Restoring the osteogenic activity of bacterial debris contaminated titanium by doping with magnesium

Abstract

Immobilized silver nanoparticles (Ag NPs) can give titanium contact-killing action against the colonization of bacteria. However, this antimicrobial process may produce bacterial debris, which is likely to compromise the osteogenic property of the material. Magnesium, as an effective stimulus for boosting the osteogenic performance of bone cells, may neutralize this adverse impact. Accordingly, the present study illuminates the effect of bacterial debris (BD) and lipopolysaccharide (LPS, a kind of endotoxin in microbial debris) on the osteogenesis of titanium, which was surface-modified by magnesium (Mg) and/or silver (Ag) plasmas. It was found that the unfavorable effects of LPS and BD on osteoblastic differentiation of human bone marrow mesenchymal stem cells (hBMSCs) can be overcome by doping with magnesium on the titanium surface. And titanium co-doped with magnesium and silver, even under the challenge of bacteria, gives the best osseointegration in vivo. This study highlights that bacterial debris can impair the osseointegration of implants and co-doping by magnesium and silver is a promising method to solve this issue.

---

1. Introduction

The skyrocketing increase in the number of joint replacement surgeries and their associated failures has raised serious concerns in the field of biomedical engineering. Among various causations, periprosthetic joint infections (PJI) are a devastating complication after arthroplasty and are associated with high morbidity and substantial cost.^1,2 PJI is estimated at 1% for hip arthroplasties and ranges between 1% and 2% after knee arthroplasties every year.^3,4 Infections were the leading cause of revision (25.2%) after knee arthroplasty and accounted for 14.8% of revisions after hip arthroplasty.^5,6 As the number of joint replacement surgeries increases, failure of implants due to infection has caused an increasing burden to health-care systems. Infections can occur through direct seeding from external contaminants or contiguous spread, haematogenous spread from other body sites and recurrent infection. Infection susceptibility is increased in settings of foreign implants,^7,8 and might result in the formation of biofilm, which can protect a microorganism from antibiotic treatment or host defenses.⁹

Many antibacterial coatings have been developed to prevent the formation of biofilms on biomaterials.^10,11 Inorganic antimicrobial agents with good stability were extensively studied for such applications.¹² Among these biocides, silver nanoparticles (Ag NPs) are the most famous materials.¹³ And the material was adopted for a releasing defense of various biomaterials.¹⁴ However, these releasing-based methods have aroused serious concerns about their size and mobility-related toxicity to mammalian cells.^15,16 It was found that immobilized Ag NPs on titanium possess a strong contact-killing action on bacterial cells in vitro and in vivo.^17,18 Even so, the validity of this strategy still needs further evaluation because bacterial debris is likely to be produced due to the contact-killing manner of these immobilized Ag NPs. And the remaining debris on titanium may cause a deterioration in the osteogenesis property of the material. It was reported that the endotoxins in microbial debris will induce bone resorption.^19–21 And many studies confirmed that bacterial endotoxins played a significant role in the failure of implants.^22–24 Lipopolysaccharide (LPS) is a classical endotoxin in Gram-negative bacteria.²⁵ LPS has a high affinity to biomaterials,²⁶ and can inhibit osseointegration and induce bone absortion.^27–29

Previous studies have shown that the osteogenic property of Ag NPs modified titanium can be improved by co-doping with magnesium which is an effective stimulus for boosting osteogenic performances in bone cells.^30–32 Accordingly, we hypothesized that bacterial debris inhibits the osseointegration of titanium fabricated with Ag NPs and co-doping with Mg can offset this side-effect. To test the hypothesis, the present study illuminates the potential effect of bacterial debris on the osteogenesis of titanium, which was treated by magnesium (Mg) and/or silver (Ag) plasmas. It was found that bacterial debris did compromise the osteogenic property of Ag NPs armed titanium and the osteogenesis of the material can be restored by doping with magnesium.

2. Materials and methods

2.1. Preparation of materials

Commercial pure titanium (Cp Ti, Grade 2) was cut into square plates (10 × 10 × 1 mm or 20 × 20 × 1 mm in size) or rods (2 mm in diameter, 7 mm in length), acid-etched and ultrasonically cleaned in distilled water. Three groups of doped substrates were prepared by carrying out plasma immersion ion implantation (PIII) at 30 kV for 1.5 h with different plasmas, i.e. titanium treated with magnesium plasma (designated Mg-PIII), titanium treated with silver plasma (designated Ag-PIII), and titanium in situ treated with magnesium and silver plasma (designated Mg/Ag-PIII). The PIII parameters were detailed in previous studies.¹⁷ The surface morphology was examined by field-emission scanning electron microscopy (FE-SEM, JEOL JSM-6700F, Tokyo, Japan), and the chemical state of the doped components was determined by X-ray photoelectron spectroscopy (XPS, PHI 5802, Physical Electronics Inc., Eden Prairie, MN).

In order to study the effect of bacterial debris on mammalian cells, microbial cells were pre-cultured on both the bare titanium and on PIII-treated titanium, and killed by illumination with ultraviolet (UV) light. Briefly, Escherichia coli cells (E. coli, ATCC 25922, in a concentration of 1 × 10⁴ cfu ml⁻¹) were cultured on the fabricated substrates at 37 °C for 24 h. After that, the samples were illuminated with UV for 4 hours in a biosafety cabinet, and then the samples were retrieved and kept in the dark overnight for further evaluation.

