GO powder (Cat. No. NM000060, Solarbio, Beijing, China) was dispersed in deionized water via ultrasonication to obtain a homogeneous suspension. The suspension was subjected to carboxylation at 4 °C, followed by repeated centrifugation and washing until the pH stabilized at 7.0–7.5. Subsequently, the GO suspension was mixed with pre-dissolved 6-NH₂-PEG solution and further ultrasonicated. To promote covalent conjugation, 1.0 mg/mL of 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) was added, and the reaction mixture was stirred at room temperature to obtain the GO-PEG composite solution.

Functional groups were analyzed by Fourier transform infrared spectroscopy (FTIR; Thermo Electron Corporation, United States) over a spectral range of 4,000–400 cm⁻¹. Microstructural morphology was examined using transmission electron microscopy (TEM; Thermo Electron Corporation, United States) at 16,500 × magnification. Elemental composition was determined using an organic elemental analyzer (vario MICRO cube, Elementar, Germany) in CHN mode. Thermogravimetric analysis (TGA; STA 449G, Germany) was performed from room temperature to 500 °C at a heating rate of 15 °C/min.

The pBMP-2 plasmid (Cat. No. P40382, MiaoLingBio, Wuhan, China) was dissolved in sterile water according to the manufacturer’s instructions and mixed with GO-PEG solution at nitrogen-to-phosphate (N/P) ratios of 10, 20, and 40. The mixtures were incubated at room temperature to form GO-PEG/pBMP-2 complexes.

Plasmid loading efficiency was determined by measuring unbound pBMP-2 in the supernatant after washing. The concentration of free plasmid was quantified using a dsDNA fluorescence assay kit (Cat. No. SY0259, Bio-Lab, Beijing, China) following the manufacturer’s protocol. Fluorescence intensity was measured using a fluorescence spectrophotometer (RF-1501, Shimadzu, Japan), and concentrations were calculated from a standard curve.

PEEK disks (Victrex plc, Thornton Cleveleys, United Kingdom) were immersed in 98% sulfuric acid at room temperature and sulfonated under magnetic stirring (100 rpm) for 5 min (22). Samples were immediately removed and ultrasonically cleaned alternately in deionized water and acetone to remove residual acid (23). After sterilization and drying, sulfonated PEEK (SPEEK) was obtained.

For in vitro release study, SPEEK-(GO-PEG/pBMP-2) samples were immersed in 1 mL of PBS (pH 7.4) at 37 °C with shaking (100 rpm). At predetermined time points (1, 3, 5, 7, 10, 14, 21, 28, and 35 days), the entire medium was collected and replaced with fresh PBS. Samples were stored at −20 °C until analysis. The amount of released pBMP-2 was quantified using the same dsDNA quantitation kit described above, with concentrations calculated from a standard curve prepared using known amounts of pBMP-2.

Surface morphology was observed using field emission scanning electron microscopy (SEM, JSM-7500F, JEOL, Japan). Surface hydrophilicity was assessed by measuring the static water contact angle (CA, JY-82B, Kruss DSA, Germany) under ambient temperature and humidity.

A total of 20 male Sprague–Dawley (SD) rats (6-week-old, 135–150 g), provided by the Animal Research Center of Xinjiang Medical University, were used for the isolation of bone marrow–derived mesenchymal stem cells. rBMSCs were isolated from each individual rat and cultured independently without pooling, thereby preserving distinct biological origins. All SD rats were fed with redistilled water and sterilized food and housed in a pathogen-freeroom with environmental conditions of good ventilation, 55 + 5%humidity, 20 + 2 °C temperature, and proper light intensity (4 W/m²) for 12 h per day. Six weeks later, all rats were killed using a rat euthanasia box filled with carbon dioxide (LC-800-S1, Shanghai Yuyan Scientific Instrument, China), and their femurs and tibias were harvested together with surrounding soft tissues. Both ends of each bone were removed to expose the medullary cavity, and bone marrow was aspirated using complete culture medium. The marrow was repeatedly flushed to ensure maximal cell recovery. The collected suspension was centrifuged at high speed, the supernatant discarded, and the pellet resuspended in fresh medium. Cells were cultured in a humidified incubator at constant temperature, with medium replaced on the third day after primary isolation. Experiments were performed using cells at the third passage. All sections of this protocol are conformed to the Animal Research Reporting in Vivo Experiments (ARRlVE) guidelines and approved by the Ethics Committee at the First Affiliated Hospital of Xinjiang Medical University (approval no. IACUC-JT-20230515-09).

Rat BMSCs were seeded onto different material samples in 24-well plates. After 3 days of culture, cells were fixed with 3% glutaraldehyde for 3 h, dehydrated through a graded ethanol series, and dried. Cell adhesion and morphology on the material surfaces were examined by field emission scanning electron microscopy (FE-SEM, JSM-7500F, JEOL, Japan).

