Integrating silver compounds and nanoparticles into ceria nanocontainers for antimicrobial applications †

Silver compounds and nanoparticles (NPs) are gaining increasing interest in medical applications, specifically in the treatment and prevention of biomaterial-related infections. However, the silver release from these materials, resulting in a limited antimicrobial activity, is often difficult to control. In this paper, ceria nanocontainers were synthesized by a template-assisted method and were then used to encapsulate silver nitrate (AgNO3/CeO2 nanocontainers). Over the first 30 days, a significant level of silver was released, as determined using inductively coupled plasma optical emission spectroscopy (ICPOES). A novel type of ceria container containing silver NPs (AgNP/CeO2 containers) was also developed using two different template removal methods. The presence of AgNPs was confirmed both on the surface and in the interior of the ceria containers by X-ray diffraction (XRD), transmission electron microscopy (TEM) and scanning electron microscopy (SEM). Upon removal of the template by calcination, the silver was released over a period exceeding three months (>90 days). However, when the template was removed by dissolution, the silver release was shortened to #14 days. The antimicrobial activity of the silver-containing CeO2 containers was observed and the minimum bactericidal concentration (MBC) was determined using the broth dilution method. Investigation on human cells, using a model epithelial barrier cell type (A549 cells), highlighted that all three samples induced a heightened cytotoxicity leading to cell death when exposed to all containers in their raw form. This was attributed to the surface roughness of the CeO2 nanocontainers and the kinetics of the silver release from the AgNO3/CeO2 and AgNP/CeO2 nanocontainers. In conclusion, despite the need for further emphasis on their biocompatibility, the concept of the AgNP/CeO2 nanocontainers offers a potentially alternative long-term antibactericidal strategy for implant materials.


Introduction
Implant-associated bacterial infections remain a signicant issue.In the U.S.A., 4.3% of implants inserted into humans become infected. 1,2Bacteria that adhere at the site of implantation can create a biolm, 3 competing with the host tissues for integration with the implant as well as proper host cell integration, both essential aspects for the biocompatibility of any implant.Once bacteria adhere to a biomaterial surface, they are very difficult to be removed by the host immune system or displaced by the host cells. 4Thus preventing bacterial contamination during and aer surgery improves the host tissue integration of the implant.It has been demonstrated that the decisive period for preventing bacterial invasion relates to only 6 hours aer surgery. 5Infections can lead to additional pain for the patient as well as malfunction of the implant, which in turn leads to a need for its replacement, or in extreme cases to amputation or even death of the patient.From an economical point of view, bacterial infections also create considerable additional medical costs which could be avoided if bacterial invasion is prevented. 2Moreover, implants are being increasingly used and at the same time bacteria are becoming more resistant to conventional antibiotics. 6Therefore, there is an essential need for new developments in preventing and curing implant-related bacterial infections.
In this context, silver compounds and nanoparticles (NPs) are gaining more interest from the scientic community as a possible replacement for commonly used antibiotics, such as penicillin.Silver-containing materials may be used to coat implants in order to prevent bacterial biolm formation.However, these compounds may be too soluble and even promote adverse effects towards the host if the silver release is not controlled. 7Ideally, the release of silver ions should be able to kill bacteria without affecting biocompatibility.The release of silver ions, however, largely depends on the environment (i.e.pH, ionic strength, organic matter) and is therefore difficult to control in the human body.As we have shown in previous reports, it is possible to design the light stability and solubility of silver coordination polymers as efficient antimicrobial coatings for implant materials. 8,9However, with these compounds the silver release occurs permanently, irrespective of the presence of bacteria.
Another strategy to control the release of silver is the integration of silver NPs (AgNPs) into a biocompatible coating, 10 or directly into the biomaterials, such as in bone cements. 11In addition to such outlooks, encapsulation of AgNPs can be advantageous in order to control and prolong the release of different compounds, 12,13 in turn increasing the stability and biocompatibility of silver drugs.Encapsulation is also of great interest for many medical applications, such as targeted drug delivery. 14For example, it has been shown that encapsulated drugs are thermally more resistant. 13,15eria (CeO 2 ) is an excellent material for encapsulating drugs since it is highly stable for long time periods in aqueous solutions and elicits limited adverse biological effects. 16,17It has been demonstrated that CeO 2 -NPs can elicit certain antioxidant properties that protect cells against the onset of oxidative stress. 16,18The aim of the present study, therefore, was to determine if it were possible to specically encapsulate silver compounds and/or NPs within ceria-based container systems in order to slow down the silver release and thus making silver coatings increase their active antimicrobial period.Two approaches were envisaged to ll nanocontainers with silver compounds.A rst method was to make hollow spheres and then ll them with the desired compound by soaking, resulting in the AgNO 3 /CeO 2 nanocontainers.Secondly, it is also possible to include the AgNPs during the synthesis of the nanocontainers.Both methods are hereby described with the resultant material's characterization.The silver release was measured over a period of 3 months.The antimicrobial activity and the cytotoxicity of the novel AgNO 3 /CeO 2 and AgNP/CeO 2 containers were evaluated in prokaryotic and eukaryotic systems respectively.

