Gentamicin

Effects of gentamicin and gentamicin–RGD coatings on bone ingrowth and biocompatibility of cementless joint prostheses: An experimental study in rabbits
Volker Alt a,⇑, Achim Bitschnau a, Felicitas Böhner a, Katharina Elisabeth Heerich a, Erika Magesin a, Andreas Sewing b, Theodoros Pavlidis a, Gabor Szalay a, Christian Heiss a, Ulrich Thormann a,
Sonja Hartmann c, Wolfgang Pabst d, Sabine Wenisch c, Reinhard Schnettler a
aDepartment of Trauma Surgery, University Hospital of Giessen-Marburg GmbH, Campus Giessen, 35385 Giessen, Germany
bBiomet Deutschland GmbH, Gustav-Krone Straße 2, 14167 Berlin, Germany
cLaboratory of Experimental Trauma Surgery, Justus-Liebig-University Giessen, 35394 Giessen, Germany
dInstitute of Medical Computer Sciences, Justus-Liebig-University Giessen, 35394 Giessen, Germany

a r t i c l e i n f o

Article history:
Received 9 September 2010
Received in revised form 4 November 2010 Accepted 10 November 2010
Available online 1 December 2010

Keywords: Gentamicin Hydroxyapatite Bone formation Arthroplasty Biocompatibility
a b s t r a c t

Antimicrobial coatings are of interest as a means to improve infection prophylaxis in cementless joint arthroplasty. However, those coatings must not interfere with the essential bony integration of the implants. Gentamicin–hydroxyapatite (gentamicin–HA) and gentamicin–RGD (arginine–glycine–aspar- tate)–HA coatings have recently been shown to significantly reduce infection rates in a rabbit infection prophylaxis model. The purpose of the current study was to investigate the in vitro elution kinetics and in vivo effects of gentamicin–HA and gentamicin–RGD–HA coatings on new bone formation, implant integration and biocompatibility in a rabbit model. In vitro elution testing showed that 95% and 99% of the gentamicin was released after 12 and 24 h, respectively. The in vivo study comprised 45 rabbits in total, with six animals for each of the gentamicin–HA, gentamicin–RGD–HA group and control pure HA coating groups for the 4 week time period, and nine animals for each of the three groups for the 12 week observation period. A 2.0 mm steel K-wire with one of the coatings under test was placed in the intra- medullary canal of the tibia. After 4 and 12 weeks the tibiae were harvested and three different areas (proximal metaphysis, shaft area, distal metaphysis) were assessed by quantitative and qualitative histol- ogy for new bone formation, direct implant–bone contact and the formation of multinucleated giant cells. The results exhibited high standard deviations in all subgroups. There was a trend towards better bone formation and better direct implant contact in the pure HA coating group compared with both gentami- cin coatings after 4 and 12 weeks, which was, however, not statistically significant. The number of mul- tinucleated giant cells did not differ significantly between the three groups at both time points. In summary, both gentamicin coatings with 99% release of gentamicin within 24 h revealed good biocom- patibility and bony integration, which was not statistically significant different compared with pure HA coating. Limitations of the study are the high standard deviation of the results and the limited number of animals per time point.
ti 2010 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

1.Introduction

Infections of total joint prostheses remain a devastating prob- lem in arthroplasty and all effort should be made to optimize infec- tion prophylaxis. In primary cemented hip arthroplasty the use of antibiotic-loaded bone cements has clearly shown positive effects with a reduction in post-operative infections rate in the Norwegian Arthroplasty Register [1,2]. This principle of local delivery of anti- biotics by bone cements was introduced by Buchholz in total joint arthroplasty in 1970 [3] and with the treatment of bone infections

with gentamicin-loaded polymethylmetacrylate beads by Klemm in 1979 [4].
In uncemented arthroplasty this principle of local delivery of antibiotics to improve infection prophylaxis is more complex. Coating of the surface of the implants with an antimicrobial agent is one option in that context. Two major preconditions of the coat- ings must be fulfilled. First, the coating has to exhibit good antimi- crobial properties. Second, it should not negatively interfere with the decisive process of osteointegration of the uncemented im- plants, which determines the long-term survival in arthroplasty. Regarding antimicrobial activity, there have been several studies

⇑ Corresponding author. Tel.: +49 641 99 44 601; fax: +49 641 99 44 609. E-mail address: [email protected] (V. Alt).
reporting on different successful antibiotic strategies for implants in general [5–18]. Hydroxyapatite (HA) was shown to have a

