The insertion of titanium implants during orthopaedic surgery is a common and relatively safe procedure today. However, bacterial infections around an implant remain difficult to control. Staphylococcus aureus and several other strains of Staphylococcus are frequently associated with the colonization of orthopaedic implants and are responsible for subsequent infections. Peri-implant infections may be dependent on the microbial flora at the operative site, and hospital-acquired multi-drug-resistant bacteria play an increasing role in causing these infections1,2.
Antibiotics such as vancomycin, when covalently attached onto titanium surfaces, reduce in vitro colony formation of gram-positive bacteria3, but the prophylactic use of local antibiotics during implant placement is controversial. Treatment with antibiotics may trigger allergic reactions4 and could promote the induction of multi-drug-resistant bacteria5,6. Thus, the development of surfaces and coatings that can actively kill bacteria remains a challenge. Probably the oldest and most widespread coatings are silver ions, which have been successfully applied against methicillin-resistant Staphylococcus aureus (MRSA)7. However, a drawback of this approach is the cytotoxicity of silver ions toward mammalian cells8.
More recently, surface functionalizations such as adsorption, coating, or covalent binding with artificial antimicrobial peptides and several cationic, polycationic, and steroid antibiotics have been described9-12. The hydrophobic and positively charged side groups of these small molecules can interact with the microbial membrane, leading to its disintegration and the metabolic breakdown of the microbe. However, the production of such surfaces for medical use is still expensive. In addition, medical acceptability of these molecules is low because of their unknown interactions with mammalian cells. Therefore, a promising strategy to prevent infection around orthopaedic implants is the use of naturally occurring cationic antimicrobial peptides (CAMPs) for bioactive coatings.
CAMPs represent a part of the innate immune system and have a broad antimicrobial spectrum13. The cationic nature of CAMPs mainly contributes to their cell selectivity because the surface of bacterial membranes is more negatively charged than that of mammalian cells. Therefore, CAMPs exhibit a strong antimicrobial activity but no cytotoxicity to mammalian cells14. This class of antimicrobial peptides is active even against multi-drug-resistant bacteria such as MRSA. As the mechanism of their action differs from that of classical antibiotics, the formation of resistant microbes has not been observed15-18. The CAMP used in the present in vitro study is recombinant human ß-defensin-2 (rHußD2), a molecule that is expressed in human epithelial cells after stimulation by injury or different inflammatory factors19. Moreover, rHußD2 takes part in the recruitment of cells of the adaptive immune response in chemoattractant macrophages and mast cells promoting wound-healing20-23. Recombinant HußD2 exhibits antimicrobial activity against both gram-positive and gram-negative bacteria, fungi, and viruses24-26.
The purpose of the present study was to test the antimicrobial activity of rHußD2 coated on different functionalized titanium surfaces. Furthermore, the release behavior in relation to the antimicrobial properties of rHußD2 on two different collagen titanium scaffolds was investigated.
Titanium Pins
One hundred custom-designed oxidized titanium pins were used in the present study. Cylindrical pins measuring 1 mm in diameter and 5 mm in height were made from pure titanium (ASTM Grade 4) according to ISO (International Organization for Standardization) 5832-2. The surfaces were produced by means of a sandblasting process with corundum particles, followed by a strong acid-etching procedure (a mixture of HCl/H2SO4) at room temperature for several minutes. This led to fine micropits of 2 to 4 µm superimposed on the sand-blasted surface.
Preparation of the Titanium Surface
Prior to surface functionalization, all of the pins underwent ultrasonic treatment (Sonorex Super 10P; Bandelin, Berlin, Germany) for ten minutes in 5-M KOH, for ten minutes in 69% HNO3, and finally for fifteen minutes in a 2:1 H2SO4/H2O2 mixture at room temperature. Titanium pins were then extensively rinsed with distilled water. Next, the pins were incubated for one hour at 65°C in an oxidizing solution (NH4OH/H2O2/distilled water in a ratio of 1:1:1) and were stored for sixteen hours at 4°C in 70% ethanol.
