By examining the effects of AgNO₃, CuCl₂, ZnCl₂, and MgCl₂ diluted in different concentrations in cell culture medium on the cell metabolism of L929 mouse fibroblasts, we found that silver, copper, and zinc salts were quite toxic to mammalian cells (Figure 1) at micromolar concentrations, in contrast to MgCl₂(millimolar concentration). All of the metal salts analyzed displayed a concentration-dependent cytotoxicity. The inhibitory concentration (IC50) values of AgNO₃, CuCl₂, and ZnCl₂ were very similar to each other (0.03, 0.08, and 0.07 mM, respectively). The IC50 values obtained for CuCl₂ and ZnCl₂ matches the values found by other groups [35, 36], whereas the IC50 value of AgNO₃ was 10 times higher than the value reported by Heidenau or Yamamoto. This might be because different test methods (MTS vs. WST vs. colony forming assay) and different incubation times (72 h vs. 8 h vs. 24 h) were used. In contrast to AgNO₃, CuCl₂, and ZnCl₂, MgCl₂ was tolerated by the L929 cells (Figure 1) in relatively high concentrations of up to 20,000 µM. The IC50 value of 39,600 µM for MgCl₂ was approximately 500-fold to 1400-fold higher, compared with CuCl₂, ZnCl₂, and AgN0₃, respectively. The relative low toxicity of magnesium has also been demonstrated in contact with other cell lines [37], showing comparable results. Primary human cells accept even higher concentrations of MgCl₂ [37, 38]. In addition, experiments with pure magnesium particles demonstrated matchable toxicity levels using rat osteosarcoma-derived cells [39].

Figure 1

Dissolved metal ion toxicity: Effect of (A) silver nitrate, (B) copper chloride, (C) zinc chloride, and (D) magnesium chloride on the metabolic activity of L929 mouse fibroblasts.

The antibacterial efficacy of AgNO₃, CuCl₂, ZnCl₂, and MgCl₂ was tested by using an agar diffusion test according to DIN 58940-3, revealing a decrease of toxicity in the order of Ag>Cu/Zn>Mg, which has been previously observed in the cytotoxicity test. With this method, all of the metal salts tested (except for MgCl₂) showed zones of inhibition to Staphylococcus epidermidis, Staphylococcus aureus, and Escherichia coli. The susceptibility of the bacteria to AgNO₃ with a concentration of 10 mM was 10 times higher than to CuCl₂ and ZnCl₂, in which concentrations of up to 100 mM were needed to form a zone of inhibition. However, MgCl₂ did not show any antibacterial efficacy when applied to bacteria in concentrations of up to 100 mM.

In a second step we tested TPU nanocomposites in terms of cytocompatibility and antibacterial efficacy. The nanocomposites were generated through a solution-compounding method based on pulsed laser ablation of metal targets (silver, copper, magnesium, and zinc) in organic solvents mixed with dissolved polymer. It was shown that polymer molecules stabilize the colloidal nanoparticles and prevent agglomeration processes along the whole process chain from particle generation until final product shaping [22]. Therefore, this technique enables the generation of nanocomposites with nanoparticles well dispersed in the polymer matrix without contact to each other and agglomeration [24]. As an example, homogeneously distributed silver nanoparticles in TPU are shown in Figure 4E. For noble metals like gold, platinum, silver or copper, pulsed laser ablation in liquid results in metal nanoparticles and no indication of bulk oxidation [40], with plasmon resonance spectra typical for elemental silver and copper nanoparticles and high ion-release capacities [27]. During laser ablation of base metals like magnesium or zinc in organic solvents, partial oxidation of the nanoparticles cannot be excluded, although nanocomposites with elemental magnesium embedded in a polymer have been fabricated recently by this method [41].

