Multifunctional antibiotic- and zinc-containing mesoporous bioactive glass scaffolds to fight bone infection

Abstract

Bone regeneration is a clinical challenge which requires multiple approaches. Sometimes, it also includes the development of osteogenic and antibacterial biomaterials to treat the emergence of possible infection processes arising from surgery. This study evaluates the antibacterial properties of gelatin-coated meso-macroporous scaffolds based on the bioactive glass 80%SiO₂–15%CaO–5%P₂O₅ (mol-%) before (BL-GE) and after being doped with 4% of ZnO (4ZN-GE) and loaded with both saturated and the minimal inhibitory concentrations of one of the antibiotics: levofloxacin (LEVO), vancomycin (VANCO), rifampicin (RIFAM) or gentamicin (GENTA). After physical-chemical characterization of materials, release studies of inorganic ions and antibiotics from the scaffolds were carried out. Moreover, molecular modelling allowed determining the electrostatic potential density maps and the hydrogen bonds of antibiotics and the glass matrix. Antibacterial in vitro studies (in planktonic, inhibition halos and biofilm destruction) with S. aureus and E. coli as bacteria models showed a synergistic effect of zinc ions and antibiotics. The effect was especially noticeable in planktonic cultures of S. aureus with 4ZN-GE scaffolds loaded with VANCO, LEVO or RIFAM and in E. coli cultures with LEVO or GENTA. Moreover, S. aureus biofilms were completely destroyed by 4ZN-GE scaffolds loaded with VANCO, LEVO or RIFAM and the E. coli biofilm total destruction was accomplished with 4ZN-GE scaffolds loaded with GENTA or LEVO. This approach could be an important step in the fight against microbial resistance and provide needed options for bone infection treatment. Statement of Significance: Antibacterial capabilities of scaffolds based on mesoporous bioactive glasses before and after adding a 4% ZnO and loading with saturated and minimal inhibitory concentrations of levofloxacin, vancomycin, gentamicin or rifampicin were evaluated. Staphylococcus aureus and Escherichia coli were the infection model strains for the performed assays of inhibition zone, planktonic growth and biofilm. Good inhibition results and a synergistic effect of zinc ions released from scaffolds and antibiotics were observed. Thus, the amount of antibiotic required to inhibit the bacterial planktonic growth was substantially reduced with the ZnO inclusion in the scaffold. This study shows that the ZnO-MBG osteogenic scaffolds are multifunctional tools in bone tissue engineering because they are able to fight bacterial infections with lower antibiotic dosage.

1. Introduction

The emergence of infection processes is one of the most common and challenging complication during and after a surgical intervention [1,2]. Bone bacterial infection is a serious issue with important clinical and socio-economic implications. It is an inflammatory process which provokes osteolysis [3,4]. Post-operative implant infections are one of the most serious complications associated with bone diseases and fractures, treated with bone grafts and prostheses [5,6]. Bacteria of principal concern include Staphylococcus aureus, Staphylococcus epidermidis and Escherichia coli [5,7,8] Treatment usually involves surgery, antibiotic systemic administration and implant removal which has limitations and repercussions in patients live quality: additional surgical interventions, prolonged hospital stays, high side effects and higher mortality [9,10]. The appearance of bacterial infection derived from surgical procedures (332000 total hip and 719000 total knee arthroplasties performed in 2010, just in the United States) [11] represent a substantial problem. The numbers are projected to reach 572000 and 3.48 million by 2030 for hips and knees, respectively [11]. In Europe, the situation is analogous, but larger number of patients undergoes primary hip arthroplasty than knee arthroplasty. Modelling data predicts the incidence of prosthetic joint infection will increase to greater than 6% by 2030 [11] owing to factors such as increased demand for surgery, the aging population, and the obesity epidemic [12–14]. Moreover, the number of people over 50 years suffering from bone related diseases, was envisaged to increase two-fold from 2012 to 2020 [15]. One of the principal reasons of the failure is the appearance of antibiotic-resistant strains [16,17] and the bacteria capability to develop a biofilm. Biofilms are groups of bacteria growing together which are able to produce an extracellular matrix accomplishing in a cooperative manner [16,17]. In a biofilm, bacteria grow protected from environmental stresses, develop antibiotic resistance, and skip the innate and adaptive immune system host strategies [18,19]. Being aware that oral and injection antibiotic administration involve several problems, higher antibiotic concentrations are required which is not a recommendable approach. Moreover, the administration of antibiotics below the minimal inhibitory concentration (MIC) leads to the development of antibiotic-resistant species. Therefore, the development and use of local strategies, including antibiotic release systems, able to eliminate or reduce the bacterial attachment, growth, and biofilm formation have attracted significant attention in bone tissue engineering [20].