Functionalization of the Titanium Surface
Prior to coating of rHußD2 on the titanium pins, the surfaces of the pins needed to be functionalized. Four different self-assembled monolayers (SAM1 through SAM4) were produced by means of direct silanization of the hydrophilized pin surfaces27. Thus, twenty titanium pins were incubated in 250 µL toluene consisting of a 10% solution of hexadecyltrimethoxysilane (SAM1), dimethoxymethyloctylsilane (SAM2), or allyltrimethoxysilane (SAM3). In addition, forty pins were incubated in 250 µL toluene consisting of a 10% solution of 3-aminopropyltrimethoxysilane (SAM4). Both silanization procedures were performed for twenty-four hours at room temperature. Then the pins were extensively rinsed with toluene and were dried at room temperature. Additional carboxylic acid (COOH) groups on the SAM3 titanium pins were generated by oxidation of the CH=CH2 functional group from allyltrimethoxysilane by incubation with 5% KMnO4 acid aqueous solution and finally were washed with distilled water.
Collagen Binding to the Functionalized Titanium Surface (SAM4)
To bind collagen in a stable manner onto the functionalized titanium oxide surface, the protein fibrils were cross-linked with the 3-aminopropyltrimethoxysilane (SAM4) layer in two different ways.
First, forty SAM4 titanium pins were pre-incubated in 200-mM 2-(N-morpholino)ethane sulfonic acid (MES) (pH 5.5) for one hour at room temperature. Twenty-five microliters of collagen solution from calf skin type I (0.1%, C8919; Sigma-Aldrich, St. Louis, Missouri) were applied by pipette to the pins, and the pins were dried at room temperature. Cross-linking with N-hydroxysuccinimide (NHS) and 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide HCl (EDC) was performed by incubating twenty pins in a solution of 10-mM NHS and 30-mM EDC in 200-mM MES (pH 5.5) for six hours at room temperature. The surface modification was referred to as SAM4::Col-NHS.
Second, the dried collagen fibrils on the SAM4 titanium pin surface were cross-linked by incubating the remaining twenty pins in a 25-µL glutaraldehyde solution (25%) for one hour at room temperature. Subsequently, the pins were rinsed in a 0.1% w/v bovine serum albumin (BSA) and 200-mM phosphate-buffered saline (PBS) solution for twenty minutes, followed by incubation in 100-mM Na2HPO4 for one hour, and finally were washed in distilled water, dried in a desiccator, and stored at 4°C. The surface modification was referred to as SAM4::Col-Glu. The successful binding of collagen was monitored with a collagen-specific stain after treatment with 0.1% w/v DirectRed 80 in 0.5% v/v acetic acid. All chemicals were purchased from Sigma-Aldrich if not described otherwise.
Coating of rHußD2 to the Functionalized Titanium Surfaces
Five different groups of functionalized titanium pins were produced and investigated in this study: three silanized titanium surfaces (SAM1, SAM2, SAM3) and two SAM4::collagen titanium surfaces (SAM4::Col-NHS, SAM4::Col-Glu). Recombinant human ß-defensin-2 (rHußD2) was produced by PLANTON (Kiel, Germany) as described previously25. Coating of this peptide was performed by directly adding 10 µL of a 1 mg/mL or an 8 mg/mL rHußD2 solution (in 0.01% acetic acid) onto the functionalized titanium surfaces. Finally, the pins were dried in a desiccator and were stored at 4°C prior to use.
In this way, two sets of SAM1, SAM2, and SAM3 layers were generated by coating either 10 µ g or 80 µ g of rHußD2 on each surface. The SAM4::collagen functionalized titanium surfaces were coated with 80 µ g of rHußD2.
Antibacterial Assays
The antibacterial activity of the rHußD2-coated pins was tested by means of a microdilution assay as described recently27, but with the following modifications. Escherichia coli strain DB3.1 (Cat. No. 11782-018; Invitrogen, Carlsbad, California) was used. Escherichia coli was streaked from a glycerol stock onto a Brain Heart Infusion (BHI) agar plate (3.7% w/v BHI bouillon, 1.5% w/v agar), was grown overnight at 37°C, and subsequently was used to inoculate 40 mL of 3.7% w/v BHI bouillon without any antibiotics in 100-mL flasks. Fresh cells were harvested by means of centrifugation and were washed twice with 10-mM sodium phosphate buffer (pH 7.2). The optical density was adjusted with 3.7% w/v BHI bouillon to 104 cells/mL. For antibacterial testing, a single pin coated with rHußD2 was incubated in one well of a microtiter plate containing 100 µL of the bacterial suspension.
Functionalized titanium pins without rHußD2 served as negative controls (SAM1-n, SAM2-n, SAM3-n, SAM4::Col-NHS-n, SAM4::Col-Glu-n). In addition, five independent control reactions consisting of rHußD2 in a final amount of 0 µg (NC), 0.02 µg (PC1), 0.2 µg (PC2), 2 µg (PC3), and 10 µg (PC4) were examined in parallel. To determine the specificity of the antibacterial activity of rHußD2, 10 µL of polyclonal monospecific anti-rHußD2 antibodies (10 µg/mL in PBS) were co-incubated with 10 µg of rHußD2 and bacterial cells (PC4/ab) (Table I).