In order to study the effect of lipopolysaccharide (LPS) on mammalian cells, samples contaminated with LPS (Escherichia coli serotype O26:B6; Sigma-Aldrich, St. Louis, MO, USA) were prepared. For screening, human bone marrow mesenchymal stem cells (hBMSCs, Stem Cell Bank, Chinese Academy of Science, Shanghai, China) were seeded on acid-cleaned titanium, and different concentrations of LPS (10, 1 and 0.01 μg ml⁻¹) were added. At various time points (days 1, 4, 7), the proliferation and extracellular matrix mineralization activity of the cells at day 14 were assessed. According to the results (Fig. S1 and S2†), LPS in a concentration of 1 μg ml⁻¹ can notably inhibit osteogenic differentiation of hBMSCs, but does no apparent harm to their proliferation. Therefore, all the subsequent experiments concerning LPS were done at this concentration.

2.2. Cell culture

The human bone marrow mesenchymal stem cells (hBMSCs, Stem Cell Bank, Chinese Academy of Science, Shanghai, China) were maintained in a culture medium consisting of Dulbecco's Modified Eagle Medium (DMEM; Gibco, USA) supplemented with 10% heat-inactivated fetal bovine serum (FBS; Hyclone, USA) and 1% antibiotic/antimycotic solution (antibiotic–antimycotic, Hyclone, USA). Cells at passages 3 to 5 were used for downstream experiments. Cells (10 000 cells per well) were seeded onto each plate (10 × 10 × 1 mm) in a 24-well plate to evaluate the proliferation and spreading of hBMSCs. In the extracellular matrix mineralization assay and alkaline phosphatase (ALP) quantification, cells (5000 cells per well) were incubated onto different plates (10 × 10 × 1 mm) for 14 days. And for real-time reverse-transcriptase polymerase chain reaction (real-time RT-PCR) measurements, cells (10 000 cells per well) were co-cultured on different plates (20 × 20 × 1 mm) for 14 days.

2.3. Cell proliferation

In order to assess cell proliferation, the cells were cultured on samples for 1, 4 and 7 days and were assessed using the Cell Counting Kit-8 assay (CCK-8, Beyotime). At prescribed time points, the culture medium was removed and fresh medium and CCK-8 solution in a ratio of 10 : 1 was added for an incubation period of 4 h at 37 °C. The absorbance was measured with an enzyme-labeling instrument (BIO-TEK, ELX 800) at a wavelength of 450 nm.

2.4. Cell spreading

To assess early cell spreading, samples were rinsed with phosphate buffer saline (PBS) three times and the cells were fixed in 4% paraformaldehyde for 10 min at room temperature after 24 hours of co-culture. Cells were then permeated by Triton X-100 (0.1% v/v) in PBS for 5 min. The permeabilized cells were incubated with rhodamine phalloidin (Sigma, St. Louis, MO) and DAPI (Sigma, St. Louis, MO) at 37 °C in darkness for 1 h and 5 min, respectively. Cells were viewed using a fluorescence microscope (Olympus, Japan).

2.5. Extracellular matrix mineralization assay

On day 14, the hBMSCs on the samples were fixed with 70% ethanol for 1 hour, and incubated with 40 mM of Alizarin Red S for 10 min. Unbound stain was removed by washing three times with phosphate-buffered saline (PBS), and the optical images were acquired. For the quantitative evaluation, 10% cetylpyridinium chloride in 10 mM of sodium phosphate (pH 7.0) was used to dissolve the mineralized nodules, and optical density (OD) values were measured at 600 nm.

2.6. Alkaline phosphatase activity

Alkaline phosphatase (ALP) staining was performed on day 14. The culture medium was removed, and the cells were washed with PBS and fixed by citrate-buffered acetone. A working solution was prepared by fast blue RR salt (Sigma-Aldrich) and naphthol AS-MX phosphate (Sigma-Aldrich). After gently rinsing with PBS, the cells were incubated with the working solution for 30 minutes, and then stained with Mayer's hematoxylin reagent (Sigma-Aldrich) for 10 minutes. The stained cells were observed under a fluorescence microscope.

For the quantitative assay, the total cellular protein on the samples was lysed by an alkaline lysis buffer. The ALP activity in the lysate was determined by the measurement of p-nitrophenyl phosphate (pNPP) using a commercial assay kit (Beyotime Institute of Biotechnology, Shanghai, China) according to the manufacturer's protocol. Briefly, the cell lysate of each sample was incubated with a solution containing 2.5 mM of pNPP at 37 °C for 30 min. After the reaction was terminated by the addition of 0.2 M NaOH, the OD value was read at a wavelength of 405 nm in a spectrophotometer. The concentration was calculated by projecting the optical densities on the standard curve. The cellular ALP activities were normalized against the total protein concentration measured by the BCA protein assay.