Polystyrene nanosphere (PS NS) synthesis
Anionic PS NSs were synthesized via emulsion polymerization. 19he reaction mixture was prepared by mixing styrene (3.70 g, 35.5 mmol), KPS (0.30 g, 1.1 mmol), and SDS (0.09 g, 0.3 mmol) in water (250 mL).The solution was stirred at 80 C under argon for 42 hours.The PS NSs were then washed three times by centrifugation (30 min at 10 000 rpm).The supernatant was subsequently discarded and the pellets re-suspended in water.

CeO 2 container synthesis
The CeO 2 containers were formed via a sol-gel method: 13 PS NSs (0.3 g), Ce(acac) 3 (0.7 g, 1.6 mmol), PVP (0.3 g, 0.0075 mmol), urea (0.3 g, 5.0 mmol) and water (40 mL) were mixed for 5 minutes.The reaction was then allowed to proceed at 100 C for 4 to 5 days without any agitation.Urea was used to prevent aggregation but had no effect on the reaction. 13The suspension was centrifuged for 30 min at 15 000 rpm.The supernatant was discarded and the pellets were re-suspended in water using sonication.This washing step was repeated 3 times and the PS NSs coated with CeO 2 were dried in an oven at 40 C for 1 day.Hollow CeO 2 containers were obtained by removing the PS core by calcination in air in an oven at 600 C for 4 hours.

Encapsulation of AgNO 3 in CeO 2 containers
AgNO 3 was encapsulated into CeO 2 containers using a method adapted from Kartsonakis et al. 13 Briey, CeO 2 containers were placed in a closed vial and submitted to a vacuum.They were then rapidly immersed in a saturated solution of AgNO 3 in ethanol and the mixture was stirred at room temperature for 2 hours in the closed vial.The mixture was centrifuged and the supernatant was discarded.The AgNO 3 /CeO 2 containers were washed three times with fresh ethanol and collected by centrifugation (10 min at 10 000 rpm).They were then dried overnight at room temperature.

Silver nanoparticle (AgNP) synthesis
The method used to synthesize AgNPs was adapted from Evanoff and Chumanov. 20A 250 mL ask was lled with PS NSs (0.140 g) and silver(I) oxide (0.045 g, 0.2 mmol) in water (60 mL).The ask was then placed into an autoclave.The autoclave was lled with 10 bar H 2 and the reaction was allowed to proceed at 70 C for 4 hours with constant stirring.Aer cooling, the mixture was centrifuged (15 min at 15 000 rpm).The supernatant was discarded and the pellets were re-suspended in water (5 mL).