1742-7061/$ – see front matter ti 2010 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.actbio.2010.11.012

positive influence on implant integration and implant survival in clinical settings [19,20]. Furthermore, arginine–glycine–aspartate (RGD) coatings have been the focus of several reports on improving the bony integration of orthopaedic implants [21–28]. Studies by Elmengaard et al. [26] and Rammelt et al. [27] showed significant increases in new bone formation compared with control implant surfaces without RGD.
Alt et al. [5] published a study on the antimicrobial effects of gentamicin–HA and gentamicin–RGD–HA coatings in a rabbit infection prophylaxis model which showed a statistically signifi- cant reduction in post-operative infection rates compared with antibiotic-free control implants. Antibiotics [29–32], including gentamicin, have been reported to impair the functioning of oste- oblasts and osteocytes that could possibly influence the bony inte- gration of implants. To the authors best knowledge there are no data available on the bony integration, new bone formation and biocompatibility of implants with gentamicin–HA and gentami- cin–RGD–HA coatings. The purpose of the current study was to evaluate the in vitro elution kinetics and new bone formation, im- plant integration and biocompatibility of gentamicin–HA and gen- tamicin–RGD–HA coating in a rabbit model. The hypothesis was that both gentamicin coatings will not show any difference com- pared with a pure HA coating regarding the three above mentioned criteria.

2.Materials and methods

2.1.Implants and coating technique

Gentamicin–HA- and gentamicin–RGD–HA-coated K-wires were tested. Pure HA-coating served as negative control group. All implants had a diameter of 2 mm and a length of 11 cm with a grit blasted surface (corundum, Ra = 2.7 lm). The HA coating was applied by an electrochemically assisted process onto the steel surface of all K-wires producing a 5 lm layer of HA on the implant [33]. In brief, HA was precipitated onto the implants from a Ca2+/
HxPO4(3 ti x)ti -containing electrolyte at 37 ti C and a pH value of 6.4 by cathodic polarization of the sample. This layer consisted of a dense but porous network of needle-like HA crystallites typically 300 nm in length and 60 nm in diameter and with a crystallinity of 72%.
For the additional RGD coatings cyclic RGD peptide highly selec- tive for avb3 and avb5 integrin receptors was used. Anchorage of RGD onto HA was by specific connection with phosphonate group as described previously [24,34].
Additional gentamicin coating at a concentration of 250 lg cmti2 active gentamicin was applied on both the HA- and RGD–HA-coated implants as previously reported for the assess- ment of the antimicrobial activity of gentamicin coatings [5]. Briefly, the layer consisted of rapidly released gentamicin sulphate (Fujian Fukang Pharmaceutical, Fuzhou, China) and more pro- longed release gentamicin crobefate (Merck KGaA, Darmstadt, Ger- many) that were placed on the implant surface using inkjet technology. The amounts of the two gentamicin salts were chosen in such a way as to achieve half of the total release of active gen- tamicin from each of the components, which corresponds to a weight fraction of crobefate to sulphate salt of around 7/2. All sam- ples were sterilized by c-irradiation at 25 kGy.

2.2.In vitro elution kinetics

The kinetics of gentamicin release from the sample surfaces was tested in an in vitro experiment. The K-wires were incubated in 15 ml of phosphate-buffered saline (PBS) at 37 tiC. The PBS elution buffer was exchanged every hour up to 12 h and after 24 and 48 h.

Measurements were taken in triplicate. The gentamicin concentra- tions were determined by an agar diffusion zone of inhibition test with respect to a defined gentamicin standard.

2.3.Study design

All animal experiments were approved by the responsible local animal care committee (RP Thüringen, Erfurt, Germany, registra- tion No. 14-03/04). There were 45 rabbits in total, with six animals each for the gentamicin–HA, gentamicin–RGD–HA and control pure HA coating groups for the 4 week time period and nine ani- mals for each of the three groups for the 12 week observation period.
2.0 mm steel K-wires with the different coatings were placed in the intramedullary canals of rabbit tibiae. After 4 and 12 weeks the tibiae were harvested and three different areas (proximal metaph- ysis, shaft area, distal metaphysis) were assessed by quantitative and qualitative histology for new bone formation, direct im- plant–bone contact and the formation of multinucleated giant cells.

2.4.Surgical procedure

All procedures were undertaken under general anaesthesia with ketamine (60 mg kg body wtti1), xylazine (6 mg kg body wtti1) and atropine (0.1 mg kg body wtti 1). The test animals had similar body weights of approximately 3 kg.
The right lower limb was disinfected and the knee region was then draped in sterile gauze. A 1 cm incision over the tibial tuber- osity was followed by longitudinal splitting of the patellar tendon. The tibial tuberosity was then penetrated such that the K-wire could subsequently be introduced into and pushed into the distal part of the intramedullary canal. The overlapping proximal end of the K-wire was then was cut at the level of the tibial tuberosity to avoid interference with the surrounding soft tissue. Wound clo- sure was performed in multiple layers followed by post-operative X-ray to ensure correct positioning of the device.
Biomechanically stable conditions allowed for full load bearing of the operated limb. The animals were kept in double cages with free diet and water ad libitum.
The observation period was either 4 or 12 weeks and the tibiae were removed after the animals had been sacrificed. The K-wires were left in the intramedullary canal of the tibia for histomorpho- metrical and histological assessment of direct implant–bone con- tact and new bone formation on the implant.

