The system was prepared as described previously (16). Briefly, IC (National Institute for the Control of Pharmaceuticals and Biological Products, Beijing, China) was dissolved into ethanol to prepare an IC solution at a concentration of 8 mg/ml. A solution of vancomycin hydrochloride (VA; molecular formula, C₆₆H₇₅C_l2N₉O₂₄·HCl; molecular weight, 1485.71 g/mol; Sigma, Shanghai, China) was also prepared by dissolving VA in phosphate-buffered solution (PBS, pH 8.0) at a concentration of 80 mg/ml. The two solutions were then mixed at a ratio of 1:1 (v/v). The IC-VA mixture was then stirred for 2 h, followed by the slow addition of CPC powder (Shanghai Rebone Biomaterials Co., Ltd., Shanghai, China) to prepare an IC-VA/CPC paste precursor at a liquid/powder ratio of 1:1 (ml/g). Subsequently, the precursor was homogenized by 4 h vigorous stirring. The resulting solution was cast by a glass mold (4 mm in diameter, 15 mm in length) and placed at 4°C for 6 h, −10°C for 3 h, then freeze-dried to obtain IC-VA/CPC cylinders. Each cylinder contained ~2 mg of IC and 20 mg of VA. We also prepared a cylinder of IC/CPC or VA/CPC at the same concentration of VA or IC for the animal experiments. A pure CPC cylinder was used as a control in the animal experiment and biocompatibility tests. All cylinders were sterilized with 20 kGy ⁶⁰Co and stored in vacuum packages at room temperature prior to subsequent use.

The microstructure of the cylinder samples was examined using a scanning electron microscope (SEM, S-3000N; Hitachi, Japan) after the samples were sputter-coated with gold under a vacuum.

Biocompatibility was determined by evaluating the toxicity of sample extracts (according to ISO10993-5) and the survival of cells seeded directly onto the sample surfaces (17). Briefly, sample extracts of IC-VA/CPC and CPC were created in Dulbecco’s modified Eagle’s medium (DMEM; Sigma) without fetal calf serum (FCS) at a concentration of 100 mg/ml (extract/DMEM culture medium) for 24 h at 37°C in a 5% CO₂ atmosphere, following the guidelines of ISO standard 10993-12. Balb/c 3T3 cells (American Type Culture Collection, Manassas, VA, USA) were cultured in DMEM containing ampicillin (0.025 g/l) and streptomycin (0.1 g/l) supplemented with 10% FCS (Sigma) at 37°C in a 5% CO₂ atmosphere. The tests were performed using 24-well plates, by plating 3T3 cells at a density of 3×10⁴ cells/well. Cells plated at the same density in wells without any extracts were used as blank controls. The total cell number of the four groups at days 1, 3, 5 and 7 was estimated by quantifying the double-stranded DNA (dsDNA) content of each sample using a PicoGreen dsDNA Quantification kit (Molecular Probes, Eugene, OR, USA). The total dsDNA was extracted by enzymatic digestion and assayed according to the manufacturer’s instructions. The proliferation of the cells was interpreted by changes in dsDNA quantity. In order to observe the morphology of cells cultivated on the surface of granules, 3T3 cells were plated onto the samples at a density of 5×10⁴ cells/sample. At 24 and 72 h after plating, the cells were fixed with modified Karnovsky’s solution and post-fixed with 1% osmium tetroxide in 0.05 M cacodylate buffer (pH 7.2) for 1 h at room temperature. The samples were washed with distilled water, dehydrated using increasing solutions of ethanol and critical point drying, metallized with gold and analyzed under the SEM.

The release behaviors of IC and VA from the IC-VA/CPC system were measured by high-performance liquid chromatography (HPLC) with UV detection at 230 nm (Agilent1200; Agilent Technologies, Santa Clara, CA, USA) according a previous method with modifications (18). Briefly, samples of IC-VA/CPC were soaked in 5 ml PBS (pH 7.4, 37°C) and agitated gently at 10 rpm. At 0, 1, 3, 5, 7, 10, 15, 20, 25 and 30 days, 5 ml PBS was collected (stored at 4°C for HPLC examination) and replaced by adding the same amount of fresh PBS. For analysis of IC concentration, the samples were centrifuged (1,000 × g) for 10 min and 0.5 ml supernatants were isolated. After the addition of 0.5 ml acetonitrile, the mixed solution was re-centrifuged (4,000 × g) for 10 min and 20 μl supernatants were applied for HPLC analysis. Peak area data was calculated by integration using Empower 2 software (Waters Co., Milford, MA, USA), which identifies the absorbance peaks of different proteins and calculates the area of the peak, therefore determining the concentration of the proteins. IC or VA released from the sample was calculated according to the standard IC level of VA solutions and the percentage of IC or VA released from IC-VA/CPC was assessed. Each test was replicated three times (n=5).