2.7. Quantitative real-time polymerase chain reaction (RT-PCR)

Total RNA was extracted using TRIzol reagent (Invitrogen) on day 14 for RT-PCR detection. Three hundred nanograms of total RNA were used for the synthesis of complementary DNA using the PrimeScript RT reagent Kit (Takara) following the manufacturer's instructions. RT-qPCR primers (Table 1) were designed based on cDNA sequences from the NCBI Sequence database. SYBR Premix Ex Taq II (Takara) was used for detection, and the target mRNA expressions were assayed on Bio-Rad C1000. The mean cycle threshold (Ct) value of each target gene was normalized against the Ct value of the housekeeping gene GAPDH to gain the relative expression.

Table 1 Primers for RT-PCR

Gene	Prime sequence (F, forward; R, reverse; 5′-3′)	Product size (bp)
Runx-2	F: GAGATCATCGCCGACCAC	135
	R: TACCTCTCCGAGGGCTACC
ALP	F: CCGTGGCAACTCTATCTTTGG	79
	R: GCCATACAGGATGGCAGTGA
OCN	F: AGCAAAGGTGCAGCCTTTGT	63
	R: GCGCCTGGGTCTCTTCACT
Col-1	F: AGGGCCAAGACGAAGACATC	138
	R: AGATCACGTCATCGCACAACA
GAPDH	F: ATGGGGAAGGTGAAGGTCG	119
	R: TAAAAGCAGCCCTGGTGACC

---

2.8. Implant-related tibia osteomyelitis model in rabbits

All the procedures regarding animal maintenance and experiments are in strict accordance with the guidelines of the Institutional Animal Care and Use Committee (IACUC) of Shanghai JiaoTong University affiliated Shanghai sixth people's hospital, the National Institutes of Health Guide for Care and Use of Laboratory Animals (GB14925-2010) and the Regulations for the Administration of Affairs Concerning Experimental Animals (China, 2014). The IACUC has approved this study. Twenty male New Zealand white rabbits aged about 8 months, weighing 2800 ± 300 g, were chosen for the in vivo experiments. The rabbits were anaesthetized by intravenous injection with 3% pentobarbital sodium solution (0.9 ml per 1000 g). Then the operation site (left tibia) was shaved, sterilized, and the body was covered with sterile sheets. The tibia was exposed by the medial approach. A hole between the tibial plateau and tibial tuberosity was drilled by a Kirschner wire (1.5 mm in diameter) to access the medullary cavity. The medullary cavity was further reamed to the distal part of the tibia by a Kirschner wire 2.0 mm in diameter. 500 μl of bacterial suspension (E. coli) with a concentration of 1 × 10⁴ cfu ml⁻¹ was transferred into the medullary cavity with a microsyringe. After bacterial inoculation, four parallel holes at an interval of 1 cm were drilled by a Kirschner wire (2.0 mm in diameter). Subsequently, the sterile ion-implanted or titanium rods were implanted into the prepared holes, and then the incision was sutured carefully. Following surgery, the rabbits were housed in separate cages and allowed to eat and drink ad libitum. The rabbits were euthanized 6 weeks post-surgery.

2.9. Radiographic evaluation

Radiographic images were obtained at 0, 2, 4, 6 weeks after surgery. The radiographic performances around the implants were assessed in a blind manner. The key points according to the literature, including periosteal reaction, osteolysis and general impression were considered. In order to specify the changes in the bone around the implants, the tibia harvested at 6 weeks were scanned using a micro-CT device (Skyscan 1172, Bruker micro CT, German). The 2D planes were reconstructed using the NRecon program. The bone mineral density (BMD), bone volume fraction (bone volume/total volume, BV/TV) and bone-implant contact rate (bone-implant contact, BIC) were measured by the CTAn program.

2.10. Histological analysis

The harvested tibia with implants were fixed and decalcified. After the implants were pushed out, the tibia were cut into four parts based on the position of the implants. The specimens were dehydrated in gradient ethanol solution and embedded in paraffin. The embedded tissue was sectioned to a thickness of 6 mm and stained by Masson's trichrome staining for optical microscopy.

2.11. Statistical analysis

The data were presented as means ± standard deviation. Statistical analysis was performed with SPSS 17.0 (SPSS Inc., Chicago, USA). One-way or two-way ANOVA followed by Bonferroni post-hoc tests were applied to determine the level of significance in the differences between groups, which were based on the normality distribution (Kolmogorov-Smirnov test with Lilliefors' correction) and equal variance testing (Levene median test). P values less than 0.05 were considered to be significant.