The cytoskeleton of rBMSCs was visualized using confocal laser scanning microscopy (CLSM, Nikon C2, Japan). Cells were permeabilized with 0.1% Triton X-100 (Cat. No. IR9073, Solarbio, Beijing, China) for 5 min, followed by staining with rhodamine–phalloidin (Cat. No. CA1620, Solarbio, Beijing, China) for F-actin and 4′,6-diamidino-2-phenylindole (DAPI; Cat. No. C0065, Solarbio, Beijing, China) for nuclei.

Cell viability was evaluated using the Cell Counting Kit-8 (CCK-8Cat. No. C0038, Beyotime, Shanghai, China). At days 1, 3, 5, and 7 of culture, a mixture of serum-free medium and CCK-8 reagent (10:1) was added to each well of 96-well plates. After incubation, the optical density (OD) was measured at 450 nm using a microplate reader.

Cytotoxicity was further assessed by Live/Dead staining. After 3 days of culture, 250 μL of a Calcein AM/PI working solution (Cat. No. C1371M, Beyotime, Shanghai, China) was added to the cells and incubated at 37 °C in the dark for 30 min. Stained cells were then visualized using a laser confocal microscope.

The osteogenic differentiation of rBMSCs on different materials was evaluated by alkaline phosphatase (ALP) staining and activity assays. Cells were cultured in osteogenic induction medium (Cat. No. PD-008, Procell, Wuhan, China) for 7 and 14 days, respectively. After culture, cells were fixed with 4% paraformaldehyde and stained with a BCIP/NBT staining solution (Cat. No. C3206, Beyotime, Shanghai, China). Following incubation in the dark, stained samples were imaged and analyzed using a stereomicroscope.

Cells were lysed by adding 1 mL of 1% Triton X-100 solution to each well and gently pipetting. After centrifugation, the supernatant was collected, and ALP activity in the lysates was determined using an ALP activity assay kit (Cat. No. P0321S, Beyotime, Shanghai, China) according to the manufacturer’s instructions. Absorbance was measured at 405 nm with a microplate reader.

The osteogenic potential of the materials was further evaluated by mineralized nodule staining and quantitative analysis. Cells were cultured in osteogenic induction medium for 14 and 21 days, respectively. After culture, cells were fixed with 4% paraformaldehyde and stained with an Alizarin Red S solution (Cat. No. C0148S, Beyotime, Shanghai, China) for 30 min, followed by thorough rinsing. Extracellular matrix mineralization nodules were imaged using a stereomicroscope. For quantification, a 10% cetylpyridinium chloride solution (Cat. No. Y175536, Beyotime, Shanghai, China) was added to dissolve the bound dye. After incubation for 1 h and gentle mixing, appropriate volumes of the eluates were transferred to a 96-well plate, and absorbance was measured at 562 nm with a microplate reader.

The expression of osteogenesis-related genes, including runt-related transcription factor 2 (Runx2), ALP, type I collagen (COL-1), and osteopontin (OPN), was analyzed by quantitative real-time PCR (qRT-PCR), with GAPDH serving as the housekeeping gene. After 7 days of culture in osteogenic induction medium, total RNA was extracted from each group using Trizol reagent and reverse-transcribed into cDNA. PCR amplification was conducted under the following conditions: pre-denaturation at 95 °C for 3 min; denaturation at 95 °C for 10 s; annealing/extension at 60 °C for 25 s; 40 cycles in total. Gene expression levels were quantified using a qRT-PCR kit with gene-specific primers (sequences listed separately). Relative expression was calculated by the 2^(-ΔΔCt) method, and statistical analysis was performed. All experiments were performed with three independent biological replicates (n = 3) (Table 1).

The antibacterial activity of the materials was evaluated using a dilution plating method. Four groups of samples—PEEK, SPEEK, SPEEK-GO-PEG, and SPEEK-(GO-PEG/pBMP-2)—were placed in 12-well plates. Bacterial suspensions of Escherichia coli and Staphylococcus aureus were adjusted to an OD₆₀₀ of 0.15, and 1 mL of each suspension was added to the wells for co-culture at 37 °C for 24 h. After incubation, the suspensions were removed, diluted to 10⁶ CFU/mL in LB broth, and 30 μL of each dilution was spread evenly onto LB agar plates using sterile glass beads. Plates were incubated at 37 °C for 24 h, and the number of colonies was counted.

Simultaneously, bacterial adhesion and morphology on the material surfaces were examined by scanning electron microscopy (SEM). Samples from each group were incubated with 1 mL of E. coli or S. aureus suspension at 37 °C for 24 h to allow biofilm formation. After incubation, samples were washed with PBS, fixed, dehydrated as described previously, dried at room temperature, and sputter-coated with gold. Bacterial morphology on the sample surfaces was observed under SEM at a magnification of 5,000 × .