We found that extracts of TPU with nanosilver embedded in concentrations from 0.01 to 1.0 wt% were not cytotoxic to the L929 cells (Figure 2A). This effect was probably based on the relatively low silver ion release from these composites during the extraction time (Figure 3C), an effect that has been reported also by Hahn et al. [25]. The nanocomposite with 0.5 wt% silver released approximately 250 ng/cm² silver ions after 3 days of incubation in water (Figure 3C). This amount of silver corresponds to 0.007 mM, a concentration that is four times lower than the IC50 value of 0.03 mM calculated for AgNO₃. In contrast to the cytocompatible silver nanocomposites, extracts of the nanocomposites containing particles made of magnesium, zinc, or copper showed a concentration dependent behavior in the range of 0.01 to 1.0 wt%, in terms of cytocompatibility. For magnesium, low nanoparticle concentrations of 0.01 to 0.05 wt% increased the vitality of the L929 fibroblast (Figure 2G), whereas higher concentrations decreased the cytocompatibility. For copper (Figure 2C) and zinc (Figure 2E), concentrations of up to 0.1 wt% were nontoxic, while toxic effects were observed in the extracts of the composites containing 0.5 wt% or more nanoparticles.

Figure 2

Effects of single nanoparticles; cytotoxicity (A, C, E, G) and antibacterial efficacy (B, D, F, H) of TPU composites with different concentrations of embedded metal nanoparticles (silver, copper, zinc, and magnesium): (A, B) Ag-TPU, (C, D) Cu-TPU, (E, F) Zn-TPU, and (G, H) Mg-TPU.

Figure 3

Effects of nanomaterial mixtures: (A, D) cytotoxicity, (B, E) antibacterial efficacy, and (C, F) cumulative ion release of TPU composites with embedded nanoparticle mixtures of two elements, each element 0.5 wt% (silver, copper, zinc, and magnesium).

In contrast to the common view that eukaryotic cells are more sensitive than bacteria to toxic agents, we found that this was not the case for Ag-TPU nanocomposites. The cytocompatible-silver nanocomposites containing 0.05 to 1.0 wt% silver nanoparticles were efficient in reducing bacterial growth of S. epidermidis (Figure 2B). In this range, the “therapeutic window” is widest at 0.5 and 1.0 wt% silver. An antibacterial effect without cytotoxicity has been previously reported by Alt et al. [42]. One reason for this effect could be the interaction of the silver ions with cellular proteins [43, 44], thereby interfering with the metabolism and energy storage [45] and the membrane integrity [46]. Eukaryotic cells cope better when challenged with silver ions because they are a bigger target, and they show a higher functional and structural redundancy than bacteria. Another reason might be that the medium used for culturing mammalian cells contains more protein than the medium required for bacteria growth [47, 48], and the serum protein albumin has been shown to reduce the biological effect of silver nanoparticle composites [49]; therefore, in the case of testing the effects on L929 fibroblasts, fewer free silver ions might be available. One also has to consider that we compared effects on cell metabolism with effects on bacterial proliferation.

Compared with silver-TPU, nanocomposites containing nanoparticles made of copper (Figure 2C), zinc (Figure 2E), or magnesium (Figure 2G) were less cytocompatible to L929 in high particle concentrations; however, they did not reduce S. epidermidis proliferation (Figure 2D, 2F and 2H). This result was rather unexpected in the case of copper because its antibacterial efficacy has been known for centuries [43, 50]. Regardless, agar diffusion tests indicated a 10 times lower susceptibility of S. epidermidis to copper salts compared with silver salts (data not shown). Taking into account that the release of copper ions from the TPU nanocomposites was in the same mass range as the release of silver ions (compare Figures 3C and 3F), a different susceptibility of S. epidermidis to the corresponding metal salts might explain the observed effect.

To test our hypothesis that mixtures of metal nanoparticles of different biological efficacy embedded in TPU change the outcome of single metal nanocomposites, we analyzed in a third approach TPU nanocomposites containing mixtures of metals. Therefore, we mixed a noncytotoxic but antibacterial concentration of 0.5 wt% nanosilver with 0.5 wt% copper, 0.5 wt% zinc, or 0.5 wt% magnesium, respectively, to see how this might alter the therapeutic window of the nanosilver TPU composite. For comparison, we also tested how the addition of 0.5 wt% zinc, 0.5 wt% silver, or 0.5 wt% magnesium, respectively, to 0.5 wt% copper-TPU might change its ion-releasing and bioactive properties.