2.5.Histomorphometry and histology

Three different parts of the tibiae (proximal metaphyseal area, shaft region, distal metaphyseal area) with the device in place were cut using a diamond bone saw for quantitative and qualitative assessment of new bone formation, direct bone–implant contact and biocompatibility as reported previously [28]. Briefly, after explantation the tibiae were placed in 4% paraformaldehyde solu- tion for 24 h. Five micrometers of transverse sections of the three different areas of the tibia with the K-wire in place were prepared using a grinding technique [35] for toluidine blue and hematoxylin and eosin staining.

2.5.1.Histomorphometry
Histomorphometrical assessment was performed on the three different regions of the harvested tibiae, including the transverse sections of both metaphyseal regions and a longitudinal section of the shaft region, using a stereo microscope (Stemi SV 11ti, Carl Zeiss, Jena, Germany) and digital image software (Media Cybernet- ics, Silver Spring, MD) as reported previously [28]. The regions of

interest (ROI) for new bone formation were defined as the area 1.5 and 1.2 mm around the implant excluding the area of the implant for the proximal and distal metaphyseal sections, respectively. For the longitudinally sectioned shaft region the ROI was defined as a distance of 1 mm from the centre of the implant excluding the area of the implant. The results are given as percentage new bone for- mation relative to the entire surface area by dividing the area of newly formed bone by the respective ROI surface area. Regarding direct implant–bone contact, the length of newly generated bone on the surface of the implant was related to the entire length of the implant.
For biocompatibility assessment quantitative evaluation of multinucleated giant cells was done using hematoxylin and eosin stained longitudinal sections of the shaft area [28]. The surface

100
95
90
85
80
75
70
65
60
55
50

0 2 4 6 8 10 12 14 16 18 20 22 24

area between the implant and cortical bone was determined with a grid micrometer (Carl Zeiss, Jena, Germany) in the ocular of a microscope (Axioskop 2 Plus, Carl Zeiss, Jena, Germany). Newly generated bone was excluded from the measurement of the surface area. Only complete grids of the micrometer were considered. Then all cells containing more than three nuclei and not adhering to bony surfaces were counted at 200-fold magnification. The number of multinucleated cells was divided by the surface area.

2.5.2.Histology
Detailed histological evaluation was carried out with a focus on the appearance of newly formed bone, foreign body cells, bone marrow necrosis, integrity of the bone marrow and cortical bone structures.

2.6.Statistical analysis

SPSS for Windows (v. 11.5) for one way analysis of variance was used for statistical analysis allowing direct comparison between each implant and each time point. The Kruskall–Wallis H test and the corresponding multiple comparison test were applied as a normal distribution could not be assumed when testing the assumption there is no difference between the groups concerning the parameter of interest. Data analysis was carried out in an exploratory sense, seeking to formulate hypotheses to be tested. Therefore, the computed P values are used as a tool to weight the evidence. Statistical significance was considered at P < 0.05. For all differences with P < 0.20 the exact P value is given in the Section 3.
Post hoc power analysis was conducted using G⁄Power 3 analysis (http://www.psycho.uni-duesseldorf.de/abteilungen/aap/
gpower3).

3.Results

3.1.In vitro elution kinetics

The in vitro elution kinetics revealed an initial burst release of around 65% of the gentamicin within the first elution step of 1 h, followed by slower release of an additional amount of 30% up to
Time [h]

Fig. 1. Elution kinetics of gentamicin from the implant surface at a concentration of 250 mg cmti 2. The mean standard deviation of all triplicate measurements is ±8% at each time point.

the elution kinetics. The mean standard deviation of all triplicate measurements was ±8% at each time point.

3.2.New bone formation and direct implant–bone contact

3.2.1.New bone formation after 4 weeks
Overall, there was only weak bone formation in all groups with no significant differences between the three different coatings for any of the three anatomical regions of the tibia (Table 1). The high- est bone formation rate was achieved by for HA-coated implants for the proximal metaphysis and the shaft region. Gentamicin– HA-coated K-wires were shown to have the strongest new bone formation in the distal metaphysis. However, none of the differ- ences reached statistical significance. High standard deviations were seen in all groups.
The detailed histology showed that new bone formation origi- nated from the endosteal surface and started to grow towards the implant for all implants. The newly formed bone consisted mainly of woven bone with only marginal direct contact with the implant surface in most cases (Fig. 2a) for the different groups. Multiple neighbouring osteoblasts with active osteoid production were identified.

3.2.2.Direct implant–bone contact after 4 weeks
Measurement of direct implant–bone contact as an indicator of bony integration of the implant did not show any statistically sig- nificant differences between the three different groups (Table 2).