With this purpose, the development of multifunctional 3D scaffolds based on Mesoporous Bioactive Glasses (MBGs) [21], doped with antibacterial ions [22] and loaded with low enough concentration of different antibiotics to avoid antibiotic resistance is a promising approach [23,24]... Moreover, MBGs present huge surface area and pore volume [25], which provides fast in vitro and in vivo responses, bone regenerative capability [26] and the possibility of being doped with different ions to add additional capabilities to the biomaterial [27, 28]. It has been demonstrated that the variation of intracellular ion concentration caused by the bioactive glass biodegradation, leads the activation of intracellular signalling pathways [27]. Mainly, zinc, copper, and silver ions, incorporated into the glass structure, allow the release of ions with antibacterial properties [29]. Moreover, they can influence gene expression mechanisms of osteoprogenitor cells, increasing bone regeneration [27]. Zinc ions are a good option for that purpose, due to their specific capacity to increase osteogenesis and their antimicrobial properties [30].

In a previous article, [31] we demonstrated the synergistic effect of Zn²⁺ ions and the osteogenic peptide osteostatin released from ZnO-MBG scaffolds, in the proliferation and differentiation of mesenchymal stem cells. In the present paper we investigated the antibacterial effect of Zn²⁺ ions released from the aforementioned scaffolds in S. aureus and E. coli cultures, before and after load the scaffolds with one of the following antibiotics: levofloxacin (LEVO), vancomycin (VANCO), gentamicin (GENTA) or rifampicin (RIFAM).

With this purpose, two types of MBG scaffolds were synthesised, first one with composition 80%SiO₂–15%CaO–5%P₂O₅ (mol-%) (BL), and the second one, with analogous composition, included a 4% of ZnO (4ZN). After coating the scaffolds with gelatin (BL-GE and 4ZN-GE) they were loaded with LEVO, VANCO, GENTA or RIFAM (BL-GE-L/V/G/R and 4ZN-GE-L/V/G/R). BL and 4ZN scaffolds were tested in S. aureus and E. coli strains, as bone infection bacteria models, in three types of assays: inhibition zone, planktonic growth and biofilm. To understand the interactions between the antibiotics and the scaffolds, molecular modeling calculations were performed. The objective of this study is based on checking whether the antibacterial activity of Zn²⁺ ions, released from the scaffolds, allows the amount of antibiotic used to be reduced maintaining their antibacterial function.

2. Materials and methods

2.1. Reactants

Reactants were purchased from Sigma-Aldrich (St. Louis, USA) Inc. to perform the described experiment: tetraethyl orthosilicate (TEOS), triethyl phosphate (TEP), calcium nitrate tetrahydrate (CaNO₃)₂·4H₂O), pluronic F123 (EO20PO70EO20), ε-caprolactone (Mw = 58,000 Da) and hydrochloric acid (HCl, 37 %). All antibiotics (levofloxacin, vancomycin, gentamicin sulfate and rifampicin) were also from Sigma-Aldrich (Sigma-Aldrich, St Louis, MO). Absolute ethanol (EtOH) was purchased from Panreac Quimica SLU (Castellar del Valles, Barcelona, Spain), All reagents were used as received without further purification. Ultrapure deionized water with resistivity of 18.2 MΩ was obtained using a Millipore Milli- Q plus system (Millipore S.A.S., Molsheim, France).

2.2. Fabrication of mesoporous bioactive glass scaffolds

The manufacturing process of the scaffolds began with the synthesis of two compositions of MBGs: 80%SiO₂–15%CaO–5%P₂O₅ (mol %), BL, and 82.2%SiO₂–10.3%CaO–3.3%P₂O₅–4.2%ZnO (mol %), 4ZN, by the Evaporation Induced Self Assembly (EISA) method. The MBGs obtained as powders were used to build the 3D meso-macroporous scaffolds by rapid prototyping as described by Heras et al. [31]. In short, powders were mixed with polycaprolactone (PCL) dissolved in dichloromethane to later evaporate the solvent until obtain a paste with the right rheological properties to print. As a result, 7 mm diameter 5 mm high scaffolds were obtained. Printed scaffolds were calcined to remove PCL and traces of solvent. Finally, scaffolds were coated with gelatin (GE) crosslinked with glutaraldehyde to increase their mechanical consistency and improve the diffusion of ions and antibiotics to the medium. Resultant scaffolds, used for the in vitro assays, were denoted as BL-GE and 4ZN-GE.

2.3. Meso-macroporous 3D scaffolds characterization

Scaffolds were characterized by; small angle X-ray diffraction (SA-XRD) in a X'pert-MPD system (Eindhoven, The Netherlands) equipped with Cu Kα radiation in the 0.6 to 8° 2θ range; transmission electron microscopy (TEM) in a JEM-2100 JEOL microscope (Tokyo, Japan) operating at 200 kV; Fourier transform infrared FTIR spectroscopy, in a Thermo Scientific Nicolet iS10 (Waltham, MA, USA) equipped with a SMART Golden Gate attenuated total reflection ATR accessory; and scanning electron microscopy (SEM) in a JEOL JSM-6400 (Tokyo, Japan) operating at 20 kV. In addition, thermogravimetric analysis (TGA) in a Perkin Elmer Pyris Diamond system (Waltham, MA, USA) allowed to quantify the drug loading and Nitrogen adsorption measurements in a Micromeritics 3 Flex (Norcross, GA, USA), were performed before and after loading the scaffold with the antibiotics to verify the loading process.

