**4. Discussion**

On the basis of our earlier works, we chose the anodic oxidation method as a surface modification of implants produced in 3D technology (SLS of Ti6Al4V ELI Grade 23 powder) [25]. Two types of amorphous coatings, which revealed suitable biointegration and antibacterial properties in preliminary studies, were selected for more comprehensive investigations, i.e., TNT5 (the ordered nanoporous) and TNT15 (the ordered nanotubular) [25,26]. Thanks to the anodic oxidation method, the TNT5 and TNT15 coatings covered the whole implant surface without cracks and gaps. This surface modification decreased the implants' hydrophobicity, whereas the enrichment with AgNPs caused the reverse effect (Table 3). Analysis of the XPS data confirmed that the surface of the TNT5 coating formed by a layer of titanium oxide, in which the titanium oxidation state is +4 (TiO2—100%). Meanwhile, the surface TNT15 layer should be treated as a mixture, which consists of Ti4<sup>+</sup> (TiO2—86%) and Ti3<sup>+</sup> (Ti2O3—14%) oxides (Tables 1 and 2, Figure S1) [43]. However, considering the earlier reports, we can assume that in water solutions, unstable oxides of titanium on the lower oxidation states will be oxidized up to TiO2[44],andthereforeinallbiologicalexperimentsTNT15canalsobetreatedasaTiO2layer.

The evaluation of biointegration properties of studied implants indicated that TNT5 nanoporous coating had the highest biointegration potential. TNT5 surface modification promoted proliferation of all tested cell lines while enrichment with silver nanoparticles inhibited proliferation of osteoblasts but not of fibroblasts cell lines. These results correspond to our previous findings, were we also noticed that TNT5 coatings enriched with AgNPs decreased proliferation of the MG-63 osteoblasts [30]. We have also observed that an increased nanotubes diameter (TNT15 coating) weakened the biointegration potential of the implant. Earlier works also showed that TiO2 nanoporous coatings with smaller pores diameter promoted osteoblast vitality and differentiation [45–47]. Moreover, cell growth on nanotubes of diameter larger than 50 nm was severely impaired due to the reduced cellular activity and an extensive programmed cell death [46]. As we have demonstrated, the presence of nanosilver on the surface of nanotubes has greater cytotoxic effect on osteoblasts than the diameter of the nanotubes themselves. Our results are in line with the findings of other authors, which indicate that nanosilver is toxic for osteoblasts and osteoclasts [48,49]. On the other hand, some experimental evidence show that TiO2 nanotubes coated with nanosilver are compatible to mammalian cells including osteoblasts [50,51]. The differences in the biocompatibility of biomaterials coated with nanosilver probably depend on the concentration and mode of AgNPs deployment on the surface of produced TNT coatings, and the rate of silver ion release to the body fluid environment [52]. An important part of our research was the determination of the inflammatory response elicited by the macrophage RAW 264.7 cell line cultured on the surface of modified implants in an inflammatory environment simulated with LPS. LPS is an outer membrane component of Gram-negative bacteria, recognized by the innate immune system as a sign of infection [53,54]. RAW 264.7 cells are widely used for inflammation studies due to the highly reproducible response to LPS derived from *Escherichia coli*, mimicing bacterial infection [55]. Our results clearly indicate that neither the investigated biomaterials nor the used dose of LPS (10 ng/ml) had any toxic effect on the RAW 264.7 macrophages. Moreover, all surface modification, besides AgNPs enrichment, promoted macrophages proliferation comparing to Ti6Al4V reference alloys. These findings are in line with Neacsu et al. [55], who also showed an increased macrophage proliferation on the nanotubes comparing to the unmodified titanium foils. Since macrophages play a key role in modulating early events in wound healing and interaction of macrophages with dental implant surfaces can be an important determinant of success of osseointegration [56,57], our results indicate