Encapsulation of AgNPs in PS NSs
The method used to encapsulate the AgNPs into the PS beads was adapted from Kumbhar and Chumanov. 21Aer the AgNP synthesis, the mixture (2 mL) was added to acetone (5 mL) and sonicated for 4 hours at room temperature.The resulting mixture was centrifuged (15 min at 15 000 rpm) and the pellets were resuspended in water (5 mL).

AgNP/CeO 2 container synthesis
The PS NSs encapsulating AgNPs were coated with CeO 2 , similar to the synthesis of the CeO 2 nanocontainers: the obtained AgNP/PS spheres (0.12 g), Ce(acac) 3 (0.50 g, 1.1 mmol), PVP (0.300 g, 0.0075 mmol), urea (0.20 g, 3.3 mmol) and water (30 mL) were mixed for 5 min.Then the reaction was allowed to proceed at 100 C for 4 to 5 days without any agitation.The suspension was centrifuged for 30 min at 15 000 rpm.The supernatant was discarded and the pellets were re-suspended in water using sonication.The washing step was repeated three times and the AgNP/PS NSs coated with CeO 2 were dried in an oven at 40 C for at least one day.Hollow AgNP/CeO 2 containers were obtained by calcination in air in an oven at 600 C for 4 hours, or by many successive dissolution treatments in boiling xylene containing triethylamine for 12 hours. 12Formed containers were then allowed to dry in an oven at 40 C.

Characterization
The synthesized particles were visualized by transmission electron microscopy (TEM) using a FEI/Philips CM-100 Biotwin microscope.The TEM grids used were 300 square mesh copper TEM grids covered by a carbon lm, provided by Electron Microscopy Sciences, USA.The particles were also visualized by scanning electron microscopy (SEM) using a FEI XL 30 Sirion FEG.SEM samples were prepared by spreading them on a carbon tape glued on a SEM holder.All images were obtained without sputter coating pretreatment.Images were obtained at different voltages using the Everhart-Thornley detector (EDT) or solid state detector (SSD) either in contrast or back-scattering mode.X-ray diffraction (XRD) patterns were measured using a STOE transmission X-ray powder diffraction system.Fourier transform infrared (FTIR) spectra were obtained with a Bruker Tensor 27 FTIR and a MKII Golden Gate Single Attenuated Total Reection System.Thermogravimetric analysis (TGA) was performed in a Mettler Toledo SDTA/TGA 851e ids.The silver concentration was determined by inductively coupled plasma optical emission spectroscopy (ICP-OES) using a Perkin Elmer Optima 7000DV ICP-OES.To assess the morphology of the A549 cells following exposure to the different CeO 2 nanocontainers, an inverted Zeiss laser scanning microscope (LSM) 710 Meta, Switzerland, was used.

Silver release
In order to determine the silver release from AgNO 3 /CeO 2 nanocontainers and AgNP/CeO 2 containers, 4 duplicates (75 mg) of each sample were prepared.Each duplicate was placed in a well, covered with distilled water (1.2 mL) and le at room temperature in the dark.At certain time intervals over a 3 month period, a fraction of the supernatant (0.500 mL) was collected and added to 1% nitric acid solution (5 mL) and was then ready for silver concentration determination.Freshly distilled water (0.500 mL) was placed back into the well.The silver concentration in each aliquot was determined by ICP-OES.