2.4. In vitro antibiotic loading and release

After sterilization by UV irradiation for 60 min, scaffolds were loaded with two concentrations of each antibiotic: saturated concentration and the concentration required to reach the minimum inhibitory concentration (MIC = 50%), which is the one that inhibits by 50% the bacterial growth of each strain. The loading process was carried out by immersing scaffolds of 80 mg in a 3 mL solution of each antibiotic, in their specific solvent (water or ethanol) at the desired concentration. Immersed scaffolds were kept under stirring at 100 rpm for 24 h at room temperature. Then, scaffolds were vigorously washed and dried for 48 h at 37 °C. LEVO and RIFAM are light sensitive, so in these cases the entire process was done in dark conditions to avoid degradation.

Dry scaffolds were immersed in phosphate buffered saline (PBS) solution adjusted at physiological pH of 7.4. Pieces were kept at 37 °C in stirring conditions (100 rpm) up to 18 days. 500 μL aliquots were extracted each day to measure the antibiotic release by UV spectrophotometry. The whole antibiotic release procedure was also performed at pH 6.5, which corresponds with the acidic pH generated in an infection process. There were no significant differences between both pH and because of that release results at pH 6.5 are not shown.

2.5. Molecular modelling

A molecular model based on mesoporous MCM-41 silica doped with zinc, similar to mesoporous 4ZN glass, was generated by using Spartan'14 software (Wavefunction Inc.). The building unit in this model was the Si₆O₁₂ pseudo–cell, consisting of hexagonal arrangements of Si–O–Si and Si–O–Zn units until reaching a model of 2 nm length and 4 nm diameter. The final structure was refined by geometry optimization using the MM+ force field. Docking calculations were carried out on the MCM-41 surface model by Hex 8.0 and CUDA 5.0 Nvidia GPU acceleration software with Polar fast Fourier transform (FFT) method and Shape+electro+DARS (decoys as reference state) correlation type with a cubic box of 10 nm and grid dimension of 0.6 nm. The correlation was only shape type. For electrostatic potential maps, coordinates of gentamicin 5+, zwitterionic rifampicin, zwitterionic levofloxacin and vancomycin 1+ were obtained from PubChem (U.S. National Library of Medicine) and treated by molecular dynamics simulations from Spartan'14 (Wavefunction, Inc.). The energy was optimized by AM1 Hamiltonian Method.

2.6. Ion release assay

For these assays, each scaffold was placed in a well of a transwell plate of 12 wells with 2 mL of Todd Hewitt Broth (THB) in each of them. They were kept at 37 °C with stirring (100 rpm) for 10 days. Each day, the 2 mL were extracted in full to measure in the culture medium the amount of Ca, P and Zn ions that had been released and another 2 mL of fresh THB were added to continue with the release study. Ion concentrations were determined by inductively coupled plasma/optical emission spectrometry (ICP/OES) using an OPTIMA 3300 DV device (Perkin Elmer). Each ion was determined in two different samples measured by triplicate.

2.7. Bacterial cultures

For these studies, Gram-positive Staphylococcus aureus (ATCC 29213) and Gram-negative Escherichia coli (ATCC25922 l) were used as bacteria models.

2.7.1. Agar disk-diffusion tests of drug loaded materials

Zinc-doped scaffolds (4ZN-GE) and loaded with the saturated concentration of each antibiotic, were immersed in 2 mL at pH 7.4 and 6.5 and kept at 37 °C and stirring conditions (100 rpm). 7.4, was selected because corresponds to physiological one, and 6.5 because matches with generated in infection environment. Aliquots (20 μL) were extracted at several times to load foams placed later in contact with a pre-seeded (100 μL) of S. aureus or E. coli having 2 × 10⁶ bacteria/mL) agar plates. After 24 h of culture in static conditions at 37 °C, the inhibition zone was analysed. The bacterial inhibition zone size was measured as: (outer diameter of the inhibition zone - disk diameter)/2. Each study was performed in triplicate.

2.7.2. Planktonic growth inhibition test

Pre-loaded scaffolds, at saturated or minimal inhibitory concentration (MIC) of each antibiotic for each strain (S. aureus and E. coli), were placed in a 12 transwell plate. Then, 3 mL of bacteria dissolution (2 × 10⁸ bacteria/mL) were added to put in contact bacteria culture with the zinc doped (4ZN-GE) or undoped (BL-GE) scaffold loaded with the different antibiotics. 24 h later, aliquots (20 μL) were extracted and seeded in agar plates. After 24 h of culture at 37 °C in static conditions, colony forming units (CFUs) were counted to know how the material, ions and antibiotic, affect the bacterial growth. Results of 4ZN-GE-antibiotic loaded scaffolds and BL-GE-antibiotic loaded scaffolds were compared. For the tests at saturated drug concentration, bacteria culture controls without material and controls with bacteria cultured with drug unloaded materials were carried out. For MIC tests, controls were carried out with bacteria cultures, growing in contact with the MIC of each drug for each strain, but without any material.