A single microtiter plate for each experiment, including the control reaction and the five different functionalized pins, was incubated for two hours at 37°C. The kinetic studies were performed by replacing the bacterial suspension, after two hours of cultivation, with a freshly precultured Escherichia coli solution, which contained the same bacterial concentration, and cultivating for another two hours, and so on. Colony-forming units (CFU) were determined after plating 100 µL of a 1:100 dilution of the bacterial suspension on Petri dishes with BHI medium and incubating them overnight at 37°C. For every group of the functionalized titanium pins and the controls, five independent antibacterial experiments were performed. A control (NC) of bacterial growth was examined for every independent assay. The number of colony-forming units on the control (consisting of a single Petri dish) was set to 0% bactericidal activity. The reduction in the colony-forming units resulting from the antibacterial activity of rHußD2 was related to the absolute colony-forming units of the control within every independent assay. This value, given as a percentage of the bacterial "killing" activity, was used to describe the antibacterial activity. Normalization of the antibacterial data was performed for every experimental group.
Quantification of rHußD2 with ELISA
In a parallel reaction, the amount of rHußD2 released into the BHI medium was determined by means of an enzyme-linked immunosorbent assay (ELISA). BHI medium was prepared as described above without bacterial cells. Titanium pins were incubated in the BHI medium for two hours at 37°C, corresponding to the antibacterial assay. The same set of controls was examined as described above. Microtiter plates were coated with 50 µL of polyclonal monospecific rabbit anti-rHußD2 antibodies (100 ng/mL) in 0.05-M Na2CO3 (pH 9.6) and then were blocked with 0.5% BSA in PBS for two hours at 37°C. After each step, the wells were washed three times with PBS (pH 7.4) containing 0.1% v/v Tween 20. Fifty microliters of the sample, consisting of BHI medium that contained either the titanium pins or the control reactions, diluted in PBS/0.1% v/v Tween 20, was added in duplicate and incubated for sixty minutes at 37°C. After washing, 50 µL of the rabbit biotinylated anti-rHußD2 antibody (100 ng/mL) was added, followed by incubation for sixty minutes. After washing, horseradish peroxidase (HRP)-conjugated streptavidin was added and incubated for forty-five minutes at 37°C. Finally, the reaction was visualized by adding 50 µL of tetramethylbenzidine (TMB) substrate for ten to twenty minutes. The reaction was stopped with 0.5-M H2SO4, and the absorbance was determined at 450 nm with use of an ELISA plate reader. As a reference for quantification, a standard protein curve was established by a serial dilution of rHußD2 (45 pg/mL to 1 µg/mL).
Source of Funding
The present study was supported by the European Union, MyJoint, FP6-NEST-028861. The funds were used for staff salaries and supplies.
To determine the antibacterial activity of functionalized and subsequently coated titanium surfaces, an antibacterial assay was established. The antibacterial activity data were generated by normalizing the colony-forming units of every independent assay to the control assay. The resulting percentage of the killing activity of five independent assays was averaged. Thus, the variability of the colony-forming units of the individual antibacterial assays was reduced.
The results confirmed the efficacy and the specificity of the recombinant human ß-defensin-2 (rHußD2) as an antibacterial agent (Table I). The bacteria were completely killed at a concentration of 10 µg of rHußD2. In contrast, the entire bactericidal activity was blocked when a specific anti-rHußD2-antibody was added to the assay. The amount of rHußD2 detected by means of ELISA in the control assays (NC, PC1 through PC4, and PC4/ab) corresponded with the amount initially applied to the anti-bacterial assay (Table I).
Titanium oxide surfaces were functionalized by self-assembled monolayers (SAM) of common silanes. In the first experiment, titanium pins functionalized by hexadecyltrimethoxysilane (SAM1), dimethoxymethyloctylsilane (SAM2), and oxidized allyltrimethoxysilane (SAM3) were coated with 10 µg of rHußD2. The SAM surfaces alone (SAM1-n, SAM2-n, SAM-3n) did not exhibit any antimicrobial activity prior to coating with rHußD2 (Table I). In contrast, functionalized titanium pins coated with rHußD2 (SAM1, SAM2, SAM3) exhibited an antibacterial activity killing rate of >90% (Table I). The biological activity of the coated titanium pins corresponded with the positive rHußD2 control (PC4) with the highest concentration (10 µg) and demonstrated successful coating of the pins with rHußD2. Furthermore, the ELISA data presented in Table I confirmed (1) that coating with rHußD2 of the functionalized pins was nearly 100%, (2) that biologically active rHußD2 was eluted from the pins, and (3) that the discrepancy between the amount of rHußD2 applied for coating and the amount eluted from the pins indicated a storage function of the SAM layers for rHußD2.
