β-Ecdysone (purity ≥98%) was sourced from MCE Biotechnology (New Jersey, United States), while PLGA (lactide-to-glycolide molar ratio 50:50, MW 160,000) was supplied by Daigang Biotechnology Co., Ltd. (Shandong, China). Polyethylene glycol (PEG) and dichloromethane were acquired from Shanghai Aladdin Biochemical Technology Co., Ltd. (Shanghai, China). The CCK8 kit was obtained from Guangzhou Saiguo Biotechnology Co., Ltd. (Guangzhou, China) and Live/Dead Staining Kit were provided by Shanghai Biyuntian Biotechnology Co., Ltd. (Shanghai, China). Streptozotocin was supplied by Shanghai McLean Biochemical Technology Co., Ltd. (Shanghai, China). Simulated body fluid (SBF), Hematoxylin and Eosin (H&E) staining kit, and Masson’s trichrome staining kit were obtained from Beijing Solepol Science and Technology Co., Ltd. (Beijing, China). RUNX2 antibody was sourced from Wuhan Doctoral Bioengineering Co., Ltd. (Wuhan, China).

PLGA scaffolds were fabricated using the solution casting/particle leaching method. Sodium chloride (NaCl) of varying particle sizes (large: 250–400 μm, small: 100–250 μm) was selected as the porogen. Small-pore, large-pore, and composite scaffolds were each prepared using this method. Polyethylene glycol (PEG), known for its excellent aqueous solubility, biocompatibility, and safety, was also used alongside NaCl as a porogen to enhance the pore connectivity of the PLGA scaffolds (Thadavirul et al., 2014). To prepare the scaffolds, a 10% w/v solution of PLGA was first created by dissolving a specific weight of PLGA and PEG (in a 1:1 ratio) in dichloromethane. Sodium chloride (NaCl) of various particle sizes was screened and added to the PLGA solution at a ratio of 1: 15 (PLGA: NaCl), followed by thorough mixing. This mixture was then poured into a circular mold with a diameter of 5 mm and a thickness of 0.5 mm, yielding two types of monolayer PLGA modules with different pore sizes. The monolayer modules were bonded together using methylene chloride as an adhesive. As a result, large-pore-size scaffolds (S1), small-pore-size scaffolds (S2), and composite scaffolds (S3) were prepared. The bonded modules were then soaked in deionized water for 3 days, with the water replaced every 8 h. After the porogen was fully dissolved, the modules were freeze-dried for 24 h. To fabricate the drug-loaded scaffolds, a 1% w/v β-Ecdysone solution was prepared, and the scaffolds were soaked in this solution for 24 h before being freeze-dried. The specific procedure is illustrated in Figure 1.

Twenty-four SPF-grade male SD rats, aged 6–8 weeks, were obtained from the Laboratory of Experimental Animals at Guangxi Medical University in Nanning, China. The rats were randomly assigned to two groups: normal (N) and diabetic (DM), with 12 rats in each group. The diabetic model was established following a previously reported method (Tawulie et al., 2023). Briefly, the rats in the N group were fed a standard chow diet, while those in the DM group were given a high-sugar, high-fat diet for 4 weeks. After this period, the DM group received an intraperitoneal injection of 1% streptozotocin (STZ) at a dose of 30 mg/kg, whereas the N group received an injection of sodium citrate buffer at the same dose. Feeding continued for an additional 4 weeks, and fasting blood glucose was measured weekly from blood samples collected from the tail vein. A fasting blood glucose level of ≥11.1 mmol/L for 4 consecutive weeks was used as the criterion for the successful establishment of the diabetic model; rats that did not meet this criterion were excluded.

