Soybean lecithin, cholesterol, octadecylamine and polyethylenimine (PEI, 25 kDa, branched) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Bombyx mori silkworm cocoons were sourced locally. Poly (ethylene oxide) (PEO, Mw = 900 kDa) was purchased from Aladdin Industrial Corporation (Shanghai, China). The pIRES2-ZsGreen1-BMP2 plasmid, containing the gene for BMP2 and a green fluorescent protein (GFP) reporter, was constructed and preserved by our laboratory. Dulbecco’s Modified Eagle Medium (DMEM), fetal bovine serum (FBS), and penicillin-streptomycin were purchased from Gibco (Grand Island, NY, USA). 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), 3,3′-dioctadecyloxacarbocyanine perchlorate (DiO), Alizarin Red S (ARS), and the alkaline phosphatase (ALP) activity assay kit were obtained from Beyotime Biotechnology (Shanghai, China). All other chemical reagents were of analytical grade and used as received. Rat BMSCs present in this study were obtained from Cell Bank and Stem Cell Bank of Chinese Academy of Sciences.

Cationic Lipo^PEI were prepared using a thin-film hydration method. Briefly, soybean lecithin (0.1 g), cholesterol (0.01 g), and Octadecylamine (0.004 g) were dissolved in ethanol in a round-bottom flask. The organic solvent was evaporated under reduced pressure to form a thin lipid film, which was then hydrated with 10 mL PEI aqueous solution (0.001 g/mL) and the Lipo^PEI suspension was obtained after stirring 30 min at room temperature. The suspension was sonicated and repeatedly extruded through a 100 nm polycarbonate membrane to obtain uniformly sized Lipo^PEI.

For the preparation of DNA-loaded Lipo^PEI, the plasmid pIRES2-ZsGreen1-BMP2 (pDNA _Bmp2 ) was mixed with the Lipo^PEI suspension at various mass ratios (Lipo^PEI: pDNA _Bmp2 = 20 : 1, 40 : 1, and 60 : 1) and incubated at room temperature to allow for complex formation. The particle size, polydispersity index (PDI), and zeta potential of the nanoparticles were measured by Dynamic Light Scattering (DLS) using a Zetasizer Nano ZS (Malvern Instruments, UK). The morphology of the nanoparticles was observed by Transmission Electron Microscopy (TEM, JEM-2100, JEOL, Japan) after negative staining with 4% phosphotungstic acid. The encapsulation efficiency (EE) of the plasmid DNA was determined indirectly. The amount of unencapsulated DNA in the supernatant after centrifugation of the complex was quantified using a UV-Vis spectrophotometer. The EE was calculated using the formula: EE (%) = (Total DNA - Free DNA)/Total DNA × 100%.

The silk fibroin (SF) solution was prepared as previously described. Briefly, cocoons were degummed, dissolved in 9 M LiBr solution, and dialyzed against deionized water to obtain a purified SF aqueous solution. For electrospinning, PEO was added to the SF solution at 10% of the SF’s mass to improve spinnability. The optimized pDNA _Bmp2 @Lipo^PEI complexes were then added to the SF/PEO solution at different concentrations (mass percentage of Lipo^PEI relative to SF was set at 0.2%, 0.4%, and 0.6% respectively, while the amount of pDNA _Bmp2 was fixed to 60 μg) and gently mixed to form a homogeneous spinning dope (Table 1). The solution was loaded into a syringe and electrospun using a standard electrospinning setup. The process parameters were optimized and set as follows: applied voltage of 10–13 kV, a tip-to-collector distance of 15–20 cm, and a solution flow rate of 0.3–0.5 mL/h. The nanofibers were collected on a grounded aluminum foil. After spinning, the scaffolds were treated with 75% ethanol vapor for 24 h to induce a conformational transition to the water-insoluble β-sheet structure. Four groups of scaffolds were fabricated: SF-PEO, Lipo^PEI@SF-PEO (empty vector), pDNA _Bmp2 @SF-PEO (scaffold with naked plasmid), and pDNA _Bmp2 @Lipo^PEI@SF-PEO.

The surface morphology and fiber diameter of the scaffolds were examined using Scanning Electron Microscopy (SEM, S-4800, Hitachi, Japan) after sputter-coating with gold. The distribution of the Lipo^PEI within the nanofibers was observed using a fluorescent microscope (DMi 8, Leica Microsystems Ltd., Wetzlar, Germany) by labeling the liposomes with DiO. The chemical structure and conformation of the SF within the nanofibers were analyzed by Fourier Transform Infrared Spectroscopy (FTIR, Nicolet 6700, Thermo Fisher Scientific, USA). The crystalline structure of the scaffolds was evaluated by X-ray Diffraction (XRD, D/Max-2550 PC, RIGAKU, Japan).