the biocompatibility of the tested nanomaterials. In the next experiments, we assessed the levels of pro and anti-inflammatory cytokines, released by macrophages growing on different implant surfaces. The pro-inflammatory IL-1β, IL-6 and TNF-α are produced predominantly by activated macrophages and are involved in the up-regulation of inflammatory reactions [58]. Similarly, NO is a prominent indicator of pro-inflammatory signal transduction in inflammatory response and antimicrobial defense [59]. In contrast, one of the major anti-inflammatory cytokines is IL-10, which inhibits the production of pro-inflammatory cytokines and mediators from macrophages and dendritic cells [60]. Our results showed that TNT5 and TNT5/AgNPs samples inhibited the LPS-induced release of pro-inflammatory cytokines and NO in comparison with the references Ti6Al4V alloy foils. In contrast, TNT15 and TNT15/AgNPs enhanced production of these mediators. Moreover, presence of AgNPs on the surface of nanotubes potentiated the anti-inflammatory activities of all tested specimens According to previous reports, silver nanoparticles show the potent anti-inflammatory effect and accelerate wound healing, however the possible cytotoxic effect on mammalian cells was also observed [49,61]. These results indicate that TNT5 coatings are good candidates for manufacture of implants with anti-inflammatory properties, since inflammation has been associated with both delayed bone healing and pathogenic bone loss [62].

The assessment of the genotoxicity of implants, which surface has been modified by producing TNT5 and TNT15 coatings, as well as their subsequent enrichment with AgNPs, was an important part of our studies. Genotoxicity is an ability of the agen<sup>t</sup> to directly or indirectly induce DNA damage. If not repaired by DNA repair system or eliminated by cell death, the damage might be retained in genetic material as a mutation and passed on to next generations. Accumulation of the mutations is causatively linked to many chronic diseases including cancer [63]. One of the severe mutagens are heavy metals, which can damage DNA directly by formation of adducts and intra- and inter- strand and DNA-protein crosslinks or indirectly through induction of massive oxidative stress. Therefore implants, especially long-term implants which remaining in the body cavity for long periods, i.e. months or years, should be scrutinized in terms of genotoxicity. Widely used in implantology titanium alloy (Ti6Al4V), a reference implant in our studies, was shown not to induce DNA damage [64]. However surface modifications of this alloy, especially with silver nanoparticles, are the source of potential DNA damaging molecules released from the implant coating, which was widely discussed in earlier reports [39,65,66]. For this reason, the biological systems enriched with AgNPs should be given special attention [40,66]. The molecular mechanism behind the genotoxic properties of AgNPs is still unsolved but involves the direct production of hydroxyl radicals and induction of oxidative stress resulting in DNA damage [66]. In the *in vivo* studies in mice the AgNPs could reach bone marrow and liver, and generate cytotoxicity to the reticulocytes and oxidative DNA damage to the liver [39]. The DNA damaging effect of NPs depends on their size, concentration and time of exposure [66,67]. Therefore it is important to analyze if silver released in a long term from the implant surface could induce DNA damage and result in mutations. In this study we took advantage of the first line in vitro gene mutation study recommended by OECD (TG 471) – bacterial reverse mutation test (Ames test), adjusted to verify mutagenicity of molecules released from the differently modified implant surfaces during 28 days incubation in PBS. Five tester strains were used to detect deletion, base substitution or frame shift mutations, depending on the tester strain's engineered genotype. None of the implant coatings tested, regardless of surface modification or AgNPs enrichment, released substances of mutagenic properties in any of the strains analyzed. This is a good prognosis for the investigated implant/TNT coating modifications discussed in this study. Their nanoporous (TNT5) and nanotubular (TNT15) morphology promotes implant biointegration and allows for controlled release of silver sufficient to kill bacteria and fungi and at the same time not inducing DNA damage.