The rats in the normal group were randomly divided into two subgroups: a blank group (Blank) and a normal control group (N + CS), with 6 rats in each. Similarly, the diabetic group was randomly divided into a PLGA composite scaffold group (DM + CS) and a β-E/PLGA composite scaffold group (DM + CS-βE), also with 6 rats per group. A cranial bone defect model was then constructed for all rats. After anesthetizing the rat with 2% sodium pentobarbital, place it in a prone position and make a 3 cm longitudinal incision along the skin of the cranial roof using a No. 11 surgical blade. Separate the skin, muscle, and periosteum layer by layer. Two 5 mm critical bone defects were created on both sides of the midline of the rat’s skull using a Lowspeed Handpiece (NSK, Japan) equipped with a 5 mm diameter deboning drill. PLGA composite scaffolds were implanted in the N + CS and DM + CS groups, β-E/PLGA composite scaffolds were implanted in the DM + CS-βE group, and no material was implanted in the Blank group. The skin and mucous membranes were tightly sutured, and penicillin was administered once daily for 3 days post-surgery (Figure 2). At 4 and 8 weeks postoperatively, the rats were euthanized by CO₂, and the skulls were collected for micro-CT scanning, HE staining, Masson staining, and immunohistochemical staining. At 8 weeks, the hearts, livers, spleens, lungs, kidneys, and brains were also collected for HE staining.

SPSS 25.0 software and GraphPad Prism8 software were used to analyze and graph the data ( x ¯ ± s ), and the one-way ANOVA method was used for analysis and comparison, while the two-by-two comparisons between groups were made using the LSD-t method, with the test criterion of α = 0.05, and the difference was considered to be statistically significant at P < 0.05.

As illustrated in Table 1, Composite scaffolds (S3) displayed a porosity of 82.19% ± 2.23%, a hydrophilic swelling rate of 926.56% ± 15.29%, and a drug loading efficiency of 2.32% ± 0.10%. These values fell between those of large aperture scaffolds (S1) and small aperture scaffolds (S2) across all three measurements. Pore size is crucial in bone formation, and research indicates that the optimal pore size for bone tissue regeneration ranges between 100 and 400 μm (Hollister, 2005; Murphy et al., 2010). During scaffold preparation, NaCl particles of two different sizes were screened: large particles (250–400 μm) and small particles (100–250 μm) were used to create suitable pore sizes for bone tissue regeneration (Lee et al., 2003). The porosity of the scaffold is recognized as crucial for enhancing cell proliferation and growth. Higher porosity facilitates cell migration and proliferation, providing ample surface area for interactions between cells and the scaffold while promoting the transport of oxygen and nutrients. However, if the porosity is too high, the scaffold may not be able to maintain optimum mechanical stiffness (Wei et al., 2009). Table 1 shows that S1 has the highest porosity, while S2 has the lowest, attributed to the increased size of the porogenic particles. Larger particles result in less PLGA content per unit volume of the scaffold, leading to larger pores upon dissolution of the porogens. However, this also results in lower mechanical strength. In contrast, S3 maintains sufficient mechanical strength while achieving high porosity. A high hydrophilic dissolution rate indicates good hydrophilicity, which enhances the rapid penetration of nutrients (Jin et al., 2021). The drug loading rate indicates the scaffold’s capacity to carry drugs. As shown in Table 1, scaffolds with greater porosity also revealed higher hydrophilic swelling and drug loading rates. This is because increased porosity allows for a greater volume of drug solution to be accommodated. In summary, S3 achieves high porosity, a substantial hydrophilic swelling rate, and an impressive drug loading rate, aligning with our expectations.

The mechanical strength of a scaffold is crucial for withstanding pressure from surrounding soft tissues during bone tissue regeneration. A universal testing machine was utilized to assess the compression strength of the three scaffold types, as illustrated in Figure 3. The highest compression resistance was shown by S2, followed by S3, while the weakest compression resistance was demonstrated by S1. This suggests that S1’s low mechanical strength hampers its ability to maintain space in the bone defect area, which is detrimental to bone repair. In contrast, improved mechanical properties were shown by S3 and S2.