3. Results

3.1. Surface characterization

By carrying out plasma immersion ion implantation (PIII) with different plasma sources, four groups of substrates were prepared, i.e. the titanium control without PIII treatment (designated Ti), titanium treated with magnesium plasma (designated Mg-PIII), titanium treated with silver plasma (designated Ag-PIII), and titanium in situ treated with magnesium and silver plasma (designated Mg/Ag-PIII). The surface microstructures of the samples were observed by SEM. The surfaces of the pure titanium control (Fig. 1A) and Mg-PIII treated titanium (Fig. 1C) were smooth at the nano scale, while nano particles (∼5 nm in diameter) can be identified on Ag-PIII (Fig. 1B) and Mg/Ag-PIII (Fig. 1D) groups. The samples were further examined by X-ray photoelectron spectroscopy (XPS). As evidenced by the full spectra (Fig. 2A–D), magnesium (Mg) and/or silver (Ag) were successfully introduced onto titanium surfaces, and the nano particles on the Ag-PIII and Mg/Ag-PIII groups are likely to be silver. According to our previous study, the highest Mg concentration on the Mg-PIII surface was about 6.0 atomic percent (atom%), while the peak Mg content at the surface on Mg/Ag-PIII was about 5.5 atom%. On the other hand, the silver concentration on the surface of Mg/Ag-PIII was 0.5 atom%, which was higher than that of Ag-PIII (0.3 atom%).³⁰

Fig. 1 Field-emission scanning electron microscopy (FE-SEM) views of the samples: (A) Ti, (B) Ag-PIII, (C) Mg-PIII, (D) Mg/Ag-PIII.

Fig. 2 The full XPS spectra of the samples: (A) Ti, (B) Ag-PIII, (C) Mg-PIII, (D) Mg/Ag-PIII; the Ag nanoparticles are marked by red arrows; the triangle and red star represent silver (Ag) and magnesium (Mg), respectively.

3.2. In vitro effects of bacterial debris and LPS on hBMSCs

3.2.1. Cell proliferation. The proliferation of human mesenchymal stem cells (hBMSCs) on bacterial debris contaminated groups was evaluated by CCK-8 assay. As the results in Fig. 3A show, impacted by bacterial debris, the OD₄₅₀ (optical density of light 450 nm) values of the Mg-PIII and Mg/Ag-PIII groups are significantly higher than those of the Ti and Ag-PIII groups at all the time points (days 1, 4, and 7), demonstrating that cell proliferation on the latter groups is inferior to that on the former groups. The screening results (Fig. S1 and S2†) demonstrated that lipopolysaccharide (LPS) in a concentration of 1 μg ml⁻¹ does no apparent harm to the proliferation of human mesenchymal stem cells (hBMSCs) but can notably depress the osteoblastic differentiation of the cells. Thus, all the subsequent experiments concerning LPS were done at this concentration. According to the CCK-8 assay (Fig. 3B), with the interference of LPS, the OD₄₅₀ values of the Ti and Ag-PIII groups are significantly lower than those of the Mg-PIII and Mg/Ag-PIII groups at all the time points (days 1, 4, and 7), indicating that cell proliferation on the latter groups is more favorable than that on the former groups.

Fig. 3 Cell proliferation of the human mesenchymal stem cells (hBMSCs) seeded on various samples with the presence of bacterial debris (BD) (A) and LPS (B). Ti-BD⁻ and Ti-LPS⁻ mean pure titanium without BD and LPS contamination. *P < 0.05, **P < 0.01.

3.2.2. Cell spreading. The hBMSCs cultured on the samples were stained by rhodamine phalloidin and DAPI to observe the cell spreading state at day 1. As the results in Fig. 4 show, the cells on substrates without any contamination of bacterial debris or LPS (BD⁻ or LPS⁻ in Fig. 4) exhibited a fully spreading state with abundant filopodia expressed, whereas the cells on the samples treated by bacterial debris or LPS (BD⁺ or LPS⁺ in Fig. 4) show fuzzy shapes with less filopodia. Besides, with the interference of BD, the number of polygonal and constrictive cells was greater on the Ti (99% ± 1%) and Ag-PIII groups (95% ± 3%) than on the Mg-PIII (35% ± 5%) and Mg/Ag-PIII groups (23% ± 3%). With the interference of LPS, the same trend was found. The numbers of polygonal and constrictive cells on the Ti, Ag-PIII, Mg-PIII and Mg/Ag-PIII were 90% ± 5%, 80% ± 3%, 25% ± 3% and 15 ± 2%. These results are evidence that the harmful effect of bacterial debris and LPS on cell adhesion and spreading can be overcome by doping of magnesium on titanium (the Mg-PIII and Mg/Ag-PIII groups).

Fig. 4 Cellular spreading shape, with or without the presence of bacterial debris (BD) or LPS, was detected by cytoskeleton stained with rhodamine phalloidin (red) and nuclei stained with DAPI (blue) at day 1. The size bar represents 50 μm.