2.7.3. Biofilm degradation

1 mL of both strains (S. aureus and E. coli) was seeded (2 × 10⁸ bacteria/mL) on a glass slide, kept at 37 °C and stirring conditions (100 rpm). During the first 24 h of culture, the medium (THB) was doped with 4% of sucrose. Then, it was replaced by a normal one. 24 h later, the biofilm was already formed. Pre-formed biofilms were placed in a 12 transwell plate were the different type of scaffolds were also located (BL-GE, 4ZN-GE and 4ZN-GE-Drug loaded scaffolds). After 24 h of culture at 37 °C and stirring conditions (100 rpm), the biofilm was extracted and stained with live and dead BacLigth bacteria viability kit (Thermo Fisher Scientific, Invitrogen™) and tested in a MC1025 confocal laser scanning microscope (Biorad). Alive bacteria were stained in green (SYTO 9), dead bacteria in red (propidium iodide, PI) and extracellular matrix in blue (calcofluor).

2.7.4. Statistical Analysis

Results are expressed as mean ± SEM (SEM: standard error of mean). Statistical evaluation was carried out with nonparametric Kruskal-Wallis test and post-hoc Dunn's test, when appropriate. A value of p < 0.05 was considered significant.

3. Results and discussion

3.1. Meso-macroporous 3D scaffolds characterization

Structural powder characterization by SA-XRD in Fig. 1, revealed that gelatin containing scaffolds, exhibited maxima in the mesoporous order indicative region. However, differences between samples as a function of the ZnO content were observed. Thus, BL-GE displayed a sharp diffraction maximum at 2θ in the region of 1.0–1.4°, assigned to the (10) reflection along with a poorly resolved peaks at around 2.0 that can be assigned to the (11) reflection. These maxima were indexed on the basis of an ordered two-dimensional (2D) hexagonal structure (plane group p6mm). The intensity of the (10) maximum decreased while the ZnO content increased, indicating a partial deterioration of the mesoporous structure in 4ZN-GE scaffolds. TEM images of zinc-doped (4ZN) and undoped (BL) scaffolds, obtained with the electron beam parallel to the mesoporous channels, are shown in Fig. 1. Therefore, both samples exhibited a typical 2D-hexagonal ordered mesoporous arrangement. FTIR profile of the 4ZN, 4ZN-GE, BL and BL-GE scaffolds showed the characteristic bands of each material, including the bands of gelatin used to coat the scaffolds (Fig. 2A). SEM micrographs from 4ZN-GE scaffolds, showed two types of channels (Figs. 2B-D). First of them were channels with sizes between 700 and 1000 μm (Fig. 2B) and the second ones were macropores with sizes between 1-10 μm (Figs. 2C-D). Higher magnification (Figs. 2C-D) shows both, surface and fracture views of the 300 μm wide bars, mentioned macropores between 1-10 μm and the homogeneous gelatin layer on the material surface (Figs. 2C-D). It should be mentioned that the mesopores (2-10 nm) are so small that they cannot be observed at the magnification that the SEM images were taken. BL-GE material showed similar SEM structure to 4ZN-GE. Therefore, BL-GE images were omitted for Fig. 2 simplification.

Fig. 1.

SA-XRD patterns and TEM micrographs of BL and 4ZN scaffolds.

Fig. 2.

A. FTIR spectra of 4ZN, 4ZN-GE, BL and BL-GE scaffolds. SEM micrographs of 4ZN-GE scaffold: B. Front view of the 700–1000 μm channels. C. Detail of the intersection of two bars where macropores of between 1-10 μm are displayed. D. Fracture detail of the intersection where the macropores are also visualised (1-10μm).

3.2. Ion release assay

To understand the possible antibacterial activity of the scaffolds, the cumulative release of calcium, phosphorus and zinc ions after soaking in THB was measured (Fig. 3). For BL-GE scaffolds, the lower release of Ca²⁺ ions compared with 4ZN-GE was attributed to its higher in vitro bioactive behaviour. Thus, a part of the Ca²⁺ ions released would precipitate as carbonate hydroxyapatite (CHA) on the scaffold surface being eliminated from medium. In contrast, 4ZN-GE scaffolds displayed slower bioactive behaviour and, consequently, most part of the released Ca²⁺ ions from these scaffolds remained in the solution. Regarding phosphorous ions, 4ZN-GE scaffolds showed lower concentration in the medium that was attributed to the precipitation of calcium and zinc phosphates [31]. Accumulated concentration of zinc increased about 1.5 ppm/day in the first 48 h and more slowly for longer times, in the same conditions than those of bacterial assays.

Fig. 3.

Amount of Ca²⁺, P(V) and Zn²⁺ ions released from BL-GE and 4ZN-GE scaffolds as a function of the soaking time in THB.

3.3. Antibiotic loading in MBG scaffolds

Antibiotic adsorption was performed at two different concentrations: saturated and the Minimal Inhibitory Concentration (MIC) for each antibiotic and tested strain.

