A controlled in vitro study was designed. Independent variables were groups and time, while dependent variables were vancomycin release, antibiotic susceptibility, early biofilm inhibition, cell proliferation, ALP activity, and osteogenic potential of the composites with qRT-PCR. The design of the composites is given in Figure 1.

PLGA and PDLLA were purchased from Evonik Industries (Essen, Germany) and vancomycin hydrochloride was purchased from Zhejiang Medicine Co. Ltd. (Zhejiang, China). β-TCP was purchased from BMT Calsis (Ankara, Turkey). Dichloromethane (JT Baker, PA, USA) was used to dissolve the polymers. In order to prepare vancomycin containing PLGA/β-TCP composites, a total of 8,574 mg of PLGA was dissolved in 40 ml of dichloromethane. Then, 5,355 mg of vancomycin hydrochloride powder was added into the solution, followed by the addition of 16,071 mg β-TCP. The mixture was stirred on a magnetic stirrer with a closed lid and dried at room temperature for 24 h. After the evaporation of dichloromethane, the remaining powdery structure was ground on a porcelain mortar. For vancomycin containing PDLLA/β-TCP composites, all procedures mentioned above were repeated, but this time, 8,574 mg of PDLLA was used instead of PLGA. The powders were hand-pressed in a tablet-pressing machine, and totally, 149 vancomycin containing PLGA/β-TCP composite discs and 160 vancomycin containing PDLLA/β-TCP composite discs were obtained. The composite discs had a 3 mm height with 6 mm diameter. The final content ratios of the composites were 53.6% β-TCP, 28.6% polymer, and 17.8% vancomycin hydrochloride.

Surface topography and composition of the composites were characterized by using a scanning electron microscope (SEM; Nova Nanosem 430, Fei, OR, USA) with a built-in X-ray energy-dispersive spectrometer (EDS). The composites were fixed on supports and coated with gold film to obtain a conducting surface before the analysis.

The vancomycin containing composite discs were further evaluated by using a Fourier transform infrared microscope, which attenuated total reflection (FTIR-ATR; Bruker Alpha, Bruker, MA, USA). PLGA, PDLLA, β-TCP, and vancomycin were also analyzed to determine the similarities and differences between the composite discs and plain materials. The infrared spectrum was collected in the range of 4,000–400 cm⁻¹ with a resolution of 4 cm⁻¹ and a scan number of 24.

Vancomycin containing composite discs from each group (n = 6) were immersed into a 50 ml phosphate buffered saline (PBS) containing polystyrene tubes. PBS was prepared by dissolving one PBS tablet (Oxoid, Hampshire, UK) in 100 ml distilled water. The tubes were placed into a hot water bath at 37°C and shaken constantly at 30 rpm. At predetermined time points (1, 2, 4, 8, 12, 24, 48, and 120 h and 1, 2, 3, 4, 5, and 6 weeks), 1 ml PBS was withdrawn and replaced with an equal amount of fresh PBS. The withdrawn solutions were analyzed spectrophotometrically with a Nanodrop ND 1000 spectrophotometer (Thermo Scientific, MA, USA) at 280 nm with a 1:10 dilution factor. A calibration curve for vancomycin was generated to calculate the amount of released vancomycin in the solution.