3.2.3. Extracellular matrix mineralization and ALP activity. The effect of bacterial debris and LPS on the expression of extracellular matrix mineralization (ECM) and alkaline phosphatase (ALP) in hBMSCs was assessed by culturing the cells on the concerned material groups for 14 days. The results demonstrated that the extracellular matrix mineralization process on all the groups was apparently inhibited by LPS or bacterial debris (Fig. 5A). However, this effect on Mg-PIII and Mg/Ag-PII is slighter than that on the Ti and Ag-PIII groups (Fig. 5B). Compared with pure titanium, in the presence of LPS, ECM on Ti and Ag-PIII was reduced to 30 ± 5% and 35 ± 6%. However, ECM on Mg-PIII and Mg/Ag-PIII recovered to 76 ± 8% and 86 ± 4%. In the presence of bacterial debris, ECM on Ti, Ag-PIII, Mg-PIII and Mg/Ag-PIII were 10 ± 4%, 20 ± 4%, 45 ± 4% and 57 ± 4% respectively, when ECM on pure titanium was standardized to 100%. A similar trend was detected in ALP. As the results in Fig. 6 show, the expression of ALP on all the groups was apparently down-regulated by LPS or bacterial debris (Fig. 6A), and the effect on Ti and Ag-PIII groups is stronger than that on Mg-PIII and Mg/Ag-PIII groups (Fig. 6B). When ALP activity on pure titanium was standardized to 100%, in the presence of LPS, ALP activity on Ti, Ag-PIII, Mg-PIII and Mg/Ag-PIII was reduced to 16 ± 3%, 25 ± 2%, 62 ± 4% and 73 ± 3%. And in the presence of bacterial debris, ALP activity on Ti, Ag-PIII, Mg-PIII and Mg/Ag-PIII was inhibited to 10 ± 4%, 16 ± 4%, 22 ± 3% and 35 ± 4%, respectively. These results demonstrated that bacterial debris and LPS did inhibit the osteogenic differentiation of hBMSCs, but doping by magnesium can significantly neutralize this adverse effect. It should be noted that the inhibitory effects of bacterial debris on ECM and ALP were stronger than that of LPS (Fig. 5B and 6B), indicating that LPS is not the only factor answering for the responses of the cells.

Fig. 5 Extracellular matrix mineralization after 14 days of culture was determined by alizarin red staining: (A) the stained area on various surfaces with or without the presence of bacterial debris (BD) or LPS (bar = 200 μm), and (B) the relative fold changes of ECM on the samples treated with LPS or bacterial debris (BD), the BD⁻LPS⁻ group had been standardized as 1; *P < 0.05, **P < 0.01.

Fig. 6 (A) Alkaline phosphatase (ALP) staining of hBMSCs on various surfaces after 14 days of culture with or without the presence of bacterial debris (BD) or LPS (bar = 100 μm); (B) the relative fold changes of ALP on the samples treated with LPS or bacterial debris, the BD⁻LPS⁻ group had been standardized as 1; *P < 0.05, **P < 0.01.

3.2.4. Expression of osteogenic gene in hBMSCs. The effect of bacterial debris and LPS on expression of osteogenesis-related genes in hBMSCs was evaluated by real-time PCR. As the results in Fig. 7 show, in the presence of LPS, the expression of ALP, OCN, Col-I and Runx-2 on Ti and Ag-PIII groups declined to 28–38% of the original state; however, the level recovered to 62–75% on the Mg-PIII and Mg/Ag-PIII groups. Meanwhile, in the presence of bacterial debris (BD), the expression of all the genes recovered from 12% to 22% on titanium and Ag-PIII, and from 40% to 62% on the Mg-PIII and Mg/Ag-PIII groups. Although all the concerned genes were down-regulated by bacterial debris and LPS, the relative folds in expression of ALP, Col-I, Runx-2 and OCN on the Mg-PIII and Mg/Ag-PIII groups were higher than those on the Ti and Ag-PIII groups, indicating that doping by Mg is effective in relieving the side effects of bacterial debris and LPS on hBMSCs.

Fig. 7 The relative fold changes of osteogenic gene expression of hBMSCs after 14 days of culture on various surfaces in the group treated with LPS or bacterial debris (BD), compared with the BD⁻LPS⁻ group (standardized as 1): (A) ALP, (B) Col-I, (C) Runx-2 and (D) OCN; *P < 0.05.

3.3. In vivo effects of bacterial attacks on osseointegration of titanium

3.3.1. Radiographic evaluation. The effect of bacterial attacks on the osseointegration of the samples was examined by using a tibia osteomyelitis model in rabbits. X-ray and micro-CT were used to assess the infection and osseointegration of the sample groups. The results demonstrated that there is no sign of infection around all the groups within 3 days (Fig. 8). After 2 weeks, osteolytic lesions, cortical bone absorption (blue arrows) and periosteal reaction (red arrows), which are in accordance with the signs of bacterial osteomyelitis,^33,34 become evident around Ti but there are no apparent changes around the other groups. After 4 weeks, the signs of osteomyelitis become much more obvious around the Ti, and similar signs are also apparent around the Mg-PIII implant. After 6 weeks, significant bone destruction and severe periosteal reaction lead to obvious bone defects and the deformity of the bone around the Ti and Mg-PIII implants. In contrast, no signs of osteomyelitis can be detected around the Ag-PIII and Mg/Ag-PIII samples within the 6 weeks. The results demonstrate that Ag doped and Mg/Ag co-doped implants exhibit good antibacterial activity in vivo.

Fig. 8 The anteroposterior X-ray films of the tibia with implants at different time points and the control group at 6 week were shown. Cortical bone absorption and periosteal reaction are marked with blue arrows and red arrows.

