde levels and glutathione levels among the control group, high glucose (HG) group, and HG + exosomes (Exos) group; D: Reactive oxygen species fluorescence images of the control group, HG group, and HG + Exos group; E: Comparison of mRNA expression levels of ferroptosis indicators prostaglandin endoperoxide synthase 2, solute carrier family 3 member 2, solute carrier family 7 member 11, glutathione peroxidase 4 among the control group, HG group, and HG + Exos group. The bars indicate the mean ± SD from three independent experiments (n = 3). ^aP < 0.05 vs control, ^bP < 0.05 vs high glucose. Con: Osteoblasts treated with normal glucose; HG: Osteoblasts treated with high glucose; HG + Exos: High glucose + endothelial cell-exosomes; ROS: Reactive oxygen species; DCF: Dichlorofluorescein; MDA: Malondialdehyde; GSH: Glutathione; PTGS2: Prostaglandin endoperoxide synthase 2; SLC3A2: Solute carrier family 3 member 2; SLC7A11: Solute carrier family 7 member 11; GPX4: Glutathione peroxidase 4.

Differentially expressed miRs in the HG + EC-Exos group and HG group

High-throughput sequencing indicated that, in contrast to the HG group, the HG + EC-Exos group had 54 downregulated miRs and 40 upregulated miRs (Figure 3). Following filtering of expression abundance, the 10 most significantly upregulated miRs in the HG + EC-Exos group compared to the HG group were miR-677-3p, miR-6240, miR-210-3p, miR-702-3p, miR-5126, miR-335-3p, miR-6538, miR-126a-5p, miR-5099, and miR-6239 (Figure 3).

Figure 3 High glucose + exosomes and high glucose group microRNAs differential expression analysis. A: Heatmap of differentially expressed microRNAs (miRs), red to blue represents expression levels from high to low; B: Volcano plot of differentially expressed miRs. Orange dots represent up regulated miRs, green dots represent down regulated miRs, gray dots represent no difference; C: Validation of the top 10 miRs that were significantly upregulated in high glucose (HG) + exosomes (Exos) compared to HG expression, ^aP < 0.05 vs high glucose; D: Differences in miR-335-3p expression among Exos, HG group, and HG + Exos group; E: Interaction diagram of miR with target proteins; F: Kyoto Encyclopedia of Genes and Genomes pathway analysis diagram. The bars indicate the mean ± SD from three independent experiments (n = 3). ^bP < 0.05 vs high glucose, ^cP < 0.05 vs high glucose + exosomes. Con: Osteoblasts treated with normal glucose; HG: Osteoblasts treated with high glucose; HG + Exos: High glucose + endothelial cell-exosomes; miR: MicroRNA; Exos: Exosomes.

Verification of the top 10 miRs and target gene screening

In the comparative analysis between the HG + EC-Exos and HG groups, upregulation of specific miRs was observed, including miR-6538, miR-6240, miR-5126, miR-210-3p, miR-126a-5p, miR-335-3p, miR-6239, miR-677-3p, miR-702-3p, and miR-5099 (Figure 3). Kyoto Encyclopedia of Genes and Genomes pathway analysis indicated the involvement of ferroptosis (Figure 3), while protein interaction analysis revealed that miR-335-3p had the potential to target PTGS2 (Figure 3). To ascertain the origins of miR-335-3p in Exos, we examined the variations in miR-335-3p levels in the EC-Exos, HG, and HG + EC-Exos groups. The concentration of miR-335-3p was significantly diminished in the EC-Exos group compared to the HG group (P < 0.05). Conversely, the levels of miR-335-3p were markedly elevated in the HG + EC-Exos group relative to the HG group (P < 0.05; Figure 3).

Luciferase assay validation of mmu-miR-335-3p binding to PTGS2 mRNA

Our predictions indicated that miR-335-3p could interact with the 321-328 region of the PTGS2 3’ UTR (Figure 4A). To confirm this interaction, we performed a dual-luciferase reporter assay. The observed dual-luciferase activity ratio resulting from cotransfection of mmu-miR-335-3p with PTGS2-3’ UTR-wild-type was markedly reduced compared to that in the control group (P < 0.05). In contrast, the dual-luciferase activity ratio when mmu-miR-335-3p was cotransfected with PTGS2-3’ UTR-mutant showed no significant difference when compared to the control group (P > 0.05; Figure 4B). These findings imply that mmu-miR-335-3p is capable of targeting PTGS2.