The release profile of the plasmid DNA from the scaffolds was evaluated in vitro. A certain mass of electrospun fibrous scaffolds (3 parallel samples in each group) was immersed in PBS buffer and placed in an oscillating shaker (temperature maintained at 37 °C and oscillation frequency at 100 rpm). Remove 1 mL of release solution every 24 h while supplementing with fresh equal volume of PBS. The absorbance value of the samples is measured by a UV spectrophotometer (NanoDrop 2000C) at 260 nm, and the average value of each group of samples is calculated according to the following formula, and the sustained-release curve can be obtained with the percentage of each release to the actual drug load as the ordinate and the sustained-release time as the abscissa. E r = V e ∑ i n − 1 C i + V 0 C n m d r u g

(Er: cumulative DNA release, %; V_e: Displacement volume of extended-release solution, 1,000 μL; V₀: Volume of release solution (μL); C_i: Concentration of DNA in the release solution at the time of the first displacement sampling (ng/μL); m_drug: total mass of DNA used for release (ng); n: Number of times the sustained-release solution was replaced).

Rat BMSCs were used for all cellular experiments. The cells were cultured in DMEM-F12 supplemented with 10% FBS and 1% penicillin-streptomycin in a humidified incubator at 37 °C with 5% CO². The culture medium was changed every 2–3 days. Cells from passages three to five were used for the experiments.

The biocompatibility of the scaffolds was assessed using the MTT assay. BMSCs were seeded onto the different scaffolds placed in a 24-well plate at a density of 5 × 10⁴ cells/well. Cells cultured on glass coverslips served as a control. After 1, 3, 5, and 7 days of culture, the MTT reagent was added to each well and incubated. The resulting formazan crystals were dissolved in DMSO, and the absorbance was measured at 492 nm using a microplate reader (MK3, Thermo, USA). Cell attachment and morphology on the scaffolds were observed by SEM after 3 days of culture. The cell-scaffold constructs were fixed, dehydrated, and prepared for SEM imaging.

To evaluate the gene transfection efficiency, BMSCs were cultured on the pDNA _Bmp2 @Lipo^PEI@SF-PEO scaffolds containing the GFP reporter gene. After 3 days of culture, the expression of GFP in the cells was observed using CLSM. Cells cultured on glass with free pDNA _Bmp2 @Lipo^PEI complexes served as a positive control. The percentage of GFP-positive cells was estimated to determine the transfection efficiency.

BMSCs were seeded on the scaffolds and cultured in an osteogenic induction medium. The osteogenic potential was evaluated by measuring ALP activity and calcium deposition.

ALP Activity Assay: After 7, 14, and 21 days of culture, the cells on the scaffolds were lysed with Triton X-100. The ALP activity in the cell lysate was measured using an ALP activity assay kit according to the manufacturer’s instructions. The total protein content was quantified using a BCA protein assay kit, and the ALP activity was normalized to the total protein content.

Alizarin Red S (ARS) Staining: After 14 and 21 days of culture, the cell-scaffold constructs were fixed and stained with 2% ARS solution (pH 4.1–4.3) for 30 min to visualize calcium deposits. For quantitative analysis, the stained calcium nodules were destained with 10% cetylpyridinium chloride, and the absorbance of the extracted solution was measured at 570 nm.

All quantitative data were presented as mean ± standard deviation (SD) from at least three independent experiments (n = 3). Statistical analysis was performed using one-way analysis of variance (ANOVA) followed by Tukey’s post hoc test for multiple comparisons. A p-value of less than 0.05 was considered statistically significant (*p < 0.05).

The physicochemical properties of the Lipo^PEI are crucial for successful gene delivery. As determined by DLS, the empty Lipo^PEI exhibited an average particle size of 115 ± 5 nm with a low PDI of 0.18 ± 0.02, indicating a uniform and monodisperse distribution (Figure 1A). The surface charge was strongly positive, with a zeta potential of +24.5 ± 2.1 mV, which is advantageous for binding negatively charged DNA (Figure 1B). Upon complexing with the pDNA _Bmp2 at the optimized mass ratio of 40:1, the resulting pDNA _Bmp2 @Lipo^PEI nanoparticles showed an increase in average size to 266 ± 12 nm (Figure 1A), while the PDI remained low at 0.22 ± 0.03. The zeta potential decreased to +16.0 ± 1.8 mV (Figure 1B), confirming the successful binding of the anionic plasmid DNA to the cationic carrier surface. TEM imaging revealed that the nanoparticles were spherical and well-dispersed (Figure 1C). The encapsulation efficiency was calculated to be 66.23% ± 1.67%, demonstrating a competent capacity for gene loading.