The typical reconstructed coronal Micro-CT sections of the implants harvested from the group with or without E. coli after 6 weeks, are shown in Fig. 9A. In the groups contaminated by E. coli, the presence of osteolytic lesions (blue arrows), cortical bone absorption (red arrows) around the Ti and Mg-PIII implant is well evidenced. A significant bone defect can be found around the Ti implant, indicating the severest osteomyelitis. In contrast, no apparent sign of infection was observed from Mg/Ag-PIII, but it shows good osseointegration even under the challenge of bacterial cells. In order to assess the osseointegration around the implants, the new bone area of Mg/Ag-PIII and Ag-PIII marked by red rectangles was evaluated by the bone mineral density (BMD), bone volume fraction (bone volume/total volume, BV/TV) and bone-implant contact (BIC), which were determined by the CTAn program. The selected area (red rectangle) is the same as in our previous study.³⁰ And the fold changes of BMD, BV/TV and BIC, compared with the results without stimulation of E. coli (standardized as 1), were calculated and are shown in Fig. 9B–D. The new bone formation around Ti and Mg-PIII was not assessed, because the necrotic bone sequestra and periosteal reaction caused by serious infection may lead to invalid results.

Fig. 9 The typical central coronal sections of the rabbit tibias with implants after 6 weeks in the group in the presence of Escherichia coli (E. coli⁺) and the group without the presence of Escherichia coli (E. coli⁻) were reconstructed by Micro-CT. (A) Formation of osteolysis (blue arrow), resorption of cortical bone (red arrow) can be observed around the Ti and Mg-PIII implants, no signs of infection can be observed around Ag-PIII and Mg/Ag-PIII implants. New bone areas around Ag-PIII and Mg/Ag-PIII implants are marked by red rectangles and analyzed by Micro-CT; (B) the relative changes in bone mineral density (BMD) of the new bone; (C) the relative changes in bone volume fraction (bone volume/total volume, BV/TV) of the new bone; (D) the relative changes in bone-implant contact rate (bone-implant contact, BIC); the E. coli⁻ group had been standardized as 1; (*P < 0.05).

3.3.2. Histological evaluation. The samples were further evaluated by Masson's trichrome staining assay (Fig. 10). Typical signs of osteomyelitis: infiltration of inflammatory cells, cortical bone destruction and severe periosteal reaction were observed around the Ti and Mg-PIII implants.³⁵ The result is consistent with the radiographic and micro-CT images shown in Fig. 8 and 9. The necrotic bone (red arrow) and abscess (black arrow) can be seen in the medullary cavity around Ti, which indicates severer inflammation than with Mg-PIII. Though there is no evidence of infection around Ag-PIII, it is worth noticing that fibrous connective tissue with infiltration of inflammatory cells (green arrow) formed at the interface. The new bone (pink arrow) formation without fibrous connective tissue at the interface indicates the good osseointegration and antibacterial ability of Mg/Ag-PIII. Besides, more new bone formation (pink arrow) can also be found around Mg-PIII than around Ti and Ag-PIII. These results demonstrated that implant-associated infection can be inhibited by Ag-PIII treatment, and co-doping of magnesium (the Mg/Ag-PIII group) can improve osseointegration of the titanium implants.

Fig. 10 The histological sections stained by Masson's trichrome staining. Higher magnifications of the red rectangle area are also displayed; the typical signs of osteomyelitis: infiltration of inflammatory cells (green arrow), cortical bone destruction and severe periosteal reaction can be found around the Ti and Mg-PIII implants. The necrotic bone (red arrow) and abscess (black arrow) can be seen in the medullary cavity around Ti. The new bone (pink arrow) formation can be detected around Mg-PIII and Mg/Ag-PIII.

4. Discussion

Magnesium (Mg) and/or silver (Ag) can be incorporated onto titanium substrates by applying the plasma immersion ion implantation (PIII) technique, and four groups of samples can be fabricated, namely the Ti control, Ag doped titanium (designated as Ag-PIII), Mg doped titanium (designated as Mg-PIII), Mg and Ag co-doped titanium (designated as Mg/Ag-PIII). Previously, it was found that the Ag-PIII group possesses strong contact-killing action to microbia in vitro and in vivo.^17,18 However, bacterial debris are likely to be produced due to the contact-killing manner of the Ag-PIII group, and the remaining debris on titanium may cause a deterioration in the osteogenesis property of the material. In this study, we do find that bacterial debris and lipopolysaccharide (LPS) cause a deterioration in the osteogenic property of the concerned groups (Ti, Ag-PIII, Mg-PIII, and Mg/Ag-PIII), but co-doping with magnesium can apparently overcome the side-effects on Ag-PIII groups.

The effects of bacterial debris or LPS on adhesion, proliferation and osteogenic differentiation of human mesenchymal stem cells (hBMSCs) on the concerned groups were evaluated. The results demonstrated that LPS and bacterial debris can apparently inhibit the adhesion, proliferation and osteogenic differentiation of hBMSCs cultured on the titanium and PIII-treated titanium substrates. And the efficacy of bacterial debris is stronger than that of LPS. This is because the bacterial debris contains multiple immunostimulatory molecules and is therefore more potent than LPS.^36,37 Even though the side-effects of bacterial debris and LPS on osteogenesis of titanium substrates can be overcome by doping with magnesium. Mg was known to enhance the proliferation of the hBMSCs and prompt early cellular spreading.^38–40 This study has further confirmed that magnesium can improve the cellular spreading and proliferation of hBMSCs on the samples in the presence of LPS and bacterial debris. LPS can inhibit osseointegration by activating Toll-like Receptor 4 (TLR4) which results in the production of bone-absorbing cytokines such as tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β) and receptor activator of nuclear factor-κ B ligand (RANKL).^30–32 In contrast, magnesium may decrease TLR-mediated cytokine production by increasing the constitutive IκBα levels and reducing NF-κB activation and nuclear localization, which can explain the bone repair ability of the Mg/Ag-PIII implant.⁴¹