Figure 4 MicroRNA-335-3p targets prostaglandin endoperoxide synthase 2 3’ untranslated region: Prediction and validation. A: MicroRNA binding site database prediction of microRNA-335-3p binding site with prostaglandin endoperoxide synthase 2 3’ untranslated region; B: Dual-luciferase reporter gene assay. ^aP < 0.05. PTGS2: Prostaglandin endoperoxide synthase 2; UTR: Untranslated region; WT: Wild type; miR: MicroRNA; Mut: Mutant; NC: Negative control.

Effect of EC-Exos on ROS, MDA, and GSH in osteoblasts induced by HG after inhibition of miR-335-3p

To evaluate the impact of miR-335-3p on osteoblast impairment induced by HG, a miR-335-3p inhibitor was introduced into HG-treated osteoblasts, leading to a notable reduction in miR-335-3p expression levels (Figure 5A). The concentrations of ROS and MDA were markedly elevated in the HG group compared to the control group (P < 0.01), while GSH levels were significantly decreased (P < 0.01). In contrast, the HG + Exos group exhibited significantly reduced levels of ROS and MDA, alongside a substantial increase in GSH levels compared to the HG group (P < 0.01). When comparing the HG + Exos + inhibitor group to the HG + Exos+ NC group, there was a significant increase in ROS and MDA levels, coupled with a marked decrease in GSH levels (P < 0.01; Figure 5B-E).

Figure 5 Inhibition of microRNA-335-3p abolishes endothelial cell-exosomes-mediated protection against high glucose-induced oxidative stress in osteoblasts. A: Detection of microRNA-335-3p levels in osteoblasts induced by high glucose after inhibition of microRNA-335-3p by reverse transcription-quantitative polymerase chain reaction; B-D: Results of reactive oxygen species, malondialdehyde and glutathione quantitative analysis; E: Detection of cellular reactive oxygen species levels in each group using dichlorofluorescein fluorescence probe. The bars indicate the mean ± SD from three independent experiments (n = 3). ^aP < 0.05 vs control, ^bP < 0.01 vs control, ^cP < 0.01 vs high glucose, ^dP < 0.01 vs high glucose + exosomes + negative control. Con: Osteoblasts treated with normal glucose; HG: Osteoblasts treated with high glucose; HG + Exos: High glucose + endothelial cell-exosomes; HG + Exos + NC: High glucose + negative control transfection + endothelial cell-exosomes; HG + Exos + inhibitor: High glucose + microRNA-335-3p inhibitor + endothelial cell-exosomes.

The effect of EC-Exos on HG-induced osteoblast ferroptosis after the inhibition of miR-335-3p

In comparison to the control group, the HG group exhibited significant elevations in mRNA and protein expression of PTGS2, alongside notable reductions in mRNA and protein expression of GPX4, SLC7A11, and SLC3A2 (P < 0.01). Conversely, compared to the HG group, the HG + Exos group showed decreased PTGS2 but increased GPX4, SLC7A11, and SLC3A2 expression at both mRNA and protein levels (P < 0.01). Following knockdown of miR-335-3p, the HG + Exos + inhibitor group showed an increase in mRNA and protein expression of PTGS2, along with a decrease in mRNA and protein expressions of GPX4, SLC7A11, and SLC3A2, compared to the HG + Exos + NC group (P < 0.01; Figure 6A-C).