MRSA is the most common pathogen isolated from the infection site, but other pathogens like Staphylococcus epidermidis, coagulase-negative staphylococci, Enterobacter species, Pseudomonas aeruginosa, and Mycobacterium species are also responsible for osteomyelitis. Staphylococcus aureus is a gram positive, facultative anaerobe. It has a spherical shape with a 0.5–1.5 μm diameter and forms bead-like clusters when colonized. S. aureus is naturally found in human skin and nostrils. It attaches to the surface with its adhesins and exotoxins and many of its strains are capable of forming biofilm (21). Early biofilm inhibition capabilities of vancomycin containing composites discs were evaluated with the tissue culture plate method. A vancomycin containing composite disc from each group (n = 3) was put into a polystyrene tube containing 10 ml of PBS and placed in a hot water bath at 37°C and shaken constantly at 30 rpm. Every week, 1 ml PBS was withdrawn and replaced with an equal amount of fresh PBS until week 6. Each time point was carried out in triplicate. A slime-forming MRSA strain obtained from Hacettepe University, Faculty of Medicine, Department of Microbiology, was used for this study. Bacterial suspensions were pipetted into sterile glass tubes containing 2 ml of trypticase soy broth (Becton Dickinson, NJ, USA), and the bacterial cultures were set to 0.5 McFarland standard (1 × 10⁻⁸cfu/ml) with a benchtop turbidity meter (Grant Instruments, Cambridge, UK). A 200 μl bacterial suspension was inoculated to fresh 2 ml trypticase soy broth and incubated at 37°C overnight. On another day, bacteria suspension turbidity was checked with a benchtop turbidity meter and the bacteria suspension with 11.0 turbidity was poured into a fresh 48 ml trypticase soy broth. This process was done in duplicate. A 200 μl bacterial culture was added into each well of round bottom 96 well plates (Corning Costar, NY, USA), and 20 μl of drug release media collected from release studies were added to the wells. The plates were incubated at 37°C for 48 h. Trypticase soy broth without any bacteria was used as negative control, while bacterial culture without any release medium was used as positive control.

After incubation, tissue culture plates were turned upside down and the planktonic bacteria were poured out. The plates were washed three times with tap water (200 μl water for each well). After washing, 125 μl 0.1% crystal violet stain was added to each well, and the plates were incubated at room temperature for 10 min. The plates were shaken and the excess stain was poured out; again, the wells were washed twice with water. The plates were placed onto a paper towel and allowed to dry. Each well was filled with 200 μl 95% ethanol, and the plates were incubated at room temperature with closed lids for 15 min. The wells were gently mixed with pipetting, and a 125 μl ethanol-crystal violet mix from each well was placed into a new 96 well plate. The new plates were spectrophotometrically analyzed in ELISA reader (Tecan Sunrise, Mannedorf, Switzerland) at 620 nm.

Composites were evaluated for their proliferation and osteogenic potential in cell culture with hBM-MSCs (passage 6, Lonza, Basel, Switzerland) and SaOS-2 (passage 17, Sigma-Aldrich, MO, USA) cells at days 1, 3, and 7 in triplicate. hBM-MSCs and SaOS-2 cells without any composites were used as control groups. The cells were cultured with the hBM-MSCs or SaOS-2 medium according to cell type. The hBM-MSCs culture medium consisted of 52.8% Dulbecco's Modified Eagle Medium (DMEM) with 1 g/l glucose (Lonza, Basel, Switzerland), 35.2% MCDB-201 medium (Sigma-Aldrich, MO, USA), 10% heat-inactivated fetal bovine serum (FBS; Sigma-Aldrich, MO, USA), 1% penicillin/streptomycin solution (Biochrom AG, Berlin, Germany), and 1% l-glutamine (Biochrom AG, Berlin, Germany), while the SaOS-2 culture medium consisted of 89% DMEM with 4.5 g/l glucose (Sigma-Aldrich, MO, USA), 10% heat-inactivated FBS, and 1% penicillin/streptomycin solution. In every 3–4 days, the media were changed. The assay was done in 24-well cell culture plates (Corning Costar, NY, USA), with analysis for three different time points (on days 1, 3, and 7) in triplicate. In a 24-well cell culture plate, 12 wells were used for MSC, while the other 12 wells were used for SaOS-2 cells. A total of 7,500 cells were seeded on each well and then the composites were placed. The plates were incubated at 37°C with relative humidity under an atmosphere of 5% CO₂. At predetermined time points, the medium was aspirated and a 500 μl fresh medium was added with 50 μl of cell proliferation agent WST-1 (Roche, Basel, Switzerland) for each well. The plates were incubated at 37°C with relative humidity under an atmosphere of 5% CO₂ for 4 h. After incubation, a 110 μl 1:10 (v/v) WST-1 containing culture medium was pipetted into a flat bottom 96-well plate, and the absorbance of the wells was measured in ELISA reader (Tecan Sunrise, Mannedorf, Switzerland) at 450 nm with 620 nm reference wavelength. Early mineralization potential of the vancomycin containing composites was evaluated with ALP activity staining for hBM-MSCs and SaOS-2 cells. Cells were cultured with the hBM-MSCs or SaOS-2 medium according to the cell type, and on day 21, the medium was discarded and a 400 μl Alkaline Phosphatase Yellow Liquid substrate system for ELISA (Sigma-Aldrich, MO, USA) was added. The plate was incubated for 30 min, and 100 μl of 3 N sodium hydroxide (NaOH) was added to stop the reaction. A 200 μl final product was pipetted to a flat bottom 96-well plate and analyzed with an ELISA reader (Tecan Sunrise, Mannedorf, Switzerland) at 405 nm wavelength.