SEM analysis was performed to evaluate the morphology of the electrospun scaffolds. As shown in Figure 2, all scaffold groups displayed a continuous, smooth, and ribbon-like nanofibrous structure, mimicking the native ECM. The pure SF scaffold had an average fiber diameter of 460.9 ± 45.2 nm. The incorporation of pDNA _Bmp2 @Lipo^PEI resulted in a slight, concentration-dependent increase in fiber diameter, with the pDNA _Bmp2 @Lipo^PEI (0.6%)@SF-PEO group reaching a diameter of 540.8 ± 55.3 nm (Figure 2A). This increase suggests the successful encapsulation of the nanoparticles within the fibers without disrupting the overall nanofibrous architecture. Microscopy of DiO-labeled liposomes confirmed their uniform distribution throughout the fiber matrix (Figure 2B). These composite nanofibrous scaffolds were treated with 75% ethanol vapor to obtain good hydration resistance and maintained their intact fiber structure after 7 days of soaking in PBS (Figure 2C). The structural changes in the SF protein upon ethanol treatment were assessed by FTIR and XRD. The FTIR spectrum of the as-spun fibers showed characteristic peaks for Amide I at 1,652 cm^-1 (random coil), which shifted to 1,627 cm^-1 (β-sheet) after ethanol treatment (Figure 2D). Correspondingly, the Amide II peak shifted from 1,533 cm^-1 to 1,525 cm^-1 (Figure 2D). These results confirm a successful conformational transition from a soluble random coil/α-helix state to a water-insoluble, stable β-sheet structure. XRD analysis supported this finding, showing a broad amorphous peak in the as-spun fibers that transformed into a sharper, more defined peak indicative of increased crystallinity after treatment (Figure 2E).

The release kinetics of the pDNA _Bmp2 from the scaffolds were monitored over 2 weeks. As shown in Figure 3, DNA release from the nanofibrous scaffolds with a rapid rate within the first 100 h and then gradually flattening. Obviously, the free DNA showed a faster release rate, and was almost fully released in 2 weeks. In contrast, the DNA encapsulated by Lipo^PEI was released more slowly, with only about 80% released in 2 weeks. This may be due to the adsorption of DNA by Lipo^PEI, which delays the release of DNA, resulting in a long-term sustained release. The sustained release pattern is highly desirable for long-term gene expression and tissue regeneration.

The cytocompatibility of the scaffolds is essential for their application in tissue engineering. As shown in Figure 4A, BMSCs cultured on all scaffold groups exhibited robust proliferation over 7 days. Notably, the proliferation on the composite nanofibrous scaffolds was significantly higher than on the glass control after 3 days (p < 0.05), indicating that the scaffold materials and the incorporated gene vectors possess excellent biocompatibility and are conducive to cell growth. SEM imaging after 3 days of culture confirmed these findings. BMSCs were observed to adhere and spread well on the surfaces of all nanofibrous scaffolds. The cells displayed a healthy, flattened polygonal morphology with numerous filopodia extensions, indicating strong cell-matrix interactions (Figure 3B). The 3D porous structure of the scaffolds allowed for some cells to begin migrating into the scaffold interior, a promising sign for eventual tissue ingrowth.

The ability of the gene-activated scaffolds to transfect BMSCs was evaluated by observing the expression of the GFP reporter gene. After 7 days of culture on the pDNA _Bmp2 @Lipo^PEI@SF-PEO scaffold, strong green fluorescence was observed in cells via confocal microscopy (Figure 5E), which was comparable to the positive control where cells were cultured on glass and treated with free pDNA _Bmp2 @Lipo^PEI (Figure 5F). In contrast, no fluorescence was detected in the negative control groups (cells cultured on scaffolds with no pDNA _Bmp2 or with naked pDNA _Bmp2 ), confirming that the Lipo^PEI was essential for successful gene delivery and expression.

To assess the ultimate biological function of the gene-activated scaffold, its ability to induce osteogenic differentiation of BMSCs was evaluated.

ALP Activity: ALP is an early marker of osteoblastic differentiation. As shown in Figure 6A, all groups supported a basal level of ALP activity. However, the pDNA _Bmp2 @Lipo^PEI@SF-PEO group exhibited a significantly higher ALP activity at day 21 compared to all control groups (glass, SF-PEO, Lipo^PEI@SF-PEO, glass + pDNA _Bmp2 ). The ALP activity in the gene-activated scaffold group was significantly higher than that of the pure SF scaffold and other control groups (p < 0.05), demonstrating a potent and sustained osteoinductive effect driven by the expression of BMP2.

Calcium Deposition: Alizarin Red S staining was used to evaluate calcium mineralization, a hallmark of late-stage osteogenic differentiation. After 21 days, intense red staining was observed in the pDNA _Bmp2 @Lipo^PEI@SF-PEO groups, indicating extensive mineralized nodule formation. This result strongly corroborates the ALP activity data and confirms that the sustained delivery and expression of the BMP2 gene from the green-electrospun scaffold effectively promotes the complete osteogenic differentiation of BMSCs.