Figure 6 Inhibition of microRNA-335-3p abolishes endothelial cell-exosomes-mediated protection against high glucose-induced ferroptosis in osteoblasts. A: The relative expression levels of mRNA for prostaglandin endoperoxide synthase 2 (PTGS2), glutathione peroxidase 4 (GPX4), solute carrier family 7 member 11 (SLC7A11), and solute carrier family 3 member 2 (SLC3A2) in each group of cells; B: Western blotting to detect the protein expressions of key ferroptosis markers PTGS2, GPX4, SLC7A11, SLC3A2; C: The relative protein expressions of PTGS2, GPX4, SLC7A11, SLC3A2. The bars indicate the mean ± SD from three independent experiments (n = 3). ^aP < 0.01 vs control, ^bP < 0.01 vs high glucose, ^cP < 0.01 vs high glucose + exosomes + negative control. Con: Osteoblasts treated with normal glucose; HG: Osteoblasts treated with high glucose; HG + Exos: High glucose + endothelial cell-exosomes; HG + Exos + NC: High glucose + negative control transfection + endothelial cell-exosomes; HG + Exos + inhibitor: High glucose + microRNA-335-3p inhibitor + endothelial cell-exosomes; PTGS2: Prostaglandin endoperoxide synthase 2; SLC3A2: Solute carrier family 3 member 2; SLC7A11: Solute carrier family 7 member 11; GPX4: Glutathione peroxidase 4; GAPDH: Glyceraldehyde 3-phosphate dehydrogenase.

Comparison of serum miR-335-3p and PTGS2 levels between DOP and T2DM patient groups

We conducted validations at the clinical level. To avoid the impact of sex on osteoporosis, we selected female patients with T2DM and DOP (comparisons of the clinical data are shown in Table 1). Variables that were significantly associated with DOP in the univariate analyses (P < 0.05), including age, body mass index, and fasting C-peptide, were then entered into a multivariable logistic regression model. Logistic regression analysis suggested that the influencing factors for DOP were miR-335-3p, body mass index, and fasting C-peptide (Table 2). Serum PTGS2 levels were significantly higher in the DOP group than in the T2DM group (P < 0.05), whereas miR-335-3p expression was notably lower in DOP group compared to T2DM group (P < 0.05; Figure 7A and B). In the DOP and T2DM groups, serum miR-335-3p levels were negatively correlated with PTGS2 levels, with a Pearson correlation coefficient of 0.954 (Figure 7C). The ROC curve for miR-335-3p was plotted, with an area under the curve of 0.875, which indicated a certain diagnostic significance (Figure 7D).

Figure 7 Altered serum levels and diagnostic significance of prostaglandin endoperoxide synthase 2 and microRNA-335-3p in diabetic osteoporosis. A: Comparison of serum prostaglandin endoperoxide synthase 2 levels between the type 2 diabetes and diabetic osteoporosis groups; B: Comparison of serum microRNA (miR)-335-3p levels between the type 2 diabetes and diabetic osteoporosis groups; C: Correlation analysis of serum prostaglandin endoperoxide synthase 2 with miR-335-3p; D: Receiver operating characteristic curve analysis of miR-335-3p. ^aP < 0.05. DM: Patients with type 2 diabetes; DOP: Patients with type 2 diabetes with osteoporosis; miR: MicroRNA; PTGS2: Prostaglandin endoperoxide synthase 2; AUC: Area under the curve; CI: Confidence interval.

Table 1 Comparison of clinical data between diabetic osteoporosis and type 2 diabetes patients, mean ± SD.

Group	n	Age (years)	BMI (kg/m2)	Fasting C-peptide (ng/mL)	FBG (mmol/L)	HbA1c (%)	MicroRNA335-3p	PTGS2 (ng/mL)
DM	32	56.63 ± 8.42	24.36 ± 1.91	2.29 ± 1.21	9.26 ± 1.78	9.19 ± 1.13	2.86 (2.25, 3.36)	10.52 (8.25, 12.91)
DOP	30	65.40 ± 8.60	21.12 ± 1.52	1.29 ± 0.59	10.75 ± 2.23	10.28 ± 2.28	1.56 (1.03, 2.52)	15.03 (12.66, 16.77)
Z/t		4.06a	-7.44a	-4.10a	2.88a	2.62a	-5.072a	-5.163a

^aP < 0.05.

DM: Patients with type 2 diabetes; DOP: Patients with type 2 diabetes with osteoporosis; BMI: Body mass index; FBG: Fasting blood glucose; HbA1c: Glycosylated hemoglobin; PTGS2: Prostaglandin endoperoxide synthase 2.