The hBM-MSCs (total 1.5 × 10⁶ cells) were cultured in T75 flasks (Corning Costar, NY, USA) with the hBM-MSCs medium at 37°C with relative humidity under an atmosphere of 5% CO₂, and in every 3–4 days, the media were refreshed. When the cells reached 60%–70% confluency, culture media in three flasks were discarded and replaced with an osteogenic differentiation medium consisting of 10% FBS, 100 nM dexamethasone, 10 mM ß-glycerophosphate (Applichem, Germany), and 0.2 mM l-ascorbic acid (Sigma-Aldrich, MO, USA) in DMEM-LG. The remaining flasks were used for two different composite discs and extraction media were used for this purpose. Briefly, 33 composite discs from each group were incubated with 30 ml of the hBM-MSCs medium. After 14 days of incubation, the cells were trypsinized with 0.25% Trypsin-Ethylenediaminetetraacetic acid (EDTA) (Invitrogen, Gibco, UK) and suspended in 200 μl PBS. mRNA was isolated with a High Pure RNA Isolation Kit (Roche, Basel, Switzerland) and complementary DNA (cDNA) was synthesized with its kit (Roche, Basel, Switzerland). A 15 μl PCR mix and 5 μl cDNA were pipetted into each well of custom plate designed with different signaling molecules (Roche, Basel, Switzerland). The final PCR reaction was quantified in a Lightcycler 480 and its software was used to calculate the crossing point (Cp) for target and reference expression with the Advance Relative Quantification method. All target genes were normalized to housekeeping genes ACTB (beta actin), GAPDH (glyceraldehyde 3-phosphate dehydrogenase), and G6PD (glucose-6-phosphate dehydrogenase). The results were given as fold change corresponding to the hBM-MSCs control group according to ΔΔCt calculation. The sequences of primers are given in Table 1.

All results were presented as average ± standard deviation and analyzed with SPSS 11.0. Statistically significant values were defined as p < 0.05 based on Student’s t-test. For determining the significance of the expression fold changes between the groups, the binary logarithm of the ΔΔCt values was calculated and ±two-fold changes were assigned as significant for the qRT-PCR study. The significant values are indicated in gray boxes in Table 2.

The surfaces of the composites contained micro cracks. The surface properties of the PDLLA/β-TCP and the PLGA/β-TCP composites were similar (Figure 2).

The adsorption bands of vancomycin were recorded at 3,252, 1,644, 1,487, 1,225, 1,014, and 426 cm−¹. The adsorption band at 3,252 cm⁻¹ was for O–H stretching, while 1,644 cm⁻¹ showed C=O stretching. The bands at 1,487 and 1,225 cm⁻¹ pointed at C=C band and C–O–C band, respectively (22). The adsorption bands of β-TCP were found at 1,212 cm⁻¹ (the pyrophosphate CPP group band), 1,017 cm⁻¹ (C–O stretching), 727 cm⁻¹ (P–O stretching), and 542 cm⁻¹ (P–O bending) (23). The characteristic peaks of PDLLA and PLGA were found at 1,746 cm⁻¹ (C=O band), 1,183 cm⁻¹ (C–O band), 1,022 cm⁻¹ (C–O band), and 540 cm⁻¹ (C–H band) (24). The peaks were recorded at 1,749 cm⁻¹ (C=O band), 1,017 cm⁻¹ (C–O band), and 538 cm⁻¹ in both vancomycin containing composites. The similarities of spectra were pointed in circles; the color red defined vancomycin, green defined β-TCP, purple defined PDLLA, and blue defined PLGA (Figure 3).