Table 1
New bone formation in the region of interest (ROI) after 4 and 12 weeks, respectively.
New bone formation (% of ROI)
4 weeks 12 weeks Proximal metaphysis

12 h, when 95% of the total release is reached (Fig. 1). After 24 h 99% of the gentamicin was eluted from the K-wires. The observed elution kinetics result from immediate release of the water soluble gentamicin sulphate and more prolonged release of gentamicin crobefate, which is a 6-hydroxy-40 -methoxy-3-(p-methoxy-ben- zylidene)-flavanone phosphate gentamicin salt with low water sol- ubility. Gentamicin is released from the crobephate salt by degradation of the flavone complex. There is no additional carrier substance applied with the coating and consequently no diffusion processes are involved, even though the release kinetics are com-
HA Gentamicin–HA
Gentamicin–RGD–HA Shaft region
HA Gentamicin–HA
Gentamicin–RGD–HA Distal metaphysis
HA Gentamicin–HA
Gentamicin–RGD–HA
0.79 ± 1.07 0.42 ± 0.48 0.12 ± 0.08

3.97 ± 3.73 3.05 ± 1.34 1.78 ± 1.64

0.59 ± 0.68 7.33 ± 12.3 4.79 ± 4.20
35.8 ± 22.0 27.3 ± 17.3 22.6 ± 15.8

12.2 ± 12.3 7.45 ± 4.97
6.0 ± 6.23

25.4 ± 14.0 4.36 ± 2.65 11.8 ± 11.5

parable with a diffusion process. Gamma sterilization did not alter There were no significant differences between the three different implant types.

Fig. 2. New bone formation (gentamicin–HA implant) after 4 weeks mainly consists of woven bone with active osteoblasts (a), whereas after 12 weeks (gentamicin– RGD implant) mainly lamellar bone with osteocytes appears (b). Toluidine blue staining. Scale bar 50 lm.

Table 2
New bone formation on the implant surface in the shaft region after 4 and 12 weeks, respectively.
New bone formation on implant surface (% of implant length)
4 weeks 12 weeks
HA 7.89 ± 6.8 33.8 ± 12.8

Fig. 3. Histological overview of new bone formation and implant bone contact after 12 weeks at the proximal metaphysis, shaft area and distal metaphysis. Hematox- ylin and eosin staining. Scale bar 1 mm.

highest new bone formation in all regions. Gentamcin–HA- and gentamicin–RGD–HA-coated implants showed similar new bone formation rates for the proximal metaphysis and the shaft. In the distal part of the tibia the gentamicin–RGD–HA group showed bet- ter bone formation than the gentamicin–HA group, however, the difference was not statistically significant. The distal metaphyseal region showed a strong trend for impaired new bone formation for the gentamicin–HA group compared with the pure HA coating (P = 0.05). However, statistical significance between the groups was not reached.
In comparison with the results for the 4 week observation per- iod, there was significantly stronger new bone formation seen for all three implants after 12 weeks. For the proximal metaphysis sta- tistically significantly greater bone formation was found after 12 weeks for the HA (P = 0.003), gentamicin–HA (P = 0.003) and gentamicin–RGD–HA (P = 0.01) groups compared with after 4 weeks. New bone formation was seen in the shaft area and the distal part of the tibia for all groups. However, statistically signifi- cant differences were reached only for the HA group in the shaft re- gion (P = 0.001).
Histologically, there was not only immature woven bone, as after 4 weeks, but also lamellar bone with all three different im-

Gentamicin–HA 11.0 ± 7.38
Gentamicin–RGD–HA 3.87 ± 4.5
24.7 ± 19.0 25.1 ± 19.3
plant types, indicating maturation and remodelling of the initially formed bone with distinct direct bonding with the implant surface

There were no significant differences between the three different implant types.

Values ranged between 3.87% for the gentamicin–RGD–HA group and 11.0% for the gentamicin–HA group with 7.89% for the HA- coated implants with high standard deviations in all groups.

3.2.3.New bone formation after 12 weeks
As seen for the 4 week observation period, there were no statis- tically significant differences between the three different implant types for any of the assessed three anatomical areas of the tibia (Table 1 and Fig. 3). HA-coated K-wires were found to have the
(Fig. 2b).

3.2.4.Direct implant–bone contact after 12 weeks
After 12 weeks there were similar results for direct implant– bone contact as for new bone formation after 12 weeks, with worse results for the gentamicin implants compared with the HA control group, which, however, were not statistically significant. The addi- tional RGD coating did not show any statistically significant differ- ence compared with the gentamicin–HA coating (Table 2).
For each of the three groups direct implant–bone contact was much greater after 12 weeks compared with the results after

4 weeks, which reached statistically significant differences for the pure HA coating (P = 0.001) and the HA–gentamicin–RGD coating (P = 0.04). The difference between 4 and 12 weeks for the gentami- cin–HA group was not statistically significant (P = 0.13).

3.2.5.Post hoc power analysis
G⁄Power 3 analysis showed a power of approximately 20% for all eight tested subgroup scenarios (two time points, 4 and 12 weeks, and four different anatomical locations, proximal metaphysis, shaft region, distal metaphysis, implant surface) for the results for new bone formation.