Antibacterial tests
The Escherichia coli (E.coli) K-12 bacterial strain was used for both zone of growth inhibition determination and minimum bactericidal concentration (MBC) determination.
Zone of growth inhibition determination.The zones of inhibition for CeO 2 , AgNO 3 /CeO 2 and AgNP/CeO 2 container powders were determined by the agar diffusion method. 22,23nitially, pellets of 90 mg of sample were made using a press under 3 tons for 1 minute.A single-cell colony was dipped in 1 mL LB broth and was incubated for 2 hours at 37 C. A total of 0.100 mL of the resulting bacterial broth was spread on an LB agar plate until dry.The container pellet of interest was then carefully placed on the plate.The plate was then incubated overnight at 37 C.The zone of growth inhibition, i.e. the zone on which no bacteria grew, was then measured according to the distance between the sample and the start of bacterial growth.
MBC determination.The broth macrodilution method was chosen for the determination of the MBC.The following method was adapted from the protocol described by Wiegand et al. 24 Briey, saturated E. coli in a LB medium was obtained by inoculating a single-cell colony in 1 mL of LB medium and incubating it at 37 C overnight.The next day, 50 mL of the saturation solution was inoculated in a fresh LB medium (5 mL) and incubated at 37 C for 2 hours.During this time, solutions containing different concentrations ranging between 6 and 240 mg mL À1 of CeO 2 nanocontainers, AgNO 3 /CeO 2 nanocontainers and AgNP/CeO 2 containers were prepared in the LB medium.Aer the two-hour incubation, the bacterial suspension was adjusted to the McFarland 0.5 standard. 25If not concentrated enough, the suspension was incubated for another hour and if too concentrated it was diluted with a fresh LB medium.This corresponded to approximately 1.5 Â 10 8 cfu mL À1 .This bacterial suspension was added to each of the nanocontainer suspensions in a 1 : 1 ratio.A sterility control, composed of a fresh LB medium, was also prepared.A total of 100 mL of each new suspension was then placed on a LB agar plate and incubated overnight at 37 C.The following day, the concentration at which total inhibition of bacterial growth occurred was recorded.

Cell culture and exposure
The A549 cell line, an adenocarcinomic human alveolar Type II epithelial cell culture (ATCC, USA), was cultured as previously described by Rothen-Rutishauser et al. 26 The experiments were performed on BD Falcon 2-chamber culture slides.Each well had a growth surface area of 4 cm 2 .The day before exposure, nanocontainer samples of 58 mg and 174 mg were suspended in a cell culture medium (2 mL) and added to a well of the chamber slide.The samples were then incubated overnight at 37 C and 5% CO 2 .The following day, the supernatant was removed and replaced with 2 mL of culture medium containing 0.5 Â 10 6 cells per mL.The samples were then incubated at 37 C and 5% CO 2 for 4 and 7 days.Aer each exposure period, the supernatant was removed and was used to determine the level of epithelial cell cytotoxicity.

Epithelial cell cytotoxicity
Cytotoxicity was determined by assessment of cell membrane damage (permeability) via measuring the release of the cytosolic enzyme lactate dehydrogenase (LDH) in the cell supernatant.
LDH is the enzyme responsible for the conversion of pyruvate into lactate in the glycolysis pathway and is present in the cytosol of eukaryotic cells.Upon disruption to the cell membrane LDH can be released from the cytosol and subsequently measured in the cell culture medium as a measure of cell damage (i.e.cytotoxicity).Following exposure to 58 and 174 mg per well for 4 and 7 days at 37 C, 5% CO 2 , LDH levels were determined using a diagnostic kit (Roche Applied Sciences, Switzerland).As a positive control, 0.2% TritonX-100 in PBS was used.A negative control of the cell culture medium only was also performed along with each sample series.All analysis was repeated in triplicate (n ¼ 3).

Epithelial cell morphology
Initially, following the exposure period, cells were xed and labelled as previously described by Lehmann et al. 27 Briey, samples were labelled with a 250 mL mix of a 1 : 50 dilution of phalloidin-rhodamine (label for cell cytoskeleton) and 1 : 100 dilution of 4 0 ,6-diamidino-2-phenylindole (DAPI stain for cell nuclei).Coverslips were then inverted and mounted onto microscope slides using Glycergel (Dako, Carpinteria, USA) and imaged by LSM.