In the case of 0.5 wt% silver nanocomposites, the addition of 0.5 wt% magnesium or 0.5 wt% copper decreased cytocompatibility (Figure 3A). The combination of silver with 0.5 wt% zinc resulted in the only noncytotoxic nanocomposite of all metal nanoparticle mixtures tested (Figure 3A). The measured cell vitality slightly decreased compared with 0.5 wt% silver in TPU (Figure 2A); and in comparison with the nanocomposite containing only 0.5 wt% zinc, it increased (Figure 2E).

The effects of mixing silver with zinc, copper, or magnesium nanoparticles on the antibacterial properties of the nanocomposites with respect to S. epidermidis were quite different from their actions on L929 cell metabolism. Although the combination of silver with magnesium increased the antibacterial efficacy of the TPU nanocomposite, the mixing of silver and zinc did not change the antibacterial behavior (Figure 3B). In contrast, the mixing of silver and copper reduced the antibacterial efficacy compared with the TPU composite with 0.5 wt% nanosilver (Figure 3B). The antibacterial efficacy of copper TPU nanocomposites increased only slightly by mixing copper with silver or magnesium (Figure 3E), whereas the mixing had no effect on L929 fibroblast vitality (Figure 3D). All nanocomposites mixed with copper nanoparticles were slightly cytotoxic (Figure 3D), with the lowest toxicity resulting from the combination with zinc, which in turn had no antibacterial properties.

The results for the antibacterial efficacy of the mixed nanocomposites (Figure 3B) can be explained by the ion release (Figure 3C). The release of silver ions resulted in a relatively strong ion release rate from bulk nanocomposites in the beginning and decreased over time, but did not exhibit a burst release known for coatings made of nanocomposites [51]. Combining silver nanoparticles with nanoparticles made of magnesium increased the silver ion release, and consequently the antibacterial properties as well. In the presence of copper nanoparticles, the silver ion release decreased drastically, which resulted in a strong decrease in antibacterial activity. The silver ion release from the mixture of silver with zinc was only slightly affected compared with the pure silver TPU nanocomposite, which resulted in almost unchanged antibacterial efficacy.

The liberation of copper ions from the mixed TPU nanocomposites was also affected by the mixture with other metals (Figure 3F). In the case of the combination with silver, there was a slight increase in the copper ion release. Anyhow, the stronger antibacterial effect for this combination (Figure 3E) was most probably caused by the release of silver and not the slightly accelerated copper ion release. The addition of zinc or magnesium led to a decrease in the copper ion release. According to Hahn et al., silver ion release kinetics differ from copper ion release, being rather limited by oxidation processes and diffusion rate, both parameters strongly influenced by the properties of the surrounding polymer matrix [25].

In general, the results concerning ion release from mixed nanocomposites are in accordance with the expectations based on the electrochemical potential. For instance, the release of copper ions from copper nanoparticles is accelerated in presence of the more noble silver, whereas at the same time the release of silver ions is decelerated compared with the pure silver nanocomposite. This electrochemical effect has been previously reported for silicone Ag-Cu nanocomposites by Hahn et al. [27]. Furthermore, the presence of less noble magnesium or zinc decelerates the release of copper ions. Overall, a decelerated release of noble metal ions and accelerated release of the less noble counterpart is observed if both are present in electrolyte solution (here, polymer composite immersed in water or cell culture media). Hence, the combination of silver with copper or silver with zinc behaved as to be expected by their position in the electrochemical series.

Anyhow, based on the electrochemical potential, it was not expected that a combination of silver with magnesium would result in a stronger silver ion release. The comparison of the effects of the metal nanoparticle doping was carried out at an identical mass load. The density of zinc is four times higher than magnesium, so the volumetric concentration of magnesium inside the polymer is four times higher than zinc (and five times higher than copper) at the same mass load. Ion release and dissolution of magnesium would possibly cause far larger voids or diffusion channels inside the polymer matrix than more dense metals. Similarly, oxidation of the identical mass of magnesium would lead to a far higher volume of metal hydroxides or other oxidized metal species formation inside the polymer compared with zinc or copper. Accordingly, we observed a more pronounced swelling of the TPU polymer composite when doped with magnesium than for all the other elements tested, which was caused by higher water uptake (Figure 4C) [20]. Hence, in addition to an electrochemical effect, the volumetric difference that causes higher voids or channels inside the polymer matrix after dissolution, causing an increased water uptake, could also contribute to the observed amplification of silver ion release by magnesium.