To examine the release behavior of rHußD2 from the different functionalized titanium pins, a time kinetic experiment was performed. As observed in the previous experiment, coating with 10 µg of rHußD2 was not sufficient to achieve complete bacterial killing, so the titanium pins were coated with an excess of rHußD2 (80 µg). Thus, a high antimicrobial activity with a killing rate of 100% was observed for all SAMs after two hours of incubation in the antibacterial assay (Fig. 1). ELISA quantification revealed that, at this time point, about 74 to 79 µg of rHußD2 was eluted from the different titanium pins into the medium (Table II). The time-dependent cultivation of the titanium pins was performed by incubating the same pin after the first two-hour cultivation in a freshly prepared Escherichia coli culture containing the same amount of bacteria for another two hours, and so on. After the second cultivation step (four hours), the antimicrobial activity still reached killing rates of 52% (corresponding to 0.41 µg of rHußD2 in the medium) to 69% (corresponding to 1.29 µg of rHußD2 in the medium) (Table II). At this time point, a stronger killing activity of rHußD2 of the SAM2 pins compared with the other functionalized surfaces could be observed (p < 0.001) (Fig. 1). After six hours, the killing rate for SAM2 was still 60% (corresponding to 0.65 µg of rHußD2 in the medium), that for SAM1 was about 29% (corresponding to 0.08 µg of rHußD2 in the medium), and that for SAM3 dropped significantly, below 5% (p < 0.001) (Fig. 1). These findings indicated that different functionalized titanium pins exhibited a different elution profile in a defined time span.
Bar graph showing the kinetics of rHußD2 release from silanized titanium surfaces as a percentage of bacteria killed. The antimicrobial activity of different silanized titanium pins coated with 80 µg of rHußD2 (hexadecyltrimethoxysilane [SAM1, black bars], dimethoxymethyloctylsilane [SAM2, fasciated bars], and oxidized allyltrimethoxysilane [SAM3, white bars]) after two, four, six, and eight hours of cultivation in an Escherichia coli suspension is shown. Bacterial killing activity is shown as the mean percentage of five independent antimicrobial assays, along with the standard error of the mean. The test of significance was performed with the Student t test. An asterisk indicates a significant difference (p < 0.001) between groups.
As it is known that collagen has a positive impact on wound-healing, functionalized titanium pins with collagen were investigated. Two different cross-linking strategies for collagen on SAM4 were applied: (1) a covalent binding strategy with the NHS/EDC cross-linking system, and (2) the use of glutaraldehyde. Successful binding of collagen was monitored with collagen-specific DirectRed staining. Functionalized collagen titanium pins without rHußD2 did not show any antibacterial activity (Fig. 2). After two hours of incubation, pins of both cross-linking strategies generated killing rates of 100% (Fig. 2). However, ELISA quantification of these pins revealed a release of only 14 to 17 µg of rHußD2 into the medium (Table II). After four hours of cultivation, the killing rate dropped to 53% (corresponding to 0.46 µg of rHußD2 in the medium) for the NHS/EDC system whereas collagen pins treated with glutaraldehyde exhibited a killing rate of almost 92% (corresponding to 6.6 µg of rHußD2 in the medium). After six hours, the antimicrobial activity of both systems dropped below 10% of the killing rate and no rHußD2 was detectable in the medium with use of ELISA.
Bar graph showing the antimicrobial activity of collagen titanium pins coated with rHußD2 as a percentage of bacteria killed. The antimicrobial activity of two different collagen titanium pins coated with 80 µg of rHußD2 (SAM4::Col-Glu [black bars] and SAM4::Col-NHS [white bars]) after two, four, six, and eight hours of cultivation in an Escherichia coli suspension is shown. Bacterial killing activity is shown as the mean percentage of five independent antimicrobial assays, along with the standard error of the mean. The test of significance was performed with use of the Student t test. An asterisk indicates a significant difference (p < 0.001) between groups.