Statistical and data analysis
The results from the silver release experiments and antibacterial tests are presented as the mean AE standard error of the mean.All cytotoxicity data are also presented as the mean AE standard error of the mean.This latter dataset was found to be normally distributed (data not shown) and so a parametric twoway ANOVA was conducted with subsequent Tukey's post-hoc test (SPSS, IBM, USA).Datasets were considered signicant if p < 0.05.Cytotoxicity analysis was performed in triplicate (n ¼ 3), while silver release experiments were performed in 4 duplicates (n ¼ 4).

Encapsulation of AgNO 3 into ceria nanocontainers
In a rst attempt, CeO 2 nanocontainers were prepared by a procedure similar to that described by Kartsonakis et al., 19 using PS NSs as a template.The slightly adapted synthesis route yields white PS NSs of 220 AE 15 nm diameter.Aer coating with CeO 2 and removing the core by calcination at 600 C, pale yellow nanocontainers with diameters between 175 and 200 nm and shell wall thicknesses of 12 to 20 nm were obtained.To introduce AgNO 3 into the containers, the nanocontainers were exposed to a vacuum in order to remove the air within the sample and were then immersed in a saturated solution of AgNO 3 in ethanol for 2 hours.In order to test if AgNO 3 was integrated into the containers, thermogravimetric analysis (TGA) was performed on the obtained grey CeO 2 nanocontainers previously dried in an oven at 60 C under a vacuum and on free AgNO 3 (Fig. 1).CeO 2 containers alone (Fig. 1A) did not have any signicant weight loss.Free AgNO 3 (Fig. 1B) underwent thermal decomposition at 370-430 C. The AgNO 3containing CeO 2 nanocontainers (Fig. 1C) showed a weight loss not only in the same temperature range, which corresponds to free AgNO 3 , but also at 30-220 C, 220-290 C and 780-900 C.This rst weight loss corresponds to desorption of physically adsorbed water.The weight loss at 220-290 C corresponds to the encapsulated ethanol that was used as a solvent during the encapsulation.Such a difference in thermal decomposition of ethanol is due to the protection of the encapsulated ethanol in the CeO 2 shell.The weight loss at 780-900 C still remains uncertain, although there is convincing evidence that this could be due to the formation of AgNPs in the CeO 2 shell from the reduction of AgNO 3 .
CeO 2 can be photosensitized. 28It may thus be inferred that simple day light is responsible for the blackening of the CeO 2 nanocontainers, reducing part of the Ag + from AgNO 3 to Ag(0).SEM images of these nanocontainers conrmed the presence of a few AgNPs (Fig. S1 †).Silver release from these AgNO 3 /CeO 2 nanocontainers showed an enhanced release during the rst 5 days aer immersion in water.Thereaer, the silver release considerably slowed and reached a threshold aer approximately 30 days (Fig. 2).This release pattern is comparable to the silver release from many other materials. 29,30