The in vivo effect of bacterial attacks on the osseointegration of the samples was further examined by using an implant-related tibia osteomyelitis model in rabbits. The X-ray, Micro-CT and histological evaluation showed that severe infection can be found around Ti and Mg-PIII, but no evidence of infection was detected around Ag-PIII and Mg/Ag-PIII. Then new bone formation around the Ag-PIII and Mg/Ag-PIII implants, around which the infection had been controlled in vivo, was measured by Micro-CT and assessed by Masson's trichrome staining. The new bone formation around Mg/Ag-PIII is significantly higher than that around Ag-PIII. Meanwhile the histological sections showed that more bone formation was found around Mg-PIII than around Ti. The trend was consistent with the results without stimulation of E. coli.³⁰ This means that Mg could promote osteogenesis no matter whether infection occurred or not. This could be ascribed to the bioactivity of Mg, which was confirmed in the in vitro study. The lower osseointegration of Ag-PIII can also be ascribed to the inflammation induced by the residual bacterial debris. Fibrous connective tissue with infiltration of inflammatory cells at the interface of the Ag-PIII implant can be found, while no evidence of inflammation can be detected on the interface of the Mg/Ag-PIII implant. So magnesium may inhibit the inflammatory reaction caused by the residual bacterial debris in vivo. The mechanism may be associated with the suppression of inflammatory cytokine production through the inhibition of the TLR pathway by magnesium.⁴¹ The present work demonstrates that both Ag and Mg are highly complementary against implant-associated infection.

5. Conclusions

In this study, the effects of bacterial debris or LPS on adhesion, proliferation and osteogenic differentiation of human mesenchymal stem cells (hBMSCs) on titanium, which was treated by magnesium (Mg) and/or silver (Ag) plasmas, were evaluated in vitro, and the effect of bacterial attacks on the osseointegration of the concerned materials was examined in vivo. According to the results, bacterial debris and LPS do have unfavorable effects on adhesion, proliferation and osteogenic differentiation of hBMSCs, and these effects can be overcome by doping with magnesium on the titanium surface. This study highlights that bacterial debris can impair the osseointegration of implants and co-doping by magnesium and silver is a promising method to solve this issue.

Acknowledgements

Financial support from National Basic Research Program of China (973 Program, 2012CB933600), National Natural Science Foundation of China (81271962, 81472109, and 31370962), Shanghai Committee of Science and Technology, China (14XD1403900), Youth Innovation Promotion Association CAS (2015204), Shanghai Rising-Star Program (15QA1404100), the Open Research Fund of State Key Laboratory of Bioelectronics, Southeast University, the Special Research Project of Health Care Industry, Ministry of Public Health, China (201302007) are acknowledged.