The new generation of implants should not only facilitate their tissue integration but also prevent microbial colonization and biofilm formation. Serious medical problems associated with the introduction of implants to the human body are infections, which can lead to increased patients failure and mortality [68–71]. To solve this problem, the implant surface is modified by the formation of bioactive nanostructures (e.g., TiO2 nanotubes, nanofibers) and/or their enrichment with metal nanoparticles (mainly silver and copper) [20,27,28,72–74]. In our previous study it has been shown that TiO2 nanotubes formed on titanium alloy (Ti6Al4V), in particular the coatings obtained using low-potential anodic oxidation, possessed in vitro anti-biofilm activity tested on *S. aureus* model [25]. In the present work both amorphous titania layers (TNT5 and TNT15) and silver nanoparticles (AgNPs) were used to modify titanium alloy surface. Their antimicrobial potential against broad range of Gram-positive and Gram-negative bacteria, as well as fungi, was tested during direct contact of the microorganisms with biomaterial samples (anti-adhesive and anti-biofilm effect). Moreover, their exposition of analyzed microbials on the supernatants probably containing the components released from biomaterial samples (biocidal effect) has been evaluated. We demonstrated that all tested modified titanium surfaces were able to inhibit microbial colonization and biofilm formation in comparison to control Ti6Al4V. Similar to our previous study [28] anti-biofilm effect strongly depended on bacterial strain used. For instance, *S. aureus* ATCC 29213 (methicillin-susceptible *S. aureus* strain, MSSA) was more sensitive to direct contact with tested biomaterials less effectively colonizing modified surfaces than *S. aureus* ATCC 43300 (methicillin-resistant *S. aureus* strain, MRSA). Previously, inhibitory effect of TiO2 nanotubes and Ag grains on *S. aureus* ATCC 29213 biofilm was demonstrated, while biofilm formation by *S. aureus* H9 MRSA clinical strain was not affected in the same conditions [28]. Interestingly, we did not observe differences in anti-adhesive/anti-biofilm activity of TNT enriched and not enriched with AgNPs. Nanostructural modification of implant surfaces was suggested to limit direct microbial cell contact with such layer, which determine the ability of nanostructures to inhibit microbial colonization and biofilm formation [73,75–77]. The mechanisms of AgNPs antimicrobial activity are more complex and multidirectional, resulting from many targets in microbial cells for Ag+ activity, such as cell wall synthesis, membrane transport, including electron transport in respiratory chain, protein functions, as well as DNA transcription and translation [73,78,79]. Thus we could have expected that modification of titanium surface by both TNT and AgNPs would potentiate antimicrobial effect of such biomaterials. Especially since the antibacterial activity of AgNPs-enriched titanium coatings was demonstrated [28,75,80]. However, for the antimicrobial activity, Ag+ should be released from the nanoparticles in the nearest proximity of the microbes. As seen in SEM images, majority of AgNPs were inside or entrapped between TiO2 nanotubes, which limited the direct contact with the microorganisms during short-time studies. Therefore, demonstrated anti-adhesive and anti-biofilm activities of both TNT- and TNT/AgNPs layers were similar in short time. However, in long lasting experiments, the TNT/AgNPs biocidal activity was higher than Ti6Al4V and TNT-modified surfaces. TNT5/AgNPs-derived supernatant exhibited bactericidal activity after 2 weeks incubation and TNT15/AgNPs-derived one after 4 weeks, suggesting that the morphology of these layers can influence the release of Ag+ and thus their concentrations in the surrounding physiological fluids and tissues. Godoy-Gallardo et al. [81] assessed antibacterial effect of Ti dental implants modified by Ag (electrodeposition) *in vivo* using dog model of ligature-induced peri-implantitis. During long-lasting experiment (peri-implantitis was initiated 2 months after implantation and the effects were observed up to next 4 months) Ag+ release and their accumulation in the tissues around dental implants were demonstrated, which probably contributed to the reduced bacteria colonization of the implant surface. Moreover, a decreased bone resorption in Ag modified impants was shown, representing ye<sup>t</sup> another positive effect of an antimicrobial modification [81]. These results confirm our assumption that after implantation Ag ions release occurs *in vivo* and may modify the conditions in micro-niche influencing microbial growth, colonization, biofilm formation, and thus limiting inflammation. Interestingly, in our in vitro study fungicidal activity of TNT/AgNPs-derived supernatants was constantly very strong (almost 100% of *C. albicans* mortality), regardless the nanotubes type (TNT5/15) or incubation time (2/4 weeks), suggesting higher sensitivity of *Candida* cells to Ag+ than bacterial cells. Besinis et al. [75] also showed highly antibacterial activity of silver plated Ti6Al4V discs coated with nano-hydroxyapatite (Ag-nHA) and silver plated Ti6Al4V discs coated with micro-hydroxyapatite (Ag-mHA), causing 100% mortality of bacteria in surrounded media, which was attributed to a small but effective slow release of Ag from the layers. Similar to our results, Besinis et al. [75] in the study on colonization of modified titanium discs layers by oral streptococci also did not observe anti-biofilm activity against *Streptococcus sanguis*. However, the enrichment with Ag strengthened anti-biofilm activity. In our studies TNT/AgNPs samples also significantly reduced *S. gordonii* and *S. mutans* adhesion and biofilm formation (although not so spectacularly). Summarizing, the enrichment with AgNPs results in anti-adhesive and anti-biofilm properties of the titanium implants against microbial strains.