Synthesis of AgNP/CeO 2 containers
In order to further control the release of silver ions from the containers, the potential to encapsulate AgNPs within a CeO 2 nanocontainer shell was investigated.Depending on the pore size of the ceria container wall, it was expected that this process would be increasingly complex and so it was subsequently assumed that this encapsulation method would not be appropriate in this specic case.Thus, to include AgNPs in the nanocontainers during their synthesis an alternative method was used.Based on the work by Evanoff and Chumanov 20 and Kumbhar and Chumanov, 21 AgNPs of approximately 40 nm were synthesized via reduction with H 2 in the presence of PS NSs before being embedded within them.The so-synthesized AgNPs easily adhered to the PS NSs (Fig. 3).Aer sonication in an acetone/water solution, they were all encapsulated inside the beads, as conrmed by TEM images (Fig. 4).Following encapsulation, the formed hybrid AgNP/PS spheres became polydispersed, and their diameters varied between 220 nm and >1 mm, as determined by TEM.This diverse size range was attributed to the incorporation of the AgNPs and the coalescence between the PS NSs.
The hybrid AgNP/PS spheres were then coated with CeO 2 as described above.The X-ray diffraction (XRD) pattern (Fig. 5) showed that the particles were composed of ceria (JCPS no.34-0394) and elemental silver (JCPS no.04-0783), while polystyrene was not detected as a crystalline component.As determined by TEM and SEM, the diameter of the containers varied from 150 nm to 12 mm and the ceria shell thickness varied from 15 nm to 20 nm.Surprisingly, some AgNPs could also be seen on the outside surface of the CeO 2 shells (Fig. 6B).In order to construct hollow spheres, the PS core was removed either by heating the spheres to 600 C or via a dissolution process with xylene.Aer the PS removal, the AgNPs were still evidently present on the surface of the CeO 2 nanocontainers (Fig. 6D and F).The difference of texture seen in these SEM images of the containers can be attributed to the treatment method used to remove the core material.When the polystyrene core was removed by calcination, the nanocontainer surface contained numerous pores and cracks, although the containers remained closed (Fig. 6C and D).Alternatively, when the PS core was removed by dissolution, the containers conserved their smooth surface (Fig. 6E and F).With both methods, the polystyrene core was removed, as shown by the loss of bands corresponding to the polystyrene on the IR spectra (Fig. S2 †).
Open containers could be observed in only a few samples.SEM analysis clearly showed that the AgNPs were also present inside the container cavity (Fig. 6F), suggesting that the AgNPs are indeed well integrated within the ceria shell.The mechanism by which the AgNPs move from the container cavity to the outside of the ceria shell is still not fully understood.However, it can be hypothesized that by the diffusion process during the coating with CeO 2 , or indeed the transport process, which involves the extraction of the polystyrene core out of the containers, be it either by heating or dissolution, may also entrain partial transport of the AgNPs to the outer side of the containers.
The silver release from these hybrid AgNP/CeO 2 containers was determined by ICP-OES.In contrast to AgNO 3 /CeO 2 nanocontainers (Fig. 2), the silver release from AgNP/CeO 2 containers prepared by calcination was much slower (Fig. 7).The AgNP/CeO 2 containers obtained by calcination therefore enabled a silver release over a period of at least 3 months.Even aer this extended period of time, silver release was still found to be considerable.It is well known that silver ions, even at low concentrations, have an enhanced antimicrobial activity, 30 which suggests that this newly designed silver-containing material could potentially prevent bacterial invasion during the most important period following implant surgery.
On the other hand, AgNP/CeO 2 containers prepared by dissolution have a lower and faster silver release, i.e. over a period of less than a month (Fig. 8).This is explained by the dissolution steps in which a considerable amount of silver was washed away.The containers prepared by dissolution would    therefore be less efficient in the prevention of infections compared to the ones prepared by calcination.
To the best of our knowledge, this is the rst time that AgNPs are encapsulated into inorganic containers using a method in which AgNPs are rst encapsulated into a PS template.This resulted in well-dispersed AgNPs throughout the CeO 2 shell and a silver release exceeding 3 months.In addition, despite the fact that orthopaedic implants are normally in place for ten to een years, 31 only a few studies described the long-term release of silver from implant materials. 32The AgNP/CeO 2 containers obtained by calcination are therefore very promising for the development of long-term antibacterial implant coatings.

Antibacterial tests
Neither tests for the zone of growth inhibition or MBC determination showed any inhibition of E. coli growth for empty CeO 2 nanocontainers (Fig. 9A).On the other hand, both silvercontaining CeO 2 containers demonstrated heightened antimicrobial activity (Fig. 9B and C), indicating the intrinsic antimicrobial properties of silver.The zone of inhibition of AgNO 3 / CeO 2 nanocontainers was 4 mm in radius, while the inhibition zone of AgNP/CeO 2 containers was 2 mm in radius.AgNP/CeO 2 containers showed a slower release of silver, explaining why the AgNO 3 /CeO 2 nanocontainers elicited a larger inhibitory zone in such a short time period (i.e.overnight incubation).AgNP/CeO 2 containers prepared by dissolution also showed an inhibitory zone as for AgNP/CeO 2 containers prepared by calcination.
The MBC determination by the broth dilution method demonstrated that AgNO 3 /CeO 2 nanocontainers have a MBC of 93 AE 5 mg mL À1 (0.54 AE 0.03 mol L À1 ) and for AgNP/CeO 2 containers, the MBC is 117 AE 6 mg mL À1 (0.68 AE 0.03 mol L À1 ).These ndings further demonstrate the heightened antibacterial activity of silver-containing CeO 2 containers.The enhanced antibacterial activity of AgNO 3 /CeO 2 nanocontainers is consistent with their higher silver load being released aer a one-day    period (Fig. 2) compared to the silver release of AgNP/CeO 2 containers (Fig. 7).