3.3. Number of multinucleated giant cells and biocompatibility after 4 and 12 weeks

There were neither qualitative nor quantitative differences regarding the appearance of multinucleated giant cells or other signs of histological biocompatibility between the three different implant groups. The number of multinucleated giant cells did not show any statistically significant differences after 4 or 12 weeks (Table 3). It remained fairly constant for the pure HA and the gen- tamicin–RGD–HA groups and decreased for the gentamicin–HA implants between weeks 4 and 12, however, the differences were not statistically significant.
The three implant types exhibited different areas of bone mar- row with different cellular reactions after 4 weeks. The major part of the bone marrow showed normal formation of adipocytes with intact haematopoietic cells. However, oedematic regions with ne- crotic areas of both adipocytes and haematopoietic stem cells were also found (Fig. 4). In those regions multinucleated giant cells could be detected, which were not seen in intact regions of the bone marrow and which seemed to be involved in phagocytosis of the necrotic tissue.
After 12 weeks the size of the necrotic bone marrow regions was reduced, however, it remained in all groups, with the typical appearance of multinucleated giant cells as found at 4 week most likely still involved in the degradation of necrotic tissue.
Fig. 4. Appearence of multinucleated giant cells (arrows) in the bone marrow of all

4.Discussion

Each measure to improve infection prophylaxis in total joint arthroplasty helps to reduce the risk of devastating periprosthetic infections. Antibiotic coating is one interesting approach in this context. At the present time being no antibiotic coating is clinically available for primary cementless total joint arthroplasty. A previ- ous work has shown a significant reduction in post-operative infec- tion rates using gentamicin–HA and gentamicin–RGD–HA coatings in a rabbit infection prophylaxis model compared with a pure HA coating [5].
The purpose of the current study was to compare new bone for- mation, implant integration and biocompatibility of two gentami- cin–HA and gentamicin–RGD–HA coatings with a pure HA coating. Observation times of 4 and 12 weeks were chosen. A 4 week period has been used by other authors [15] and is regarded as the time

Table 3
Number of multinucleated giant cells in the shaft region after 4 and 12 weeks, respectively.
Number of multinucleated giant cells per mm2 4 weeks 12 weeks
HA 3.33 ± 2.36 2.93 ± 1.52
Gentamicin–HA 5.58 ± 6.11 2.93 ± 1.52
Gentamicin–RGD–HA 3.76 ± 3.71 3.88 ± 4.13
There were no significant differences between the three different implant types.
three implant types after 4 weeks. Hematoxylin and eosin staining. Scale bar 100 lm.

period required for initial bone ingrowth and initial bony fixation of the prosthesis. The time point of 12 weeks was added as further data on medium term bony integration and new bone formation is of interest.
Neither the gentamicin–HA nor gentamicin–RGD–HA coatings showed any significant differences regarding new bone formation, direct implant–bone contact and the appearance of multinucle- ated giant cells compared with the gentamicin-free HA control implants in the present study. Furthermore, there were no signif- icant differences between the gentamicin–HA and gentamicin– RGD–HA groups. However, there was a general trend towards im- paired new bone formation after 4 and 12 weeks and direct im- plant–bone contact after 12 weeks for the gentamicin-coated implants compared with the HA implants. This trend almost reached statistical significance for new bone formation in the dis- tal metaphysis after 12 weeks for the HA group compared with the gentamicin–HA group (P = 0.05). Post hoc power analysis showed a power of approximately 20% for the results for new bone formation, mainly due to the limited number of animals and the high standard deviation in each group, which limits the study and the conclusions that can be drawn. A further weakness is that in general results from animal trials cannot be directly transferred to the clinical situation. The main difference between this animal model and the human situation of cement-free