Forty-eight male New Zealand white rabbits weighing 2.2–2.6 kg (male and 6 months old) were randomly divided into 4 groups, and each group contained 12 animals. Under general anesthesia (i.v. injection of 15 mg/kg pentobarbital) an osteoperiosteal defect of 15 mm was created through the whole thickness of the shaft of the right radius by removing the segmental bone. Staphylococcus aureus (ATCC28923) suspension (0.2 ml) at a concentration of 5×10⁶ colony forming units (CFU)/ml was injected into each defect. After 60 min, the defects were washed with sterile saline. In group A, the defects were filled with one IC-VA/CPC containing 2 mg of IC and 20 mg of VA. In groups B and C, the defects were filled with one IC/CPC (containing 2 mg of IC) and VA/CPC (containing 20 mg of VA), respectively. The defects of group D were left empty without any filling material as a blank control. The muscles held the graft in place without the need for internal fixation or external splints. At 4, 8 and 12 weeks, four animals from each group were sacrificed for radiological and histological evaluation. All animal experiments were conducted according to the Chinese Regulations of Animal Welfare and permission was granted by the Ethics Committee of the Ningxia Medical University

The samples of CPC (Fig. 1A) and IC-VA/CPC (Fig. 1B) both exhibited three-dimensional structures with numerous granules. The SEM images demonstrated that the loading of IC and VA caused no clear changes on the ultrastructure of CPC. Using SEM, it was also observed that the granules contained in the samples were ~5–10 μm and ball-like or irregular, and were formed on the surface and in the middle of the samples. Furthermore, among the granules, numerous spongy pores were formed, which is hypothesized to be helpful for the degradation of the scaffold as well as the release of drugs.

The proliferation of the cells, which were co-cultured with CPC and IC-VA/CPC, was interpreted by changes in dsDNA quantity. In the CPC and blank control groups, cells proliferated faster during the first 3 days and plateaued at day 5, as indicated by the change in dsDNA content. By contrast, the dsDNA contents of IC-VA/CPC increased rapidly from 1 to 7 days. Between days 3 and 7, the dsDNA contents of the IC-VA/CPC group were significantly higher than those of the CPC and blank control groups (P<0.05). There were no significant differences in dsDNA content between the CPC and blank groups (P>0.05; Fig. 2A). These findings were further corroborated by SEM analysis. One day after the seeding of 3T3 cells on the CPC and IC-VA/CPC, cells appeared flattened, polygonal and spindle-shaped and were distributed evenly. At 7 days, the number of cells on the surface of IC-VA/CPC (Fig. 2B) was increased compared with that on the surface of CPC (Fig. 2C). A number of cells on the IC-VA/CPC had spread and colonized patches of the CPC surface. The spread cells maintained physical contact with each other through filopodia or lamellipodia.

The drug release behavior of the IC-VA/CPC system in vitro was investigated by HPLC examination and then calculated based on a standard curve, and was demonstrated as the accumulated percentage of IC and VA release, respectively (Fig. 3). The IC-release profile exhibited a low burst of rapid drug release. From 0 to 7 days, ~35% IC was released and then the speed decreased and <50% of IC was released by 25 days. From 25 days, there was a second rapid release and 30 days later there was ~15% of IC remaining in the system (Fig. 3A). The VA-release profile showed a modest burst of rapid drug release. From 0 to 5 days, ~45% VA was released and then the speed decreased and ~90% VA was released by 30 days (Fig. 3A).

All wounds in the rabbits of group A (IC-VA/CPC) and group C (VA/CPC) healed completely within 10 days. The wounds of group D (blank control) showed delayed healing of 3–5 weeks, while the wounds in group B (IC/CPC) failed to heal after >5 weeks. The wounds of certain rabbits in groups B and D appeared red and swollen for >7 days, and there was pus exuding from the wounds. Upon bacteriological examination, S. aureus ATCC28923 was detected in the secreted pus. Pus secretion in the wound area stopped and was absorbed without any treatment in >3 weeks. Of the total 48 rabbits used in this study, three animals died; on the 50th day (one animal in group D), the 10th day (one animal of group B) and the 14th day (one animal of group B), respectively. Blood samples from the dead rabbits, collected from the heart under sterile conditions, were cultured for 48 h and S. aureus was detected. Twelve weeks after surgery, S. aureus could still be detected in the bone defects of rabbits from groups B (9/10) and D (2/11) by tissue culture, but not in any animals of groups A and C.

Through radiological observation, a standard defect shadow was observed at the radius of rabbits in group D, while in groups A, B and C, the defects were filled with high-density material, as revealed by the postoperative radiograph. In animals of group A, a bone callus formed by the 4th week, defects started repairing by the 8th week and the medullary cavity was recanalized by the 12th week. By comparison, the defects appeared to be delayed union or nonunion in group C (delayed union, 7; nonunion, 5) and group D (nonunion, 11), and osteomyelitis was observed in group B (10/10) at week 12 (Fig. 4).

In accordance with the radiological examination, different changes were observed in the four groups by histological evaluation. In group A, chondrogenesis and osteogenesis were observed, while no indications of inflammatory infiltration were found 4 weeks postoperatively. On the 8th week, copious new bone tissue was observed within and around the material. There was clear creeping substitution in some areas and the graft was disorganized, largely resorbed and combined with the new bone. The woven bone, lamellar bone and recanalization of the marrow cavity were identified in the majority of the specimens, and the defects were found to have been completely repaired by the 12th week. In group B, animals showed signs of infiltration of chronic inflammatory cells, while the remaining material degraded into small fragments and was embedded in the connective tissue. In group C, a number of animals (7/12) had bone defects that were filled with scar tissue, while the other animals (5/12) had repaired defects. In group D, the bone defects of the animals (11/11) were filled with scar tissue (Fig. 5).