Table 2 Factors affecting diabetic osteoporosis in binary logistic regression analysis.

Factors	β	Standard Error	Wald χ2	OR (95%CI)
MicroRNA-335-3p	-2.644	1.274	5.695	0.071 (0.006-0.863)a
BMI	-0.511	0.256	3.983	0.600 (0.363-0.991)a
Fasting C-peptide	-1.553	0.770	4.064	0.212 (0.047-0.958)a

^aP < 0.05.

BMI: Body mass index; OR: Odds ratio; CI: Confidence interval.

DISCUSSION

The mechanism by which diabetes leads to osteoporosis is largely unclear, but oxidative stress caused by high blood glucose is currently considered an important factor. Oxidative stress promotes the release of ROS, leading to mitochondrial dysfunction, which in turn leads to bone tissue cell death[17]. Ferroptosis represents a novel form of programmed cell death, which is identified by an accumulation of iron and lipid peroxidation that is reliant on ROS[18]. Biochemically, ferroptosis is characterized primarily by lipid peroxidation, which is often accompanied by iron overload, excessively high levels of ROS, and reduced activity of antioxidant enzymes[19]. Among these, the cystine/glutamate antiporter (also known as System Xc^-)/GSH/GPX4 system plays a crucial role as an antioxidant regulatory pathway in the process of ferroptosis. This system comprises transporter protein SLC7A11 and regulatory protein SLC3A2, which facilitate the exchange of intracellular glutamate for extracellular cysteine. GPX4 is capable of utilizing GSH as a cofactor to convert phospholipid hydroperoxides into their corresponding alcohols[19]. In a mouse model of T2DM-induced osteoporosis, ferroptosis caused by disorders of glucose and lipid metabolism was verified. In DOP mice, serum ferritin levels were significantly elevated, while SLC7A11 and GPX4 expression in bone tissue was markedly reduced. These mice also displayed pronounced trabecular bone deterioration and bone loss, consistent with decreased GPX4 and elevated PTGS2 expression[20]. Our research corroborated these findings, demonstrating that in osteoblasts cultured under HG conditions, expression of PTGS2 was elevated. Additionally, the level of MDA, a byproduct of lipid peroxidation, was increased, whereas the level of ROS was significantly increased. Conversely, GSH levels were decreased. Key ferroptosis-related markers, including SLC7A11, SLC3A2, and GPX4, showed significantly decreased expression.

PTGS2, also known as cyclooxygenase 2, is an enzyme that accelerates lipid peroxidation and is an early direct product of inflammation. Recent research has suggested that PTGS2 could be a key ferroptosis gene in diabetic retinopathy and diabetic nephropathy[21]. PTGS2 promotes the maturation of osteoclasts through prostaglandin E2, thereby facilitating the occurrence of osteoporosis[22]. Recent research has demonstrated that PTGS2 plays a crucial role in the mechanism underlying femoral head necrosis[23]. In our study, we observed that serum levels of PTGS2 in individuals diagnosed with DOP were markedly elevated compared to those in patients with T2DM alone. This finding aligns with previous investigations, indicating the significant involvement of PTGS2 in DOP pathology. Considering the role of PTGS2 in both diabetes and disorders related to bone metabolism, it is plausible that ferroptosis may serve as a shared mechanism underlying these conditions.