Notes and references

1. S. M. Kurtz, K. L. Ong, J. Schmier, F. Mowat, K. Saleh, E. Dybvik, J. Kärrholm, G. Garellick, L. I. Havelin, O. Furnes, H. Malchau and E. Lau, J. Bone Jt. Surg., Am. Vol., 2007, 89, 144–151.
2. B. H. Kapadia, R. A. Berg, J. A. Daley, J. Fritz, A. Bhave and M. A. Mont, Lancet, 2016, 387, 386–394.
3. S. M. Kurtz and O. Kllau, Clin. Orthop. Relat. Res., 2010, 468, 52–56.
4. H. Dale, G. Hallan, B. Espehaug, L. I. Havelin and L. B. Engesaeter, Acta Orthop., 2009, 80, 639–645.
5. K. Bozic, S. M. Kurtz, E. Lau, K. Ong, T. P. Vail, H. E. Rubash and D. J. Berry, Clin. Orthop. Relat. Res., 2010, 468, 45–51.
6. J. Parvizi, I. M. Pawasarat, K. A. Azzam, A. Joshi, E. N. Hansen and K. J. Bozic, Journal of Arthroplasty, 2010, 25, 103–107.
7. A. P. Puhto, T. M. Puhto, T. T. Niinimäki, J. I. Leppilahti and H. P. Syrjälä, Journal of Arthroplasty, 2014, 29, 1101–1104.
8. W. Zimmerli, P. D. Lew and F. A. Waldvogel, J. Clin. Invest., 1984, 73, 1191–1200.
9. J. W. Costerton, P. S. Stewart and E. P. Greenberg, Science, 1999, 284, 1318–1322.
10. M. Cloutier, D. Mantovani and F. Rosei, Trends Biotechnol., 2015, 33, 637–652.
11. D. Campoccia, L. Montanaro and C. R. Arciola, Biomaterials, 2013, 34, 8533–8554.
12. A. Simchi, E. Tamjid, F. Pishbin and A. R. Boccaccini, Nanomedicine, 2011, 7, 22–39.
13. M. W. Whitehouse, Prog. Drug Res., 2015, 70, 237–273.
14. S. Loher, O. D. Schneider, T. Maienfisch, S. Bokorny and W. J. Stark, Small, 2008, 4, 824–832.
15. P. V. AshaRani, G. Low Kah Mun, M. P. Hande and S. Valiyaveettil, ACS Nano, 2009, 3, 279–290.
16. A. Pratsinis, P. Hervella, J. C. Leroux, S. E. Pratsinis and G. A. Sotiriou, Small, 2013, 9, 2576–2584.
17. H. Cao, X. Liu, F. Meng and P. K. Chu, Biomaterials, 2011, 32, 693–705.
18. H. Qin, H. Cao, Y. Zhao, C. Zhu, T. Cheng, Q. Wang, X. Peng, M. Cheng, J. Wang, G. Jin, Y. Jiang, X. Zhang, X. Liu and P. K. Chu, Biomaterials, 2014, 35, 9114–9125.
19. K. Bandow, A. Maeda, K. Kakimoto, J. Kusuyama, M. Shamoto, T. Ohnishi and T. Matsuguchi, Biochem. Biophys. Res. Commun., 2010, 402, 755–761.
20. Q. Xing, Q. Ye, M. Fan, Y. Zhou, Q. Xu and A. Sandham, J. Cell. Physiol., 2010, 225, 106–114.
21. H. Kadono, J. Kido, M. Kataoka, N. Yamauchi and T. Nagata, Infect. Immun., 1999, 67, 2841–2846.
22. E. M. Greenfield, Y. Bi, A. A. Ragab, V. M. Goldberg, J. L. Nalepka and J. M. Seabold, J. Biomed. Mater. Res., Part B, 2005, 72, 179–185.
23. J. L. Nalepka, M. J. Lee, M. J. Kraay, R. E. Marcus, V. M. Goldberg, X. Chen and E. M. Greenfield, Clin. Orthop. Relat. Res., 2006, 451, 229–235.
24. J. L. Nalepka and E. M. Greenfield, BioTechniques, 2004, 37, 413–417.
25. U. Zähringer, B. Lindner, S. Inamura, H. Heine and C. Alexander, Immunobiology, 2008, 213, 205–224.
26. S. K. Nelson, K. L. Knoernschild, F. G. Robinson and G. S. Schuster, J. Prosthet. Dent., 1997, 77, 76–82.
27. L. A. Bonsignore, J. R. Anderson, Z. Lee, V. M. Goldberg and E. M. Greenfield, Bone, 2013, 52, 93–101.
28. T. D. Herath, Y. Wang, C. J. Seneviratne, Q. Lu, R. P. Darveau, C. Y. Wang and L. Jin, J. Clin. Periodontol., 2011, 38, 694–701.
29. Y. C. Lu, W. C. Yeh and P. S. Ohashi, Cytokine, 2008, 42, 145–151.
30. Y. Zhao, H. Cao, H. Qin, T. Cheng, S. Qian, M. Cheng, X. Peng, J. Wang, Y. Zhang, G. Jin, X. Zhang, X. Liu and P. K. Chu, ACS Appl. Mater. Interfaces, 2015, 32, 17826–17836.
31. H. Cao, T. Cui, G. Jin and X. Liu, Surf. Coat. Technol., 2014, 256, 9–14.
32. C. Palacios, Crit. Rev. Food Sci. Nutr., 2006, 46, 621–628.
33. Y. H. An and R. J. Friedman, J. Investig. Surg., 1998, 11, 139–146.
34. T. P. Schaer, S. Stewart, B. B. Hsu and A. M. Klibanov, Biomaterials, 2012, 33, 1245–1254.
35. M. Lucke, G. Schmidmaier, S. Sadoni, B. Wildemann, R. Schiller, A. Stemberger, N. P. Haas and M. Raschke, J. Biomed. Mater. Res., Part B, 2003, 67, 593–602.
36. M. Hirschfeld, Y. Ma, J. H. Weis, S. N. Vogel and J. J. Weis, J. Immunol., 2000, 165, 618–622.
37. T. Kawai and S. Akira, Nat. Immunol., 2010, 11, 373–384.
38. H. Zreiqat, C. R. Howlett, A. Zannettino, P. Evans, G. Schulze-Tanzil, C. Knabe and M. Shakibaei, J. Biomed. Mater. Res., 2002, 62, 175–184.
39. Y. H. Leem, K. S. Lee, J. H. Kim, H. K. Seok, J. S. Chang and D. H. Lee, J. Tissue Eng. Regener. Med., 2016, 10, E527–E536.
40. H. Zreiqat, S. M. Valenzuela, B. B. Nissan, R. Roest, C. Knabe, R. J. Radlanski, H. Renz and P. J. Evans, Biomaterials, 2005, 26, 7579–7586.
41. J. Sugimoto, A. M. Romani, A. M. Valentin-Torres, A. A. Luciano, C. M. Ramirez Kitchen, N. Funderburg, H. Renz and P. J. Evans, J. Immunol., 2012, 188, 6338–6346.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra11854b

‡ These authors contributed equally.

---