In addition, changes in the oxidation and ion-release rates resulting from local electrochemical cell formation (contact corrosion) cannot be excluded because we did not perform transmission electron microscopy (TEM) studies on mixed nanocomposites to clarify this issue. However, because of the special characteristics of the synthesis process, achieving polymer grafting on the nanoparticles in situ [24], we expect that the nanoparticle mixture is homogeneously distributed throughout the whole polymer matrix without agglomeration, avoiding direct particle-particle contact.

TEM studies would also show whether elemental magnesium or oxidized species or even passivated metal-metal oxide core-shell particles are present in the polymer matrix. However, in a previous ion release and ultrastructure analysis by TEM on silver and copper nanoparticle polymer composites, passivation layers around the nanoparticles during ion release were not reported. TEM analysis of cryo-sectioned copper and silver nanoparticle polymer composites revealed holes and edges after 84 days immersion, indicating anisotropic dissolution [27]. The analysis of the nanoparticle shape change indicated a corrosion of nanoparticles throughout the bulk of a composite [27], and second-order kinetics were identified for the ion release of the copper nanoparticle composite [25]. Regarding magnesium polymer composites, it may be of relevance that crystal structure analysis of laser-generated magnesium nanoparticle polymer composites has recently been provided evidence for the unoxidized state of the magnesium inside the polymer, at least before immersion into a liquid [41].

Finally, to investigate the effects of nanocomposites with mixed magnesium and silver in detail, we tested combinations of 1 wt% silver nanoparticles with increasing magnesium nanoparticle concentrations up to 1 wt% embedded in TPU (Figure 4). Increasing amounts of magnesium nanoparticles resulted in an increased release of both silver and magnesium ions from mixed TPU nanocomposites in a concentration-dependent manner (Figures 4A and 4B). In parallel, the presence of magnesium in the polymer matrix resulted in an increased water uptake (Figure 4C). As a consequence of the higher amount of accessible (dissolved) silver ions, the antibacterial efficacy against S. epidermidis was also improved, depending on the concentration of magnesium nanoparticles embedded in silver TPU (Figure 4D). Furthermore, this combination also exhibits an improved antibacterial effect against S. aureus, whereas silver nanoparticles alone do not prevent the growing of S. aureus on the surface of 1% Ag-TPU nanocomposites (Figure 4F). In other words, the most promising material combination was silver with magnesium.

Figure 4

Effects of nanomaterial mixtures: (A) silver and (B) magnesium cumulative ion release, (C) water uptake, and (D) antibacterial efficacy of mixed TPU composites with embedded combined nanoparticles (1.0 wt% silver with 0 to 1.0 wt% magnesium); STEM analysis of silver nanoparticles (black dots) embedded in (E) TPU matrix and of (F), white dots in polymer correspond to silver nanoparticles) S. aureus grown on 1% Ag-TPU nanocomposites.

Although magnesium did not lead to a quantitative expansion of the therapeutic window, it significantly accelerated the release of silver ions. The detected faster ion release within the first days might have the potential to reduce the risk of bacterial infection, which is especially high immediately after the treatment with the medical device. From the viewpoint of clinical use, a medical device with antimicrobial protection based on the silver ion release should be able to be fully functional at the initial state after contact with biofluids or implantation if the substitution of conventional antimicrobial pharmaceuticals by silver is desired. However, this is generally not the case, because the silver ion release has to overcome a lag phase. This is the result of first-order kinetics of the ion release [27] and the comparable slow dissolution of silver. Immediate antimicrobial function would require either (1) pre-immersion of the composite, (2) combination with conventional antibiotics, or (3) a higher silver dose, causing higher dissolution rates. Unfortunately, this higher dissolution rate in the beginning would definitely cause the risk to exceed toxic levels during longer ion-release time.