The results of biological studies indicate that Ti6Al4V implants with TNT5 or TNT5/AgNPs surface modifications exhibit most suitable properties (biocompatibility, immunological activity, lack of genotoxicity, and antimicrobial activity) for their use in the construction of implants, e.g. for the orthopedy. Therefore, these systems were chosen for surface roughness parameters (Sa) and mechanical properties determination. Sa parameter of the coatings used for implants is important in the case of human cells and tissue adhesion, cells proliferation and time of healing [76]. The high level of roughness ensures better tissue adhesion and primary stability between the implant and bone. It has also been proven that surfaces with higher roughness have a positive e ffect on the time of healing after implantation [77,78]. On the other hand, the increased roughness results in an increased surface area, which can encourage bacterial adhesion (such as *S. aureus)* and increase peri-implantitis occurrence [79]. Therefore, when designing the new generation of implants it is important to enrich their surface with the antibacterial protection, which in our case consisted from AgNPs. The deposition of silver nanoparticles on the surface of TNT5 layers led to smoothing of the surface and roughness reduction. A similar e ffect was noticed by Bahadur et al. for TiO2 layers doped by Ag nanoparticles [80]. In order to determine the biomechanical compatibility of biomaterials used in the construction of implants, especially long-term ones, it is important to determine Young's Modulus [26,42,82,83]. The results of earlier works revealed the influence of this factor on the surrounding living tissue, such as bone [84–87]. The significant di fference in Young's Modulus between implants and human bone (especially cortical human bone ~ 20 GPa) can induce bone loosening and reduced bone quality in the implant surrounding and in consequence loosening of the implant in the bone [88,89]. Considering obtained results, the lower value of Young's Modulus of the implant/TNT5 coating system, the more biocompatible it is. On the other hand higher nanohardness value obtained for Implant TNT5/AgNPs was similar for results reported for TiO2 [82,90]. Analysis of the distribution of nanomechanical property (nanohardness and Young's Modulus) confirmed the uneven distribution of the tested properties on the surface of the implants (Figure 11). The same e ffect was reported by Rayón et al. [42]. Obtaining a homogeneous distribution of nanomechanical properties was impossible due to the roughness of the samples and the geometry and structure of the nanotube. The results obtained confirm that the increase in the nanohardness value causes an increase in the Young's modulus. Increase of the nanomechanical properties values (H and E) increased H/E ratio, which describes the resistance to wear. The relationship between wear resistance and value of H/E ratio was reported [35]. Moreover, an increase in fracture toughness is attributed to higher values of Young's Modulus (E) and nanohardness (H) [91]. The obtained H/E ratio value correlated with nanoscratch-test results. The nanoscratch-test technique was used to study the adhesion properties of thin coatings or layers [42,82,92]. The forces used during implantation procedure may provoke the coating full delamination; therefore the coatings should have proper adhesion to the metallic substrate. Higher adhesion was obtained for the TNT5/AgNPs, which is attributed to the stronger metallic bonds, which occur. An important aspect in the context of implant modification is the determination of their compression resistance, but unfortunately conventional tests do not include nano-scale modification tests. The following parameters can indirectly indicate the strength of the coating: H/E, H<sup>3</sup>/ E2. Both parameters can be determined indirectly from the results obtained during nanoindentation measurements. The first parameter allows determining the wear resistance, while the second parameter allows determining the material's ability to propagate energy at plastic deformation during loading [93]. For the studied modifications, the value of the H/E parameter was 0.0054 ± 0.0039 and 0.0190 ± 0.0133, respectively for TNT5 and TNT5/AgNPs and H<sup>3</sup>/E<sup>2</sup> ~ 4.71 Pa and 4768.97 Pa for TNT5 and TNT5/AgNPs, respectively. These results indicate and confirm that the

presence of silver nanoparticles on the surface of TiO2 nanotubes significantly a ffects both the wear resistance as well as the material's ability to propagate energy at plastic deformation during loading which suggests better tribological and strength properties of the tested surface.