3.3.1. Thermogravimetric Analysis

Thermogravimetric analysis (TGA) of scaffolds provided the amount of drug loaded in each case (Table 1). This analysis showed the satisfactory drug loading into the scaffolds by detecting losses of mass which correspond to the loaded drug.

Table 1.

Textural properties of BL-GE and 4ZN-GE before and after to be loaded with antibiotic (L. levofloxacin, V: vancomycin, G: gentamicin or R: rifampicin) determined by nitrogen adsorption. S_BET: BET surface area; V_p: pore volume. D_P: pore diameter. The amounts of saturated antibiotic loaded in each sample determined form TGA were also included.

MATERIAL	BL-GE	BL-GE-L	BL-GE-V	BL-GE-G	BL-GE-R	4ZN-GE	4ZN-GE-L	4ZN-GE-V	BL-GE-G	4ZN-GE-R
SBET (m 2/g)	188	146.5	180.5	180.7	126.5	151.4	111.3	107.5	124.5	142.2
VP(cm 3/g)	0.21	0.15	0.21	0.21	0.12	0.14	0.06	0.11	0.11	0.13
DP (nm)	3.7	3.5	3.7	3.7	3.4	3.5	3.3	3.5	3.4	3.5
mol drug / mg scaffold	0	1.3 × 10 −4	1.1 × 10 −5	1.3 × 10 −5	8.4 × 10 −6	0	1.3 × 10 −4	1.1 × 10 −5	1.3 × 10 −5	8.4 × 10 −6

3.3.2. Nitrogen Adsorption Porosimetry

Nitrogen adsorption–desorption isotherms and pore size distributions of BL-GE and 4ZN-GE scaffolds, before and after loaded with each antibiotic, are shown in Fig. 4. All curves can be identified as type IV isotherms, characteristic of mesoporous materials (Fig. 4.A). Observed type H1 hysteresis loops in the mesopore range are characteristic of open cylindrical pores. Fig. 4.B correspond to pore size distributions. Variations were detected when the scaffolds were loaded with the antibiotics. Loaded scaffolds showed a decrease both in surface area and pore volume verifying the drug loading into the mesopore structure.

Fig. 4.

A Nitrogen adsorption isotherms and B pore size distributions of BL-GE and 4ZN-GE scaffolds before and after be loaded with the antibiotics.

Table 1 collects the textural properties, i. e., surface area (S_BET), pore diameter (D_P) and pore volume (V_p) of loaded and non-loaded scaffolds. In general, higher values were found for antibiotic unloaded samples. Thus, BL-GE and 4ZN-GE antibiotic-loaded scaffolds, presented a slight decrease in surface area and pore volume because of the drug loading. These data allowed to verify the successful drug loading into the mesopores of MBG scaffolds.

3.4. Molecular modelling

MBGs investigated have compositions based in SiO₂–CaO–P₂O₅ (BL) or SiO₂–CaO–P₂O₅–ZnO (4ZN) systems. There, silicon and phosphorus elements, as SiO₂ and P₂O₅, are the basis of the glass network in tetrahedral units that bond each other through covalent bonds by their oxygen atoms. Calcium, as CaO, is a network modifier bonded through ionic interactions to the oxygen atoms. Moreover, our research group demonstrated from ²⁹Si NMR spectroscopy that 4ZN-GE scaffolds contain [ZnO₄] tetrahedra [31]. These tetrahedra exhibit negative charge (2–), which justifies the attraction of Ca²⁺ and Zn²⁺ ions acting as network compensators of charge instead of as network modifier cations. Accordingly, the number of non bonding oxygen (NBO) decreases with the increase of Q⁴ species and the Q³ species decrease Moreover, these MGBs present in their surface ionized silanol Si–OH groups which produce a negative charge that interact with the loaded drugs.

Electrostatic potential mapped density (Fig 5.A) and molecular modelling interaction studies (Fig 5.B) of 4ZN-GE model loaded with LEVO VANCO, GENTA, and RIFAM were calculated. Molecular modelling indicates a positive total charge in VANCO and GENTA and a zwitterionic nature in RIFAM and LEVO. Moreover, modelling showed strong 6 and 4 hydrogen bonds with VANCO (-295.7 kcal/mol) and GENTA (-250.1 kcal/mol), respectively. In case of RIFAM (-226.1 kcal/mol) and LEVO (-368.7 kcal/mol) modelling exhibited 2 and 0 hydrogen bonds, respectively. Although VANCO has 6 hydrogen bonds that favourably interact with 4ZN-GE surface, it is a large molecule (1450.3 uma) which causes low antibiotic loading. Zwitterionic LEVO presents no hydrogen bonds but it is the smallest molecule (823.9 uma) allowing the formation of a stable complex and achieved a high loading yield.

Fig. 5.

A. Electrostatic potential mapped density calculations carried out with Hex 8.0 software. B. Molecular modelling interaction studies where 4ZN-GE model was used as receptor and gentamicin 5+, levofloxacin zwitterion, rifampicin zwitterion or vancomycin 1+ as ligands.