Research suggests that Exos are important effector molecules and carriers for intercellular material exchange and information transmission, regulating the cellular microenvironment and participating in various biological processes of target cells, such as promoting angiogenesis and anti-immunogenicity, inhibiting cell apoptosis, and affecting tumor cell proliferation, antifibrosis, and various anti-inflammatory processes[24-26]. Recent evidence suggests that Exos may not only facilitate intercellular communication but also modulate ferroptosis in recipient cells, highlighting their potential as a novel therapeutic strategy[27]. Blood vessels are crucial for bone formation, with endothelial cells lining their inner surface actively internalizing and secreting biologically active molecules. Studies have shown that EC-Exos can inhibit osteoclast differentiation by upregulating miR-155, which affects the process of osteoporosis[8]. Compared with Exos secreted by osteoblasts and bone marrow mesenchymal stem cells, EC-Exos exhibit stronger bone-targeting properties, which are related to their high expression of pregnancy zone protein on the surface. Therefore, the unique targeting of EC-Exos is expected to provide new insights and methods for the treatment of osteoporosis. Multiple studies have suggested that miRs play important roles in the biological functions of osteoblasts, and we hypothesized that EC-Exos might regulate the miRs of osteoblasts under HG conditions[11-13]. Therefore, we analyzed the differential miR expression in the HG + Exos and HG groups and, through ferroptosis enrichment and target gene prediction, we found that EC-Exos might affect PTGS2 by regulating miR-335-3p to improve ferroptosis induced by HG in osteoblasts. Next, we analyzed the differences in miR-335-3p in the EC-Exos, HG, and HG + EC-Exos groups. The results suggested that the level of miR-335-3p was lower in the EC-Exos group but significantly greater in the HG + EC-Exos group, indicating that the miR-335-3p in the HG + EC-Exos group did not directly originate from the EC-Exos. However, it was certain that the EC-Exos could function through miR-335-3p in osteoblasts. Consequently, we proceeded to suppress miR-335-3p in osteoblasts exposed to HG levels. The results indicated that the ability of Exos to mitigate ferroptosis in osteoblasts triggered by HG was notably diminished. This suppression led to elevated levels of ROS, a reduction in MDA and GSH, and an increase in PTGS2 expression. Additionally, the expression of pivotal molecules involved in the ferroptosis pathway, such as GPX4, SLC7A11, and SLC3A2, increased. Recent research[28] has suggested that Exos from human adipose-derived stem cells induce osteogenic differentiation through the miR-335-3p/apelin receptor axis and improve osteoporosis. Our study suggested that EC-Exos can alleviate ferroptosis in osteoblasts cultured in HG through miR-335-3p/PTGS2, providing new insights for the treatment of DOP. These findings corroborate our hypothesis.

In diabetes-related research, miR-335-3p has been shown to improve insulin resistance by affecting macrophage polarization[29] and play a role in protecting islet function[30]. In patients with diabetic retinopathy, the level of miR-335-3p in plasma is significantly reduced[31]. In bone and joint diseases, in advanced knee osteoarthritis, miR-335-3p might be a regulatory gene target in the infrapatellar fat[32]. In the nucleus pulposus and blood of patients with lumbar disc degeneration, the expression of miR-335-3p is significantly reduced, which activates the NOD-like receptor protein 3 inflammasome and promotes the progression of inflammation[33]. Comprehensive analysis of miRs in the peripheral blood of humans with osteoporosis has suggested that miR-335-3p is significantly downregulated[34]. These studies suggest that miR-335-3p is a key target for regulating inflammatory bone diseases and diabetes. Consistent with previous studies, our research suggested that miR-335-3p levels were negatively correlated with PTGS2 in the serum of both DOP patients and T2DM patients, indirectly confirming that miR-335-3p might improve DOP by regulating PTGS2. The ROC curve analysis indicated that miR-335-3p had diagnostic value; however, the limited sample size in this study precluded age-matched subgroup analyses and necessitates further validation in larger cohorts to confirm its diagnostic potential.

Although this study indicated that EC-Exos had therapeutic effects on HG-induced osteoporosis, the exosomal components involved in this effect have not yet been identified. Exos contain a large number of functional proteins, mRNAs, long noncoding RNAs, and miRs, which can be taken up and internalized by osteoblasts to regulate transcription. However, it remains unclear which molecules in Exos played a significant role, and further research including in vivo validation in animal models, is needed to confirm these findings and assess their translational potential. In summary, our study showed that EC-Exos improve HG-induced osteoblast ferroptosis by activating miR-335-3p/PTGS2, providing new insights into the treatment of DOP.

CONCLUSION

EC-Exos affect ferroptosis in osteoblasts induced by HG by activating miR-335-3p/PTGS2. Serum miR-335-3p may be a novel diagnostic biomarker.