The results of the stent degradation experiments (Figure 4A) indicated that the degradation rate of S3 reached approximately 60% by week 8, while S1 had degraded nearly 90% and S2 showed only about 30% degradation. The findings from the drug release experiments (Figure 4B) demonstrated that the drug release rate is closely correlated with the scaffold degradation rate. S4 revealed a pronounced burst release effect, with nearly 20% of the drug released in the first week and a cumulative release of about 90% by week 8. This rapid release was attributed to S4’s large pore size and high porosity. In contrast, S5 released the drug slowly, with only about 20% released by week 8, while S6 achieved a cumulative release rate of around 60% by that same time. Previous studies have shown that the time required for bone defect repair is approximately 3 months (Saul and Khosla, 2022). The degradation time and drug release profile of PLGA composite scaffolds align well with the timeline of bone defect repair. This ensures that the scaffolding material provides consistent support in the bone defect area throughout the healing process while also degrading appropriately. PLGA composite scaffolds also facilitates prolonged drug release, contributing to sustained therapeutic efficacy.

The FTIR spectra of the β-E/PLGA composite scaffolds, PLGA composite scaffolds, and β-Ecdysone are presented in Figure 5. The absorption peaks of the β-E/PLGA composite scaffolds largely aligned with those of the PLGA composite scaffolds; however, the absorption peak at 3,470 cm⁻¹ for the blank scaffolds shifted to 3,490 cm⁻¹ in the β-E/PLGA composite scaffolds. This shift suggests that the loading of the drug disrupts the hydrogen bonding system within the PLGA. In hydrogen bonding, the O-H bond length increases, leading to lower energy requirements for vibration and, consequently, absorption peaks at lower frequencies. Following the disruption of these hydrogen bonds, the peaks shift back to higher frequencies. Further, the absorption peak at 1,760 cm⁻¹ shifted to 1,750 cm⁻¹, likely due to overlap with the absorption peak at 1,650 cm⁻¹ in the drug. Furthermore, the absorption peak at 1,100 cm⁻¹ shifted to 1,090 cm⁻¹, attributed to the stretching vibration of the alcohol C-O in the drug. These spectral changes indicate successful encapsulation of the drug, demonstrating improved encapsulation efficacy.

The bulk morphology and electron microscope images of composite scaffolds are displayed in Figure 6. Overall, a sandwich-like layered structure is characterized by a large pore size in the middle layer and smaller pores in the outer layers. Through scanning electron microscopy, it was demonstrated that the internal pores of the composite scaffold are interconnected and exhibit both large and small sizes, which is beneficial for drug loading and slow release.

Next, the in vitro biosafety of the PLGA composite scaffolds was tested. The identification results of BMSCs were shown in Supplementary Figure S1. The results of the CCK8 assay (Figure 7A) indicated that the extracts of the composite scaffolds displayed no toxicity to rat bone marrow MSCs when compared to the blank control group, in both low and high-glucose media. The results of the live-dead cell staining assay (Figure 7B) demonstrated that the bone marrow MSCs survived and proliferated on the composite scaffolds in both low and high-glucose environments. These findings suggest that the composite scaffolds possess reliable in vitro safety and cytocompatibility.

In summary, the PLGA composite scaffold was successfully developed that demonstrates high porosity, a significant hydrophilic swelling rate, an effective drug loading capacity, good mechanical properties, and an appropriate degradation time along with favorable slow-release characteristics. In addition, the PLGA composite scaffold also had good biosafety in vitro, making it suitable for further animal experiments.

The animal experiments complied with the ethical requirements of Guangxi Medical University Laboratory Animal Center. To evaluate the effect of drug-carrying composite scaffolds in the treatment of diabetic bone defects, a diabetic rat cranial bone defect model was constructed. Based on the previously established criteria for the establishment of a diabetic rat model (Sun et al., 2022), a successful diabetic model is indicated when fasting blood glucose levels are maintained at ≥11.1 mmol/L for 3–4 weeks. The results (Table 2) showed that the fasting blood glucose levels of the rats in the diabetic group consistently exceeded 11.1 mmol/L for the entire duration of 4 weeks. The weight of the rats in the diabetic group continued to decline compared to their pre-modeling weights, which aligns with the typical symptoms associated with diabetes mellitus (Barragán-Bonilla et al., 2019).