Statistical analyses were performed using SPSS 26.0 and GraphPad Prism 9. All experiments were performed with three independent biological replicates (n = 3). Data are expressed as mean ± standard deviation (SD) for continuous variables with normal distribution and homogeneity of variance. Comparisons between two groups were conducted using independent-samples t tests, while comparisons among multiple groups were assessed by one-way analysis of variance (ANOVA) followed by least significant difference (LSD-t) post-hoc tests. A p < 0.05 was considered statistically significant.

FTIR confirmed the functionalization of GO with PEG. In the infrared spectrum, characteristic absorption peaks were observed at 2,800–2,900 cm⁻¹ (C–H stretching), 1,736 cm⁻¹ (C=O stretching), and 3,427 cm⁻¹ (O–H stretching). Compared with GO, GO-PEG exhibited marked spectral changes, most notably a pronounced increase in the C–H absorption band at ~2,800 cm⁻¹, indicating successful PEG modification of the GO surface (Figure 1A).

TEM revealed distinct morphological differences between GO and GO-PEG. GO exhibited an irregular polygonal shape with a multilayered, wrinkled structure. Following PEG modification, the nanosheets displayed a more circular morphology and a markedly reduced particle size, confirming substantial structural changes (Figure 1B).

The PEG loading on the GO surface was quantified by combining TGA with EA. TGA revealed the temperature-dependent mass loss profiles of GO and GO-PEG (Figure 1C). EA showed the elemental compositions of C, H, and N for both groups (Figure 1D). Notably, the nitrogen content increased from 0.04% in GO to 5.82% in GO-PEG, corresponding to a PEG loading of approximately 16.8%.

The transfection efficiency of pBMP-2 at different N/P ratios (10, 20, 40) was evaluated by qRT-PCR. pBMP-2 effectively transfected cells and enhanced BMP-2 mRNA expression across all N/P ratios tested. Expression levels increased progressively from N/P 10 to N/P 20, indicating improved transfection efficiency (p < 0.001). Compared with the positive control (lipo8000/pBMP-2), the N/P 20 group showed slightly lower efficiency (p < 0.01). Nevertheless, BMP-2 mRNA expression in all pBMP-2 groups was significantly higher than in the untreated control (p < 0.0001) (Figure 2A).

The cytocompatibility of GO-PEG/pBMP-2 transfection complexes at different N/P ratios was assessed by CCK-8 assay. On Day 1, no significant differences in cell viability were observed between the N/P 10 and N/P 20 groups compared with the control, and no statistical difference was detected between these two groups. In contrast, viability in the N/P 40 group was significantly reduced (p < 0.01). By Day 3, cell viability in the N/P 10 and N/P 20 groups increased relative to the control (p < 0.01), with no significant difference between them, whereas the N/P 40 group remained markedly reduced (Figure 2B). These results suggest that N/P ratios of 10–20 provide favorable biocompatibility, while higher ratios substantially impair cell viability. Accordingly, the N/P 20 group was selected for subsequent experiments.

In this study, the plasmid loading efficiency of the SPEEK-(GO-PEG/pBMP-2) composite was 1.4% ± 0.5%. The in vitro release profile of pBMP-2 showed an initial burst release, followed by a sustained and controlled release. The cumulative release reached approximately 79% by day 30, indicating that the SPEEK-(GO-PEG/pBMP-2) composite successfully achieved long-term sustained release of pBMP-2 from a PEEK-based implant system (Supplementary Figure S1).

SEM was used to examine the surface morphology of samples in each group. PEEK exhibited a relatively flat and smooth surface with only minor polishing scratches, whereas sulfonated SPEEK displayed a distinct three-dimensional porous structure. After coating with GO-PEG or GO-PEG/pBMP-2, the porous structure was partially covered, and layered functionalized graphene oxide was observed on both surfaces, with no marked differences between the two groups (Figure 2C).

The surface hydrophilicity of each group was evaluated by water CA measurements. Untreated PEEK exhibited the largest contact angle, indicating strong hydrophobicity. In contrast, sulfonated PEEK showed a markedly reduced contact angle (p < 0.0001), reflecting enhanced hydrophilicity. Further coating with GO-PEG or GO-PEG/pBMP-2 decreased the contact angle even more (p < 0.0001), demonstrating superior hydrophilicity, although no significant difference was detected between the two groups (Figures 2D,E).

SEM observation of rBMSCs cultured on different material surfaces revealed distinct morphological differences. On PEEK, cells exhibited a slightly elongated spindle shape with few peripheral filopodia. In contrast, cells on SPEEK displayed pronounced spreading, with filopodia extending from the cell edges and anchoring to the porous surface. On SPEEK-GO-PEG and SPEEK-(GO-PEG/pBMP-2), cells demonstrated extensive spreading with lamellipodia formation, larger cell areas, and strong adhesion to the material surfaces (Figure 3A).