implantation of a total hip or knee joint is that the current animal model only uses an intramedullary device without joint replace- ment and therefore biomechanical loading of the implants, which is crucial for new bone formation, cannot be compared. However, the current work provides the first insights into in vivo new bone formation for the described gentamicin coating layers. In combi- nation with the previously published in vivo antibacterial effect [5] of these coatings, this and its sister article deliver a complete picture of this coating technology and target antibacterial effec- tiveness and bone ingrowth, which are the two crucial parameters for preclinical testing of antimicrobial coatings before their intro- duction into clinical practice.
In combination with the previously published in vivo antibacte- rial effect [5] of those coatings, these two articles deliver a com- plete picture of this coating technology and target antibacterial effectiveness and bone ingrowth, which are the two crucial param- eters for preclinical testing of antimicrobial coatings before their introduction into clinical practice. The results should be seen in the light of the well-known weaknesses of animal trials due to a limited number of animals combined with the high standard devi- ation of the results.
In vitro elution kinetics assessment showed that 98% of the gen- tamicin was released after 24 h. Although no in vivo measurements in the animals were made, a similar release can be assumed and re- lease over weeks or months with ‘‘chronic’’ exposure of osteoblasts and osteoblast precursor cells to gentamicin can be excluded. Therefore, this relatively short elution time of gentamicin of only 24 h is most likely related to the differences in bone formation after 12 weeks. There is only limited data on the effects of locally released antibiotics on new bone formation in the literature, par- ticularly in the context of antibiotic-coated implants. Only Moojen et al. [15] have reported a study of new bone formation and im- plant integration for tobramycin–HA implants in a rabbit model compared with tobramycin-free HA implants. They did not find any statistical significance between the tobramycin and the control groups in an infection-free environment. There were even better results for osseointegration and bone–implant contact of tobramy- cin-coated implants compared with the control group after bacte- rial contamination of the intramedullary canal. However, this effect was linked by the authors to the reduced infection rates in the tobramycin–HA group compared with the HA group as infec- tion-free bone most likely enables better bone formation than in- fected bone.
As there were no significant differences in the current study be- tween the formation and appearance of multinucleated giant cells for gentamicin-coated implants compared with the control group, the influence of the gentamicin coating does not seem to be a gen- eral foreign body reaction, as HA itself is known to cause the for- mation of multinucleated giant cells [36]. Therefore, other mechanisms are most likely involved in the effects of gentamicin on bone formation.
Some in vitro studies have reported on the negative effects of different antibiotics on the functions of osteoblasts and osteo- clasts [30–32]. Investigations on cephazolin and vancomycin showed that vancomycin at a concentration of <1000 lg mlti1 had no or only marginal effects on the replication of osteoblasts, however, very high concentrations >10,000 lg mlti1 could lead to cell death. Cephazolin at a concentration of <100 lg mlti1 had no consequences for osteoblasts, however, concentrations of >200 and 10,000 lg mlti1 were related to impaired replication and cell death, respectively [30]. A similar study on tobramycin reported intact replication of osteoblasts at a concentration of
<200 lg mlti 1 [32]. Impaired replication and cell death were found at concentrations of 400 and 10,000 lg mlti1, respectively. Isefuku et al. [31] reported a remarkable decrease in the activity of alka- line phosphatase in and 3H-thymidine uptake by human osteo-

blasts after 96 h incubation with gentamicin concentrations
>100 lg mlti1. On the other hand, Düwelhenke et al. [29] found that gentamicin, streptomycin and tobramycin at concentrations up to 400 lg mlti1 were free of negative effects on the metabolic activity on human osteoblasts.
Only cells of the proximal tubuli of the kidneys and resorptive endothelia of the ear [37] express the so-called megalin receptor, which is responsible for receptor-regulated intracellular uptake and related nephro- and ototoxicity [38,39], therefore the potential effects of gentamicin on bone cannot be explained by this receptor pathway. Both lysosomal accumulation [40,41] with subsequent overloading and apoptosis via extracellular calcium receptor-in- duced apoptosis [42,43] have also been discussed in the literature as potential mechanisms for the cell toxicity of aminoglycosides. The data of the present study cannot clarify the molecular basis of the influence of gentamicin on new bone formation and further research is needed in that field.
The relatively short release time of the two gentamicin coatings in the present work, only 24 h, must also be seen in the context of their antimicrobial effectiveness, with a statistically significant reduction in infection rates compared with gentamicin-free control implants [5]. This relatively short release indicates that the first 24 h post operation seem to be extremely important for infection prophylaxis in the context of locally released antibiotics from anti- microbial coatings. This observation is in line with clinical results for systemic infection prophylaxis from the Norwegian Arthro- plasty Register in which four administrations of intravenous (i.v.) antibiotics on the day of surgery was shown to be most effective in periprosthetic joint infections [2]. If i.v. administration was pro- longed over two or more days no further improvement in infection prophylaxis could be achieved.
The control group of the current work was also used in a similar study to assess the effects of an additional RGD coating on HA im- plants [28]. All animals in the present and above mentioned stud- ies were operated on at the same time, reducing any systematic differences between the two studies.
An RGD coating did not enhance osseointegration in combina- tion with gentamicin–HA or in combination with HA implants compared with RGD-free implants. It seems that the additional RGD layer was unable to improve the attraction of osteoblast pre- cursor cells to the implant surface as, on the one hand, the effect of HA itself seems to be considerably stronger than that of RGD and, on the other, the potential effects of gentamicin could not be bal- anced by RGD.

5.Conclusion

Overall, antimicrobial coating of implants is an important tool in improving of infection prophylaxis in total joint arthroplasty. Besides the antimicrobial effectiveness of the coating, unimpaired osseous integration of the implant is essential. The present study analysed new bone formation, bony integration and formation of multinucleated giant cells on gentamicin–HA-, gentamicin–RGD– HA- and HA-coated implants without any statistically significant differences between the three groups. However, there was a gen- eral trend towards impairment of new bone formation in both groups with gentamicin-coated implants compared with the HA control group. This study underscores the importance of analyses of bone formation and implant integration processes in the context of antimicrobial coatings of implants. Further research will be re- quired in both infection prophylaxis effectiveness and bony inte- gration of the implant before the first antimicrobial coating will be clinically available in cementless joint arthroplasty. This study is limited by the high standard deviation of the results and the lim- ited number of animals.