The rat extreme bone defect model is widely recognized for evaluating bone repair in tissue engineering. In this study, extreme bone defects with a diameter of 5 mm were created bilaterally in the skulls of the rats. To assess bone regeneration in the defect areas post-surgery, rat skulls were collected at 4 and 8 weeks for micro-CT scanning. The results (Figure 8A) indicated that new bone formation was significantly reduced in the DM + CS group compared to the N + CS group. Conversely, the DM + CS-βE group exhibited significantly greater new bone formation than the control group.

Bone volume fraction (BV/TV) serves as an indicator of bone quantity in the area of interest, while the number of trabeculae (TbN), trabecular thickness (TbTh), and trabecular separation (TbSp) are key metrics for evaluating the spatial morphology and structure of trabecular bone. High values of BV/TV, TbN, and TbTh indicate substantial new bone formation, while a high TbSp value reflects greater distances between trabeculae and suggests poor bone quality. The findings (Figure 8B) revealed that the DM + CS group had lower BV/TV, TbN, and TbTh, along with higher TbSp compared to the N + CS group. This suggests that the DM + CS group had a reduced amount of new bone and poor bone quality, indicating that the diabetic condition hindered bone defect repair. In contrast, the DM + CS-βE group showed higher BV/TV, TbN, TbTh, and lower TbSp compared to the DM + CS group, indicating that β-Ecdysone enhanced the repair of cranial bone defects in a diabetic model.

To further validate the hypothesis, H&E staining, Masson staining, and immunohistochemical staining of bone tissues from the N + CS, DM + CS, and DM + CS-βE groups was carried out. These analyses provided additional insights into bone tissue regeneration. The H&E staining results (Figure 9A) revealed sparse vascularization in most of the new bone regions in the DM + CS group at both 4 and 8 weeks, indicating that osteogenesis was somewhat inhibited. In contrast, the N + CS and DM + CS-βE groups exhibited a significant presence of new blood vessels in the bone regions, suggesting active osteogenesis. While the amount of neovascularization in the N + CS group was slightly greater than in the DM + CS-βE group, the inclusion of β-Ecdysone contributed to improved bone regeneration in the defect areas of diabetic rats. Masson staining results (Figure 9B) corroborated the findings from the H&E staining, showing more neovascularized bone tissue and abundant blood vessels in the N + CS and DM + CS-βE groups. Conversely, the DM + CS group displayed low levels of neovascularization and sparse vascularization. RUNX2, a member of the RUNX family of transcription factors, plays a crucial role in osteoblast differentiation and bone morphogenesis and is specifically expressed in osteoblasts, where it is localized in the nucleus (Komori, 2019). Immunohistochemical staining results (Figures 9C, D) showed that RUNX2 protein expression was significantly higher in the N + CS group compared to the DM + CS group (indicated in blue). Similarly, RUNX2 protein expression was substantially higher in the DM + CS-βE group than in the DM + CS group. These findings suggest that β-Ecdysone mitigates the inhibitory effects of diabetes on bone defect repair and promotes the healing of diabetic bone defects.

The H&E staining results of heart, liver, spleen, lung, kidney, and brain tissues from rats in each group at 4 weeks (Supplementary Figure S2) and 8 weeks (Figure 10) were compared. The visceral tissues of rats in the N + CS, DM + CS, and DM + CS-βE groups did not exhibit any significant pathological changes compared to the Blank group. This indicated that the composite scaffolds did not cause any toxic damage to the vital organs in vivo, thereby confirming the in vivo safety of the composite scaffolds.