3.5. Antibiotic release from MBG scaffolds

Antibiotics release assays were performed in PBS at physiological pH of 7.4. The use of THB for these studies was discarded because the bacterial growth induction in this medium at the analysis conditions. No significant differences were obtained between BL-GE and 4ZN-GE. Thus, only results corresponding to the last one will be presented. Fig. 6 shows the release from 4ZN-GE scaffolds of each antibiotic along time. Antibiotics release kinetics from the MBGs were evaluated according to zero-order kinetics and first-order kinetic model. Zero-order kinetics,

$$Q_0-Q_t=k_0·t$$ (1)

Fig. 6.

Antibiotic release curves from 4ZN-GE at physiological pH of 7.4.

Where Q_t is the amount of drug remaining as a solid at time t, Q₀ is the initial amount of drug in the pharmaceutical dosage form and k₀ is the zero-order release rate constant. As it is observed in Fig. 6, LEVO presents a release preserved in time, maintaining the release even after 15 days. It must be considered that LEVO presents a zwitterionic nature (pKs 6.1-8.2) [32] having a neutral total charge but being able to establish electrostatic interactions with the silanol groups of the MBG matrix and providing a maintained release favoured by its high load and solubility [32]. However, in case of RIFAM, VANCO and GENTA, the release data showed in Fig. 6 fit better to a first-order kinetic model with an empirical non-ideality factor (δ) [33].

$$Y=A{\left(1-e^{k1t}\right)}^{\delta }$$ (2)

Where Y is the percentage of antibiotic released at time t, A the maximum amount of antibiotic released (in %), and k₁, the release rate constant. δ values are 1 for first-order kinetics materials and 0, for materials that release the loaded drug in the very initial time of analysis. RIFAM, VANCO and GENTA present zero and positive total charge of 6+ and 5+, respectively at the physiological pH of assay. δ value provides a fidelity degree approximation to proposed model for theoretical first-order kinetics (Table 2). When δ is near to 1, first-order kinetic is more accurate and drug delivery from mesopores has smaller burst effect. This effect is usually attributed to the adsorption of drug molecules into the mesopores of the matrix. As it can be observed in Table 2, the biggest value of δ is 0.83 for RIFAM and then VANCO, 0.45 and finally GENTA 0.27. As is observed in Fig. 6, RIFAM release is maintained due to its stable interaction with the surface of MBGs and its limited solubility. However, VANCO present a less stable interaction with MBGs and GENTA a higher solubility having, in both cases, a faster release.

Table 2.

Parameters of antibiotics release kinetics from 4ZN-GE scaffolds. w₀: initial loaded mass (wt% of antibiotic loaded); A: maximum relative release; k₁: constant of Chapman model release rate and K _LEVO zero model release rate; δ: non-ideality factor in Chapman model; R: goodness of fit.

Sample	pH	w0 (wt%)	A (%)	K1(x103h−1) K0 LEVO	δ	R
4ZN-GE-L	7.4	6	71	0.05	-	0.99
4ZN-GE-V	7.4	1	87	10.5	0.45	0.97
4ZN-GE-G	7.4	2	99	8.0	0.27	0.97
4ZN-GE-R	7.4	4	78	9.5	0.83	0.99

At this pH of 7.4, interactions between the antibiotics and the thin gelatin layer coating the scaffolds were not considered, due to the pKs: pKb ~ 6.5 and pKa ~ 4.7 of NH₂ and COOH ionisable groups of gelatin that were not able to establish electrostatic interactions with the drugs.

On the other hand, the complete antibiotic release assay was also performed at pH 6.5, which corresponds to the pH generated in an infection process. There were no significant differences for results obtained at pH 7.4 and at pH 6.5. For this reason, the antibiotic release results at pH 6.5 are not shown.

3.6. Agar disk-diffusion tests of drug loaded materials

3.6.1. S. aureus

4ZN-GE scaffolds were loaded with LEVO, VANCO and RIFAM because their referenced effectiveness against S. aureus (Fig. 7.1-3) [34–36]. The antibiotic loaded scaffolds were named as 4ZN-GE-L, 4ZN-GE-V and 4ZN-GE-R, respectively. The assay allowed evaluating the capability of the scaffolds loaded samples to eradicate bacteria. Disks foams were impregnated with aliquots (20 μL) containing the antibiotic and zinc ions released from scaffolds at each time at both infection and physiological conditions, and put in contact with pre-seeded S. aureus agar plates. 4ZN-GE-L scaffolds inhibited S. aureus growth for 18 days (Fig. 7.1) and reached their maximal inhibition after 5 days with a 14 mm halo. 4ZN-GE-V scaffolds inhibited S. aureus growth for 5 days. 4ZN-GE-V scaffolds reached their maximal inhibition after 6 h showing 18 % of inhibition respect LEVO (Fig. 7.2). 4ZN-GE-R scaffolds inhibited S. aureus growth for 10 days and reached their maximal inhibition after 48 h showing 57 % of inhibition compared to LEVO (Fig. 7.3). There were no significant differences between results obtained at physiological, 7.4, and infection, 6.5, pH values.

Fig. 7.