Both PDLLA/β-TCP and PLGA/β-TCP composites maintained a sustained release of vancomycin for 6 weeks. The PDLLA/β-TCP composites released 2.3 ± 0.2 mg vancomycin, while the PLGA/β-TCP composites released 2.1 ± 0.2 mg in a day. After 6 weeks, cumulatively, 3.1 ± 0.2 and 3.4 ± 0.4 mg vancomycin were released from the PDLLA/β-TCP and the PLGA/β-TCP composites, respectively (Figure 4).

In the biofilm inhibition study, there was a statistically significant difference between the composite groups and the bacterial control (p < 0.05). Released medium added to the bacterial suspensions inhibited early biofilm formation throughout 6 weeks (Figure 5).

Cells cultured on the composites proliferated, and there was a statistically significant difference for the PDLLA/β-TCP group between day 1 and day 7 for both cell lines (p = 0.01 for hBM-MSCs and p = 0.03 for SaOS-2, respectively). On the contrary, the PLGA/β-TCP group only showed a statistically significant difference between day 1 and day 7 for the SaOS-2 cell line (p = 0.03). Both composite groups showed a statistically significant difference versus blank hBM-MSCs on day 7 (p = 0.01), but there was no such significant difference for SaOS-2. There was no statistically significant difference between the groups for day 1 to day 3 or for day 3 to day 7 (Figure 6).

PDLLA/β-TCP (p = 0.011) and the PLGA/β-TCP (p = 0.006) composites cultured with SaOS-2 cells presented a higher ALP activity compared with the SaOS-2 cell group without any composite. The ALP activity of the composites cultured with hBM-MSCs, however, was higher but was not statistically significant than the hBM-MSCs group (Figure 7).

The alkaline phosphatase (ALPL) upregulated 3.44-fold higher in the osteogenic medium group, while it downregulated in composite groups according to control. This downregulation, however, was statistically not significant. Bone morphogenetic protein-1 (BMP-1) upregulated significantly only in the osteogenic medium group, while bone morphogenetic protein-2 (BMP-2) was upregulated only in the composite groups. Bone morphogenetic protein-3 (BMP-3) downregulation was statistically significant only for PDLLA/β-TCP. Bone morphogenetic protein-4 (BMP-4) was not expressed significantly in any group. Collagen, type X, and alpha 1 (COL10A1) expressed significantly only in the PLGA/β-TCP group. The cartilage oligomeric matrix protein (COMP) was upregulated in each group, but its expression was significant only in the osteogenic medium and in the PLGA/β-TCP groups. Cathepsin K (CTSK), insulin-like growth factor 2 (IGF2), and matrix metallopeptidase 8 (MMP8) expressions upregulated only in the osteogenic medium group. The expression of insulin-like growth factor 1 (IGF1) increased in all groups, while it was significant for the PLGA/β-TCP group.

SMAD expressions decreased in the composite groups, which were significant only for SMAD1 in the PDLLA/β-TCP group.

Transforming growth factor, beta 2 (TGF-β2), downregulated in the PLGA/β-TCP group, while TGF-βR1 (receptor 1 of TGF-β) downregulated in all groups. Vitamin D receptor (VDR) downregulated in the osteogenic medium group; on the contrary, it was upregulated significantly in the composite groups. RUNX2 upregulated in the osteogenic medium group and downregulated in the composite groups; however, these expressions were not significant (Table 2).