The addition of magnesium to silver nanoparticle composites might be an alterative to solve this problem because faster ion release in the critical initial phase has been caused indirectly by magnesium without the need to increase the amount of silver in the matrix, hence reducing the duration of the lag phase by increasing the ion release in the time span during the first days after application of the medical device. Clearly, fine-tuning of the magnesium to silver ratio and overall load would be required to see whether a larger therapeutic window compared with pure silver composites can be achieved by the nanoparticle mixture; further ion release tests and in vitro studies need to be done before the indication is strong enough to start in vivo trials.

In conclusion, the results of this study show that silver ion-releasing TPU composites possess antibacterial properties against S. epidermidis without being cytotoxic to mammalian cells starting at 0.05 wt%, and with an optimal mass load of 0.5 wt% silver in the polymer. In contrast, nanocomposites with copper, zinc, or magnesium nanoparticles did not cause any antibacterial effect at nanoparticle concentrations up to 1.0 wt% in TPU.

The combination of metals in TPU nanocomposites resulted in alterations of ion release, water uptake, cytotoxicity, and antibacterial efficacy. In general, the results concerning ion release from mixed nanocomposites are in accordance with the expectations based on the electrochemical potential, with decelerated release of noble metal ions and accelerated release of the less noble counterpart. In contrast, the mixture of silver with magnesium in TPU nanocomposites led to accelerated release of silver ions, probably caused by a volumetric effect of the less dense Mg inside the polymer, resulting in an increased water uptake of the nanocomposites, which increased the silver ion release especially in the first days but did not yet result in an enhanced therapeutic window. Nevertheless, magnesium might be a tool to overcome the problem of delayed silver ion release and to further decrease the amount of silver required to gain antibacterial effects, thereby reducing the negative effect of silver on the polymer properties and the overall material costs. A more comprehensive evaluation of the antibacterial efficacy and safety of combined silver and magnesium nanocomposites would need further in vitro studies on different polymers using a wider spectrum of mammalian cell types and, if successful, additional in vivo studies.

Thermoplastic polyurethane nanocomposites were generated based on the description in Wagener et al. [24]. The polyether-based TPU polymer (BASF Polyurethanes GmbH, Elastogran 1190 A) was kindly provided by B. Braun Melsungen AG, Germany. In brief, nanoparticles were generated using picosecond-pulsed laser ablation of bulk metal targets (silver, copper, magnesium, zinc) in tetrahydrofuran (THF). The THF was doped before the ablation process with 0.3 wt% of dissolved TPU for in situ stabilization and improved nanoparticle dispersion in the solid nanocomposite. Afterward, the tetrahydrofuran was evaporated, which resulted in solid TPU nanocomposite master batches with a nanoparticle concentration of 1.0 wt%. Solid samples were fabricated using solution casting on glass plates (10-mm diameter), followed by slow evaporation of the THF in a closed vessel. After complete evaporation, the nanocomposite film samples were removed from the glass plates. For adjustment of various particle concentrations (0.01 to 1.0 wt%), defined amounts of TPU nanocomposite master batches and pure TPU were dissolved in THF with an overall fixed TPU concentration of 10%; this viscosity can be used for sample preparation using solution casting.

For the ion-release measurements, TPU nanocomposites (n=3) were placed in cavities filled with 1.5 mL of ultrapure water (Millipore) each. Samples of the supernatant were taken at day 1, 3, 6, 10, 15, 20, 24 and 30, and nitric acid (1.0 vol-%) was added for the stabilization of ions during the storage until measurement. Silver, copper, zinc, and magnesium ions were detected using a graphite platform atomic absorption spectroscope (AAS, nova 400, Analytik Jena). The metal ion concentrations (ng/cm²) represent the mean values of the corresponding triplicates. The concentration of ions released by pure TPU was analyzed in parallel and found to be <1% of the absolute metal ion concentration measured for the TPU nanocomposites (data not shown).