Epithelial cell cytotoxicity and morphology
Fig. 10 shows that CeO 2 nanocontainers elicit no signicant increase (p > 0.05) in LDH release from the A549 cells at either time point or concentration tested.However, further analysis via LSM highlighted that these nanocontainers induced signs of apoptosis (e.g.shrinkage of the cell size and membrane as well as condensed nuclear regions (Fig. 10B)).In addition, the monolayer, as clearly shown in the sample treated with cell culture media only (negative control) (Fig. 10A), was completely disrupted by the CeO 2 nanocontainers.Thus, the lack of any detectable LDH release can be attributed to the fact that the cell membrane remained intact.Although CeO 2 NPs have been shown to elicit a heightened antioxidant effect, 33 the potential for this nanomaterial to cause apoptosis in A549 cells is also well known, as recently reported by Mittal and Pandey. 34While Mittal and Pandey 34 highlighted the mechanistic pathways associated with the onset of cell death following interactions (i.e.uptake) of CeO 2 NPs with A549 cells, in the present study the apoptosis observed can be strongly attributed to the rough surface of the CeO 2 nanocontainers that the cells were cultured upon.Although an imperative understanding of the raw form of the CeO 2 nanocontainers has been gained here, it is hypothesized that the effect of CeO 2 nanocontainers on cell morphology will be signicantly reduced following the impregnation of the CeO 2 nanocontainers into a polymer-based biomaterial for their eventual application.
Following exposure to AgNO 3 /CeO 2 and AgNP/CeO 2 containers respectively, the epithelial cell monolayer was also signicantly disrupted, as evident in Fig. 10C and D. With respect to the AgNO 3 /CeO 2 nanocontainer exposure, a signicantly lower (p < 0.05) LDH release compared to control levels was observed for both time points and concentrations tested.This effect, augmented further by the noticeable unspecic rhodamine-phalloidin staining throughout the sample, can be associated with the nanocontainers inducing immediate necrosis following exposure to the A549 cells.This explains the low LDH concentration in the culture medium for AgNO 3 /CeO 2 .Since cell death occurred rapidly aer exposure, cells could not build up a LDH load over the incubation periods, as was the case for the other samples and the control.This effect is most likely due to a rapid silver ion release from these nanocontainers that causes membrane disruption and actin depolymerisation, as previously reported by Cheng and colleagues. 35ith regard to the AgNP/CeO 2 nanocontainers, similar to the CeO 2 nanocontainers, a limited, non-signicant (p > 0.05) cytotoxicity was observed aer a 4 day exposure.Following a 7 day exposure, however, the AgNP/CeO 2 nanocontainers elicited a signicant increase (p < 0.05) in LDH activity at both concentrations tested.This effect can be attributed to a higher concentration of silver in the medium aer 7 days' exposure that constantly interacts with the cells during this prolonged period.Subsequent LSM analysis supported the LDH analysis and further indicated that AgNP/CeO 2 containers could also induce necrosis in the A549 cells, showing effects similar to the AgNO 3 /CeO 2 sample, although in a slower fashion (Fig. 10D).It is important to note, however, that the AgNP/CeO 2 containers elucidated a slower cytotoxicity than the AgNO 3 /CeO 2 containers due to a gradual, instead of immediate, silver release from the nanocontainers.More detailed descriptions of AgNP effects on cells can be found elsewhere. 32As an example, Mukherjee et al. 36 have demonstrated that biosynthesized AgNPs are more toxic towards cancerous cells, such as A549 cells, due to the cancerous cells' tendency to acidify their environment.The development of silver-containing materials can therefore nd use in biomedical applications as an antibacterial or anticancer agent.