Variation with time of scaffolds inhibition zone of 4ZN-GE-L (1), 4ZN-GE-V (2) and 4ZN-GE-R (3) against S. aureus, and of 4ZN-GE-L (4), 4ZN-GE-G (5) and 4ZN-GE-R (6) against E. coli, at physiological pH of 7.4 and pH 6.5 of an infection environment. * Indicates significant differences between pH 7.4 and pH 6.5. Statistical significance: p < 0.05.

3.6.2. E. coli

4ZN-GE scaffolds were loaded with LEVO, GENTA and RIFAM active antibiotics against E. coli (Fig. 7.4-6) [34,37,38]. Loaded scaffolds were named as 4ZN-GE-L, 4ZN-GE-G and 4ZN-GE-R, respectively. Agar disk diffusion tests were performed as above. 4ZN-GE-L scaffolds inhibited E. coli growth for 18 days (Fig. 7.4). 4ZN-GE-L scaffolds reached their maximal inhibition after 72 h with a 15 mm halo. 4ZN-GE-G scaffolds inhibited E. coli growth for 72 h with a maximal inhibition at 6 h, showing 20% of inhibition respect LEVO (Fig. 7.5). 4ZN-GE-R scaffolds inhibited E. coli growth for 120 min and reached their maximal inhibition after 48 h showing 4 % of inhibition respect LEVO (Fig. 7.6). As Fig. 7 shows, the most effective antibiotics against E. coli were LEVO and GENTA. The inhibition zone created by RIFAM was just the 4% respect LEVO. There were no significant differences between physiological and infection pH values.

3.7. Planktonic growth inhibition tests

Tested scaffolds were loaded with a saturated antibiotic concentration and the loaded drug was quantified (Table 1) to assess the antibiotic effect against bacteria. The antimicrobial effect of BL-GE and 4ZN-GE scaffolds, with and without drugs, at concentrations of 27 and 80 mg scaffold/mL, was evaluated between 2 and 48 h by measuring colony forming units (CFUs)/mL. Fig. 8 shows the CFUs/mL for both strains (S. aureus and E. coli) and both materials (BL-GE and 4ZN-GE), loaded or not with the tested antibiotics. Regarding S. aureus cultures, bacteria growth for VANCO, LEVO and RIFAM loaded BL-GE and 4ZN-GE scaffolds was inhibited at a 99.9%. For unloaded 4ZN-GE scaffolds, only after 48 h the bacteria growth decreased up to a 50% compared to BL-GE scaffolds. This effect can be attributed to the zinc release, reaching 5.5 ppm of Zn²⁺ at 2 days of assay, enough amounts to inhibit S. aureus growth (Fig. 3) [39]. In E. coli cultures, for unloaded BL-GE and 4ZN-GE scaffolds, bacteria growth was inhibited only the first 2 h until a 75% due to the calcium and zinc ions release. For BL-GE and 4ZN-GE LEVO and RIFAM loaded scaffolds, almost of 100% of bacteria growth inhibition was achieved. RIFAM was no tested against E. coli because it was reported that this antibiotic is not effective against these bacteria [40]. As Fig. 8 shows, the same antibiotics that managed to inhibit bacterial growth in the inhibition zone experiments, achieved an effective antibacterial effect for the planktonic growth of S. aureus and E. coli.

Fig. 8.

Planktonic growth of S. aureus and E. coli after 2, 6, 24 and 48 h of bacteria culture in contact with BL-GE and 4ZN-GE scaffolds, loaded and unloaded with: Levofloxacin (L), Vancomycin (V), Gentamicin (G) and Rifampicin (R). * indicates significant differences between 2, 6, 24 and 48 h. Statistical significance: p < 0.05.

3.8. Minimal Inhibitory Concentration (MIC) tests

Previous planktonic growth tests showed no significant differences between zinc doped and undoped scaffolds. Antibiotics were very effective against both strains and zinc effect was no appreciable. For this reason, we considered of interest to prove, in an antibiotic resistant scenario, what role zinc ions released from the scaffold could play. Thus, to test zinc antibacterial effect and its possible synergistic effect with antibiotics, the concentration of antibiotic loaded was reduced for each strain from the saturated one to the MIC of each drug (Fig. 9).

Fig. 9.

Antibiotic released from MIC loaded 4ZN-GE scaffolds at pH 7.4 after 6 and 24 h. Release of antibiotics from BL-GE was analogous in all the cases (data not shown).

Two different concentrations (27 mg/mL and 80 mg/mL) of BL-GE and 4ZN-GE scaffolds were loaded with the antibiotics to have the MIC of antibiotic for each strain released to the culture medium. Both bacteria strains were cultured 24 h in contact with the loaded scaffolds and its antibacterial effect was compared for both strains with the MIC effect for each antibiotic. Inhibition growth was achieved by all of them, compared with the MIC [41].

In S. aureus assays (Fig. 10) when concentration was 80 mg scaffold/ mL, 4ZN-GE loaded with antibiotic achieved a higher CFUs decrease at 24 h (except for LEVO) than antibiotic loaded BL-GE scaffolds. This fact was attributed to the synergistic effect of zinc ions and the antibiotics. On the other hand, when the concentration was 27 mg scaffold/mL, the synergistic effect at 24 h was demonstrated with the combinations of zinc-containing scaffolds with VANCO or with LEVO.