Water uptake of nanocomposite samples was analyzed by differential weighting of dry and immersed samples after being placed for 1 day in ultrapure water (Millipore). For weighting, a microbalance was used.

Mouse fibroblasts (L929; DSMZ, Braunschweig, Germany) were seeded at a density of 1×10⁴ cells/mL in 96-well or 48-well cell culture plates and incubated at 37°C, 5% CO₂, and 90% relative humidity. After 24 h, the respective culture media were discarded and replaced by metal salts diluted in a range from 5 µM to 10.000 µM in a cell culture medium (RPMI with 10% FCS, both obtained from Gibco, Germany) and further incubated with L929 for 3 days (n=4 in a single experiment, each concentration was tested in two to six individual experiments). Vitality was measured photometrically at OD 490 nm by using the MTS assay (CellTiter 96^® AQueous One Solution Cell Proliferation Assay, Promega, Germany). The culture medium was used as a background control and subtracted from all of the values to obtain corrected values to calculate the median. On the basis of the medians of the single experiments, the median, upper and lower quartiles, and the maximum and minimum depicted in the chart were calculated. The cells incubated with growth media only were used as a negative control (100% vitality). Vitality values below 70% were classified as cytotoxic according to international standards [33].

To test the effect of soluble substances released in vitro from the metal nanocomposites on L929 cell metabolism, liquid extracts were prepared from nanocomposites and pure TPU as a reference in cell culture media (RPMI+10% FCS) using a surface-to-extraction volume ratio of 3 cm²/mL for 72 h at 37°C.

L929 cells were seeded at a density of 7.5×10³ cells/mL in 96-well cell culture plates and incubated at 37°C, 5% CO₂, and 90% relative humidity. After 24 h, the respective culture media were discarded and replaced by the extracts and further incubated with L929 cells for 72 h. The testing of the metabolic activity was performed as described above.

The effect of the nanocomposites on S. epidermidis (DSM 20044, DSMZ) surface attachment and proliferation were measured using a modified version of the method described in Bechert et al. [52]. The TPU nanocomposite test items were extracted in 0.9% sodium chloride (NaCl, Merck, Germany) with 0.1% bovine serum albumin (BSA, Sigma, Germany) for 24 h, dried and incubated again with 4×10⁶ CFU/mL S. epidermidis in 1 mL of cell suspension in a 48-well microtiter plate at 37°C for 1 h. For each test item, a minimum of four replicates were examined. After incubation, the test items and controls were washed three times in phosphate-buffered saline (PBS, Biochrom, Berlin, Germany) to remove loosely attached bacteria from the surface. The test items and controls with the remaining adherent cells were incubated in PBS with 0.2% (NH₄)₂SO₄(Merck, Germany), 0.25% glucose (Roth, Germany), and 1% sterile trypton soy broth (TSB, Oxoid, Germany) for 18 h at 37°C in fresh 48-well microtiter plates. The nanocomposite test items were removed and aliquots of the bacterial suspension were transferred to a 96-well microtiter plate. After adding TSB, the proliferation of the bacteria was monitored at 578 nm using a microtiter plate reader (µQuant or Synergy MX, Biotek) up to 24 h. The time needed for reaching an optical density (OD) of 0.2 was defined as the onset OD time. This time was recorded for each well of all of the nanocomposite-microorganism combinations tested, and the onset OD time of TPU (negative control) was subtracted to obtain the delta onset OD. The median, upper and lower quartiles, and the maximum and minimum were determined from the delta onset time of the replicates. Test items with a delta onset time of 6 h and more were considered to be antibacterial because this time frame corresponds to a log 3 reduction of bacterial growth (99.9%), compared with the control (TPU).

Scanning transmission electron microscopy (STEM) analysis was performed on thin silver-TPU nanocomposites film, arisen on copper grid after dipping in colloid. For visualization of S. aureus growth on 1% Ag-TPU nanocomposites, the bacteria were fixed on 1 wt% Ag-TPU nanocomposites with 3% paraformaldehyd (Fluka, Germany) for 20 min. at room temperature after the measurement of the antibacterial efficacy and further subjected to SEM analysis in low vacuum-mode.