Conclusions
In conclusion, it was demonstrated that it is possible to use CeO 2 nanocontainers to efficiently encapsulate either AgNO 3 or AgNPs.These materials demonstrated an enhanced antimicrobial activity against E. coli, whereas in human cells (A549 epithelial cells) all three samples, when tested in their raw form, were observed to induce a form of cell death (i.e.apoptosis (CeO 2 nanocontainers and AgNP/CeO 2 nanocontainers) or necrosis (AgNO 3 /CeO 2 and AgNP/CeO 2 nanocontainers)).The AgNP/CeO 2 nanocontainers are particularly interesting for the prevention of implant-related infections thanks to their longterm silver release.These effects, however, which can be attributed to the rough surface of the CeO 2 nanocontainers and the kinetics of silver release, could be controlled by encapsulating the materials into a biomaterial (i.e. a polymer).From these ndings, it can be suggested that with further emphasis towards increasing their biocompatibility, these newly designed AgNP/CeO 2 nanocontainers may offer an advantageous strategy towards potentially preventing bacterial infections following implant surgery.

Fig. 2
Fig. 2 Silver release as a function of immersion time in water for CeO 2 containers containing AgNO 3 .Data are the % mean AE standard error of the mean compared to the positive control (n ¼ 4).

Fig. 3
Fig. 3 TEM images of AgNPs attached to PS NSs after synthesis.

Fig. 4
Fig. 4 TEM images of AgNPs encapsulated in the PS NSs after sonication demonstrating the full encapsulation of nanoparticles.

Fig. 5
Fig. 5 XR diffractogram of AgNP/CeO 2 containers.Red lines and green stars correspond to the peak positions of ceria and silver XRD patterns respectively.

Fig. 6
Fig. 6 TEM images (A, C and E) and SEM images (B, D and F) of silver nanoparticles encapsulated in the PS NSs and coated with CeO 2 before the removal of the polystyrene core (A and B), and after the removal of the core via calcination (C and D) and via dissolution in xylene (E and F).Silver nanoparticles can be observed in all images.In SEM images, they clearly appear on the surface of the CeO 2 coating, as well as in the cavity of the broken containers (F).

Fig. 7
Fig. 7 Silver release as a function of immersion time for AgNP/CeO 2 containers prepared by calcination.

Fig. 8
Fig. 8 Silver release as a function of immersion time for AgNP/CeO 2 containers prepared by dissolution.Data are the % mean AE standard error of the mean (n ¼ 4).

Fig. 10
Fig. 10 Cytotoxicity (bar graph) of 58 mg per well (light grey) and 174 mg per well (black) of CeO 2 nanocontainers, AgNO 3 /CeO 2 nanocontainers and AgNP/CeO 2 containers after 4 and 7 days of exposure to A549 epithelial cells.Cell culture media only served as the negative control in both occasions (white bars).Data are the % mean AE standard error of the mean compared to the positive control (0.2% TritonX 100) (n ¼ 3).*Represents p < 0.05 compared to the negative control at 4 days.#Represents p < 0.05 compared to the negative control at 7 days.LSM images showing the XY panel and the XZ panel (below XY image) of the negative control (cell culture media only) (A), CeO 2 nanocontainers (B), AgNO 3 /CeO 2 nanocontainers (C) and AgNP/CeO 2 containers (D).The scale bar corresponds to 50 mm.The nuclei are stained with DAPI (yellow) and the F-actin with rhodamine-phalloidin (cyan).