Fig. 10.

CFUs/mL of S. aureus and E. coli for 27 and 80 mg scaffold/mL concentrations after 24 h in contact with 4ZN-GE and BL-GE antibiotic loaded scaffolds and with MIC. *and # indicate significant differences between the different samples. Comparisons between: BL-GE and 4ZN-GE (*), MIC and BL-GE (#); Statistical significance: p < 0.05.

In E. coli assays (Fig. 10) for 27 and 80 mg scaffold/mL concentrations, 4ZN-GE and BL-GE scaffolds loaded with LEVO showed almost total capability to kill bacteria at 24 h. However, for GENTA-loaded scaffolds, a superior behaviour was achieved for 4ZN-GE scaffold indicating again a synergistic effect of zinc and GENTA.

3.9. Biofilm degradation

Finally, the antimicrobial effect of antibiotic loaded and unloaded BL-GE and 4ZN-GE scaffolds was investigated in a simulate infection environment containing S. aureus and E. coli biofilms. In Fig. 11, it is possible to appreciate the typical structure of a preformed biofilm which shows colonies of living bacteria (green) covered by a protective mucopolysaccharide matrix (blue) used as a control in both strains.

Fig. 11.

Confocal micrographs of S. aureus and E. coli biofilms before (control) and after 24 h in contact with BL-GE scaffolds, 4ZN-GE scaffolds and 4ZN-GE scaffolds loaded with each tested antibiotic

After contacting Gram + and - bacteria with the tested scaffolds, notable differences were observed in the biofilms. According to each case, they were partially or totally destroyed showing extracellular matrix in blue (calcofluor), alive bacteria in green (PI) and dead bacteria in red (SYTO 9).

Regarding S. aureus assays, the scaffolds BL-GE, 4ZN-GE and 4ZN-GE-G produced a partial destruction of the biofilm at 24 h. However, RIFAM, VANCO and LEVO loaded 4ZN-GE scaffolds reached complete biofilm destruction, observing colonial killed bacteria without the presence of protective layer of mucopolysaccharides.

Regarding E. coli assays, BL-GE, 4ZN-GE, 4ZN-GE-V and 4ZN-GE-R loaded scaffolds reached a partial destruction of the biofilm after 24 h. Nevertheless, LEVO and GENTA loaded 4ZN-GE scaffolds reached complete biofilm destruction.

These results show that our therapeutic ion-drug systems formed by the combination zinc together with GENTA or LEVO for E. Coli and RIFAM, VANCO or LEVO for S. aureus are very effective for the total destruction of the biofilm in the first 24 h of incubation, which is indicative of their antimicrobial effect. These results agree with inhibition zone assays (Fig. 7) and planktonic growth experiments (Fig. 8) where it was possible to confirm the antibacterial efficiency of same antibiotics for same strain [34–38]. When biofilms were in contact with the unloaded 4ZN-GE scaffolds a partial biofilm destruction was found in both strains which show zinc antibacterial effect by itself.

4. Conclusions

Zn²⁺ ions in combination with antibiotics, included in the mesoporous bioactive glass scaffolds increased the antibacterial effect of the system. In vitro studies of Zn-free (BL-GE) and Zn-containing (4ZN-GE) scaffolds with S. aureus and E. coli, showed a synergistic antibacterial effect of zinc with the antibiotics LEVO, VANCO, RIFAM or GENTA loaded in the scaffolds.

In planktonic assays, BL-GE and 4ZN-GE scaffolds loaded with saturated antibiotic concentration, exhibited total inhibition growth of S. aureus and E. coli cultures at 2 h. However, when scaffolds were loaded only with the MIC of antibiotics, zinc ions coming from 4ZN-GE scaffolds, showed synergistic antibacterial effect with LEVO and VANCO against S. aureus and with LEVO and GENTA against E. coli cultures.

In the inhibition growing zone tests, inhibition halos were achieved in S. aureus cultures at 18, 9 and 5 days with 4ZN-GE scaffolds loaded respectively with LEVO, RIFAM and VANCO. Nevertheless, in E. coli cultures, inhibition halos were maintained at 8 and 3 days with 4ZN-GE scaffolds loaded with LEVO and GENTA.

Partial biofilm destruction was reached with 4ZN-GE scaffolds, and total biofilm destruction with VANCO, LEVO and RIFAM 4ZN-GE loaded scaffolds for S. aureus and LEVO and GENTA 4ZN-GE loaded scaffolds for E. coli.

Therefore, the study demonstrated that the addition of antibacterial zinc ions inhibits planktonic bacterial growth and destroys biofilms with minimal antibiotic loaded concentration. There was an antibacterial effect of same antibiotics for same strains, potentiated by zinc. This approach reduces antibiotic resistance mechanisms eradicating bacteria in the surroundings of the implant site. These results provide significant insights for designing bone osteogenic implants capable of playing simultaneously a dual role against infection.