Acknowledgements

The authors like to thank Prof. H. Kessler, Technische Universi- tät Munich, for RGD peptide development and the team of the Lab- oratory of Experimental Trauma, University Giessen, Germany, with special mention of Anne Hild and Ursula Sommer. The study was supported by Biomet Deutschland, Berlin, Germany.

Appendix A. Figures with essential colour discrimination Certain figures in this article, particularly Figures 1–4, are diffi-
cult to interpret in black and white. The full colour images can be found in the on-line version, at doi:10.1016/j.actbio.2010.11.012.

References

[1]Engesaeter LB, Espehaug B, Lie SA, et al. Does cement increase the risk of infection in primary total hip arthroplasty? Revision rates in 56,275 cemented and uncemented primary THAs followed for 0–16 years in the Norwegian Arthroplasty Register. Acta Orthop Scand 2006;77:351–8.
[2]Engesaeter LB, Lie SA, Espehaug B, et al. Antibiotic prophylaxis in total hip arthroplasty: effects of antibiotic prophylaxis systemically and in bone cement on the revision rate of 22,170 primary hip replacements followed 0–14 years in the Norwegian Arthroplasty Register. Acta Orthop Scand 2003;74:644–51.
[3]Buchholz HW, Engelbrecht H. Über die depotwirkung einiger antibiotika bei vermischung mit dem kunstharz palacos. Der Chirurg 1970;40:511–5.
[4]Klemm K. Gentamicin–PMMA-beads in treating bone and soft tissue infections. Zentralbl Chir 1979;104(14):934–42 [in German].
[5]Alt V, Bitschnau A, Osterling J, et al. The effects of combined gentamicin– hydroxyapatite coating for cementless joint prostheses on the reduction of infection rates in a rabbit infection prophylaxis model. Biomaterials 2006;26:4627–34.
[6]An YH, Bradley J, Powers DL, et al. The prevention of prosthetic infection using a cross-linked albumin coating in a rabbit model. J Bone Joint Surg Br 1997;79:816–9.
[7]Antoci Jr V, King SB, Jose B, et al. Vancomycin covalently bonded to titanium alloy prevents bacterial colonization. J Orthop Res 2007;25:858–66.
[8]Campbell AA, Song L, Li XS, et al. Development, characterization, and anti- microbial efficacy of hydroxyapatite chlorhexidine coatings produced by surface-induced mineralization. J Biomed Mater Res 2000;53:400–7.
[9]Darouiche RO, Mansouri MD, Zakarevicz D, et al. In vivo efficacy of antimicrobial-coated devices. J Bone Joint Surg Am 2007;89:792–7.
[10]De Jong ES, De Berardino TM, Brooks DE, et al. Antimicrobial efficacy of external fixator pins coated with a lipid stabilized hydroxyapatite/
chlorhexidine complex to prevent pin tract infection in a goat model. J Trauma 2001;50:1008–14.
[11]Dunn DS, Raghavan S, Volz RG. Gentamicin sulphate attachment and release from anodized Ti–6A1–4V orthopedic materials. J Biomed Mater Res 1993;27:895–900.
[12]Gollwitzer H, Ibrahim K, Meyer H, et al. Antibacterial poly(D,L-lactic acid) coating of medical implants using a biodegradable drug delivery technology. J Antimicrob Chemother 2003;51:585–91.
[13]Kälicke T, Schierholz J, Schlegel U, et al. Effect on infection resistance of a local antiseptic and antibiotic coating on osteosynthesis implants: an in vitro and in vivo study. J Orthop Res 2006;24:1622–40.
[14]Lucke M, Schmidmaier G, Sadoni S, et al. Gentamicin coating of metallic implants reduces implant-related osteomyelitis in rats. Bone 2003;32:521–31.
[15]Moojen DJ, Vogely HC, Fleer A, et al. Prophylaxis of infection and effects on osseointegration using a tobramycin–periapatite coating on titanium implants – an experimental study in the rabbit. J Orthop Res 2009;27:1002–7.
[16]Price JS, Tencer AF, Arm DM, et al. Controlled release of antibiotics from coated orthopedic implants. J Biomed Mater Res 1996;30:281–6.
[17]Stigter M, de Groot K, Layrolle P. Incorporation of tobramycin into biomimetic hydroxyapatite coating on titanium. Biomaterials 2002;23:4143–53.