Previous reports have shown that recombinant human ß-defensin-2 (rHußD2) exhibited antibacterial activity at a minimal bactericidal concentration of >10 mg/L against gram-positive bacteria28,29 and at a minimal bactericidal concentration of <12 mg/L against gram-negative bacteria25. These findings make rHußD2 a very good candidate as a naturally occurring antimicrobial peptide for bioactive implant coating. The antibacterial tests in the present study were performed with use of a sensitive indicator organism, Escherichia coli, in order to establish the concept of an effective antimicrobial drug delivery system using rHußD2 on titanium surfaces. The results corroborated the efficient antibacterial activity of rHußD2 and, in addition, showed its bioactivity when coated on metallic surfaces. The present study demonstrated that titanium surfaces functionalized by different SAM (self-assembled monolayers) could store and deliver antimicrobially active rHußD2.
The aliphatic SAM layers (SAM1 and SAM2) and also the negatively charged SAM3 layer could store and release sufficient rHußD2 for antimicrobial activity in vitro (Table I). However, there were differences between the three functionalized surfaces. The rHußD2 coating exhibited continuous release of rHußD2 for several hours (Fig. 1). Hexadecyltrimethoxysilane (SAM1) and dimethoxymethyloctylsilane (SAM2) differ in the length of their exposed aliphatic chains. The aliphatic chain of dimethoxymethyloctylsilane (SAM2) is shorter and might exhibit a better rHußD2 delivery system, as indicated by an antimicrobial activity of 60% killing after three cultivation steps (six hours). The other hydrophobic layer (SAM1) exhibited a killing rate of only 29%.
Both hydrophobic surfaces (SAM1 and SAM2) exhibited a higher capacity for storage and peptide release compared with the SAM3 pins (Fig. 1). In addition, the functionalized SAM3 surface exhibited the fastest decrease of activity over time. After six hours, almost no killing of Escherichia coli was observed and no rHußD2 was detected with ELISA (Fig. 1 and Table II). The reason for this finding may be an ionic interaction between the SAM3 surface and rHußD2. However, excessive and unbound rHußD2 was released immediately, resulting in the delivery of most of the rHußD2 during the first two hours of cultivation.
Bacterial contamination of implants and subsequent infections often occur immediately after device implantation. Recent protocols have recommended short-term antibiotic prophylaxis30,31. The silane-functionalized titanium surfaces enable the immediate release of rHußD2 in a short time in vitro, which may prevent early postoperative infections around the implants in clinical use.
A positive effect of tissue regeneration occurred when surfaces of biomaterials were treated with collagen. Morra et al.32 reported that biochemical alteration of titanium surfaces by collagen can increase the healing rate. From this point of view, an experiment was designed to cover the titanium specimen with bovine type-I collagen prior to rHußD2 coating.
The direct comparison of collagen-treated surfaces and pure silane monolayers revealed that less rHußD2 was eluted into the medium from collagen pins (Table II). This effect might be associated with a strong and irreversible binding of rHußD2 onto the collagen matrix. Within the first two hours of the antibacterial assay of the collagen titanium pins, only 14 to 17 µg of rHußD2 was detected with ELISA, which led to a complete killing of the bacteria. However, after the second cultivation step (four hours), a significant difference was observed between rHußD2 release from the glutaraldehyde cross-linked collagen surface (SAM4::Col-Glu) and the other collagen surface (SAM4::Col-NHS) (Table II and Fig. 2) (p < 0.001). SAM4::Col-Glu delivered about 6.6 µg of rHußD2 and caused 92% of bacterial killing. In contrast, SAM4::Col-NHS released only 0.46 µg of rHußD2 with an antibacterial activity of 53% (Table II and Fig. 2).
It was not possible to determine the bioactive rHußD2 remaining in the collagen matrix, but we speculate that as a consequence of gradual collagen catabolism in the surrounding tissue in vivo, bioactive rHußD2 might be slowly released. On the basis of this hypothesis, a peptide delivery system that will yield defined antimicrobial dosages in vivo could be developed in the future by using different collagen binding systems.
Various approaches have been described for bioactive coating onto titanium surfaces3,10,11,33,34. However, most antibiotic substances are difficult to apply because of their toxicity, their putative promotion of bacterial resistances, or their lack of availability. Here, we demonstrated, for the first time, that the use of a recombinant human ß-defensin-2 as an antimicrobial substance is suitable for bioactive coating of titanium surfaces.
Future studies will be performed to investigate these types of antibacterial coatings on orthopaedic implants and their effects on osseointegration in vivo.