Fluorescent staining of rBMSCs was performed with Rhodamine–Phalloidin (red) for F-actin and DAPI (blue) for nuclei to evaluate cell adhesion on different material surfaces. On PEEK and SPEEK, cells predominantly displayed a spindle-like morphology with visible F-actin filaments arranged in a relatively disordered pattern. In contrast, cells on SPEEK-GO-PEG and SPEEK-(GO-PEG/pBMP-2) exhibited well-spread, polygonal morphologies. F-actin expression was markedly increased and organized into a regular network aligned with the direction of cell extension (Figure 3B).

Cell proliferation on different materials was evaluated by CCK-8 assay at days 1, 3, 5, and 7. At all-time points, proliferation in the SPEEK, SPEEK-GO-PEG, and SPEEK-(GO-PEG/pBMP-2) groups was significantly higher than that in the PEEK control (p < 0.05). No significant differences were observed among the three modified groups on day 1. From day 3 onward, proliferation in the SPEEK-GO-PEG and SPEEK-(GO-PEG/pBMP-2) groups was markedly greater than in the SPEEK group. Moreover, at days 3 and 5, cell numbers in the SPEEK-(GO-PEG/pBMP-2) group exceeded those in the SPEEK-GO-PEG group (p < 0.01), indicating superior cell viability (Figure 3C).

Cytotoxicity was further evaluated by live/dead staining. In the PEEK group, abundant PI-positive cells were observed, indicating higher cytotoxicity. The SPEEK group exhibited fewer dead cells (p < 0.05). In contrast, the SPEEK-GO-PEG and SPEEK-(GO-PEG/pBMP-2) groups showed negligible PI staining, demonstrating markedly reduced cytotoxicity (p < 0.001) (Figures 3D,E).

After 7 and 14 days of osteogenic induction, ALP staining revealed blue–purple signals in all groups, with intensity increasing over time. At both time points, the SPEEK-(GO-PEG/pBMP-2) group displayed markedly stronger staining than the other groups, indicating the highest ALP expression (Figure 4A). Consistently, ALP activity assays performed at days 1, 4, 7, 10, and 14 showed no significant differences among groups on day 1. From day 4 onward, ALP activity in the SPEEK-(GO-PEG/pBMP-2) group was significantly higher than in all other groups (p < 0.001), confirming superior osteogenic activity (Figure 4B).

After 14 and 21 days of osteogenic induction, alizarin red staining revealed orange–red mineralized nodules in all groups, with staining intensity increasing over time. At both time points, the SPEEK-(GO-PEG/pBMP-2) group exhibited markedly stronger staining and a greater number of mineralized nodules than the other groups. Semi-quantitative analysis confirmed these findings, showing significantly higher mineral deposition in the SPEEK-(GO-PEG/pBMP-2) group (p < 0.05) (Figures 4C,D).

The osteoinductive effects of the composites were further assessed by examining the expression of osteogenesis-related genes, including COL-1, ALP, Runx2, and OPN, using qRT-PCR, with expression levels normalized to GAPDH. Gene expression profiles across the four material groups were consistent with the osteogenic staining results. Notably, the SPEEK-(GO-PEG/pBMP-2) group exhibited significantly higher expression of COL-1, ALP, Runx2, and OPN compared with the other groups (p < 0.05) (Figures 4E–H).

The antibacterial properties of the materials against S. aureus and E. coli were assessed using the plate count method. Compared with PEEK, all modified groups exhibited enhanced antibacterial activity. Both SPEEK-GO-PEG and SPEEK-(GO-PEG/pBMP-2) demonstrated stronger antibacterial effects than SPEEK, with no significant difference between the two groups (p < 0.01) (Figures 5A–J).

SEM was used to further evaluate bacterial morphology on the surfaces of different materials. For S. aureus, bacteria on PEEK exhibited intact spherical morphology, with many in the division phase and densely distributed across the surface, indicating active proliferation. On SPEEK, bacterial numbers decreased, but morphology remained largely unchanged, and distribution was irregular due to the porous structure. In contrast, SPEEK-GO-PEG and SPEEK-(GO-PEG/pBMP-2) surfaces showed markedly reduced bacterial counts. Most bacteria displayed morphological alterations, including wrinkled envelopes, with partial membrane rupture observed in some cells. For E. coli, bacteria on PEEK exhibited intact rod-shaped morphology with smooth edges and dense surface coverage. On SPEEK, bacterial numbers decreased without major morphological changes. On SPEEK-GO-PEG and SPEEK-(GO-PEG/pBMP-2), bacteria exhibited visible membrane indentations, irregular edges, and surface wrinkling (Figures 5K,L).