[18]Stigter M, Bezemer J, de Groot K, et al. Incorporation of different antibiotics into carbonated hydroxyapatite coatings on titanium implants, release and antibiotic efficacy. J Control Release 2004;99:127–37.
[19]Havelin LI, Engesaeter LB, Espehaug B, et al. The Norwegian Arthroplasty Register: 11 years and 73,000 arthroplasties. Acta Orthop Scand 2000;71:337–53.
[20]Oosterbos CJ, Rahmy AI, Tonino AJ, et al. High survival rate of hydroxyapatite- coated hip prostheses: 100 consecutive hips followed for 10 years. Acta Orthop Scand 2004;75:127–33.
[21]Bernhardt R, van den Dolder J, Bierbaum S, et al. Osteoconductive modifications of Ti-implants in a goat defect model: characterization of bone growth with SR muCT and histology. Biomaterials 2005;26:3009–19.
[22]Ferris DM, Moodie GD, Dimond PM, et al. RGD-coated titanium implants stimulate increased bone formation in vivo. Biomaterials 1999;20:2323–31.
[23]Kantlehner M, Schaffner P, Finsinger D, et al. Surface coating with cyclic RGD peptides stimulates osteoblast adhesion and proliferation as well as bone formation. ChemBioChem 2000;18:107–14.
[24]Kantlehner M, Finsinger D, Meyer J, et al. Selective RGD-mediated adhesion of osteoblasts at surfaces of implants. Angew Chem Int Ed 1999;38:560–2.
[25]Elmengaard B, Bechtold JE, Soballe K. In vivo effects of RGD-coated titanium implants inserted in two bone-gap models. J Biomed Mater Res A 2005;75:249–55.
[26]Elmengaard B, Bechtold JE, Soballe K. In vivo study of the effect of RGD treatment on bone ongrowth on press-fit titanium alloy implants. Biomaterials 2005;26:3521–6.
[27]Rammelt S, Illert T, Bierbaum S, et al. Coating of titanium implants with collagen, RGD peptide and chondroitin sulfate. Biomaterials 2006;27:5561–71.
[28]Bitschnau A, Alt V, Böhner F, et al. Comparison of new bone formation, implant integration, and biocompatibility between RGD–hydroxyapatite and pure hydroxyapatite coating for cementless joint prostheses – an experimental study in rabbits. J Biomed Mater Res B Appl Biomater 2009;88:66–74.
[29]Düwelhenke N, Krut O, Eysel P. Influence on mitochondria and cytotoxicity of different antibiotics administered in high concentrations on primary human osteoblasts and cell lines. Antimicrob Agents Chemother 2007;51:54–63.
[30]Edin ML, Miclau T, Lester GE, et al. Effect of cefazolin and vancomycin on osteoblasts in vitro. Clin Orthop Relat Res 1996;333:245–51.
[31]Isefuku S, Joyner CJ, Simpson AH. Gentamicin may have an adverse effect on osteogenesis. J Orthop Trauma 2003;17:212–6.
[32]Miclau T, Edin ML, Lester GE, et al. Bone toxicity of locally applied aminoglycosides. J Orthop Trauma 1995;9:401–6.
[33]Rössler S, Sewing A, Stolzel M, et al. Electrochemically assisted deposition of thin calcium phosphate coatings at near-physiological pH and temperature. J Biomed Mater Res A 2003;64:655–63.
[34]Auernheimer J, Zukowski D, Dahmen C, et al. Titanium implant materials with improved biocompatibility through coating with phosphonate-anchored cyclic RGD peptides. ChemBioChem 2005;6:2034–40.
[35]Donath K, Breuner G. A method for the study of undecalcified bones and teeth with attached soft tissues. The Sage–Schliff (sawing and grinding) technique. J Oral Pathol 1982;11:318–26.
[36]Gottlander M, Johansson CB, Albrektsson T. Short- and long-term animal studies with a plasma-sprayed calcium phosphate-coated implant. Clin Oral Implants Res 1997;8:345–51.
[37]Lundgren S, Carling T, Hjalm G, et al. Tissue distribution of human gp330/
megalin, a putative Ca(2+)-sensing protein. J Histochem Cytochem 1997;45:383–92.
[38]Moestrup SK, Cui S, Vorum H, et al. Evidence that epithelial glycoprotein 330/
megalin mediates uptake of polybasic drugs. J Clin Invest 1995;96:1404–13.
[39]Schmitz C, Hilpert J, Jacobsen C, et al. Megalin deficiency offers protection from renal aminoglycoside accumulation. J Biol Chem 2002;277:618–22.
[40]Silverblatt FJ, Kuehn C. Autoradiography of gentamicin uptake by the rat proximal tubule cell. Kidney Int 1979;15:335–45.
[41]Tulkens P, van Hoof F. Comparative toxicity of aminoglycoside antibiotics towards the lysosomes in a cell culture model. Toxicology 1980;17:195–9.
[42]Davey PJ, Haslam JM, Linnane AW. Biogenesis of mitochondria 12. The effects of aminoglycoside antibiotics on the mitochondrial and cytoplasmic protein- synthesizing systems of Saccharomyces cerevisiae. Arch Biochem Biophys 1970;136:54–64.
[43]Ward DT, Maldonado-Perez D, Hollins L, et al. Aminoglycosides induce acute cell signaling and chronic cell death in renal cells that express the calcium- sensing receptor. J Am Soc Nephrol 2005;16:1236–44.