Effect of graphene oxide loaded on TiO2-nanotube-modified Ti on inflammatory responses

Although a titanium matrix modified with titanium dioxide nanotube (TNT) arrays can have anti-inflammatory effects in vitro, these effects are limited. In this study, the TNT surface was modified by electrodepositing graphene oxide (GO) to enhance the anti-inflammatory effect of the material. Scanning electron microscopy (SEM), Raman spectroscopy, and x-ray diffraction were used to characterize each of these materials. Cell Counting Kit-8 (CCK-8) was used to determine the cell proliferation status. Enzyme-linked Immunosorbent Assay (ELISA), immunofluorescence staining, and RNA sequencing were used to assess the regulation of inflammation in each group. Raman spectroscopy confirmed that GO was successfully loaded onto the surface. The SEM, ELISA, fluorescence staining, and RNA sequencing results indicated that TNT-GO can effectively inhibit the inflammatory response and induce the M2 polarization of macrophages. TNT-GO can weaken the surface inflammatory responses of materials, suppress the secretion of pro-inflammatory factors, and promote the M2 polarization of macrophages. These advantageous properties render TNT-GO a promising material for dental implants.


Introduction
Although titanium (Ti) is widely used in dental and orthopedic implants because of its good biocompatibility, corrosion resistance, and excellent mechanical properties, Ti implants may perform unsatisfactorily because of weak bonding to bone tissue, bacterial infection, and poor tissue healing [1,2].The insertion of a Ti implant into bone may cause host proteins to non-specifically adsorb on the surface of the Ti implant, which could trigger a foreign body reaction [3].In this regard, macrophages, as the first cells to colonize the surfaces of implants after implantation, play a decisive role in tissue healing.Macrophages receive inflammatory signals in an autocrine and paracrine manner and transmit signals by producing inflammatory factors and growth factors, thereby regulating the microenvironment around damaged tissue [4,5].Excessive inflammatory response often leads to active osteoclasts, poor bone healing, or alveolar bone resorption, which would cause implant failure.An appropriate inflammatory response, on the other hand, is conducive to the removal of foreign body pathogens, and inflammatory factors such as nitric oxide can also promote vascular regeneration [6,7].Therefore, the osseointegration process of implants may depend on a balanced and controllable immune response.
Macrophages are activated upon migration to the site of injury and exhibit two polarization states.The M1 phenotype promotes inflammatory response, whereas the M2 phenotype weakens inflammatory response and accelerates tissue repair and remodeling [8][9][10].The M2 phenotype plays a significant role in wound healing and the immune regulation of implants, which serves to accelerate tissue remodeling.Therefore, it is assumed that modifying the surface of the implant to endow it with immunomodulatory activity would be an effective approach to suppress inflammatory reactions, accelerate tissue healing, and promote soft tissue closure.
Graphene oxide (GO) [11][12][13], a derivative of graphene with a large specific surface area and abundant functional groups, has good hydrophilicity, dispersibility, and biocompatibility.The rough microscale folding structure of GO is known to be capable of regulating the tension of the cytoskeleton [14,15].These desirable properties suggest that GO would be a promising surface coating for implants.However, the regulatory effect on macrophages of the presence of a GO coating on the surface of titanium dioxide nanotube (TNT) is unclear.
In this study, a new material, TNT-GO, by electrodepositing GO onto the surface of TNT.Then, mouse monocyte macrophages (RAW 264.7) were cultured in vitro on the surfaces of pure Ti sheets, TNT, and TNT-GO to investigate the regulatory ability of the modified materials to assess their inflammatory responses.

Materials and methods
A Ti sheet (10 mm × 10 mm × 0.2 mm) was placed in a centrifuge tube (50 ml) and washed ultrasonically with acetone, anhydrous ethanol, and deionized water for 10 min, after which it was allowed to dry at room temperature.TNTs were prepared via anodic oxidation.In brief, pure Ti sheets were anodized in an ethylene glycol electrolyte containing 10 vol.% deionized water and 0.5 wt% ammonium fluoride at 50 V for 15 min, followed by annealing at 550 °C for 2 h to obtain TNT arrays.After ultrasonic dispersion of sheet GO (0.1 mg ml −1 , 100681, XFNANO, China) for 2 h, it was loaded onto the TNT surface via electrodeposition at a load voltage of 50 V for 5 min to obtain TNT-GO.

Sample characterization
The surface morphology of the material was examined via scanning electron microscopy (SEM, S4800, Japan).The reverse sides of the three groups of specimens were coated with conductive adhesive, marked, and pasted on the sample disk in sequence.The acceleration voltage was 10 kV, and the surface morphology of the specimens was observed at magnifications of 30,000 × and 50,000 ×.Raman spectra were recorded using a LabRAM HR800 spectrometer (HORIBA, USA) in the range of 1000-3000 cm −1 .The surface roughness of the sample was examined using atomic force microscopy (AFM, Nanoscope V, USA).The contact angle of the sample was measured using an optical contact angle measuring device (OCA15pro, Germany).X-ray diffraction (XRD, TTRAX III, Japan) was used for analyzing the crystal phases.

Cell culture
Commercial RAW264.7 (China Cell Line Resource Infrastructure, China) cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin under standard culture conditions (37 °C, 5% CO 2 ).All reagents used in this experiment were provided by Thermo Fisher Scientific (United States).

Macrophage morphology
RAW 264.7 cells were seeded on the different samples in 24-well plates at a density of 1 × 10 4 cells per well for 4 and 24 h, followed by cell fixation, dehydration, drying, and Au plating.Then, the morphology of the RAW 264.7 cells was examined via SEM.

Cell viability
The cell immunofluorescence method was used to detect the state of the RAW 264.7.RAW 264.7 cells were seeded on the different samples in 24-well plates at a density of 4 × 10 4 cells per well for 24 and 48 h.After incubation for 24 or 48 h, diluted calcein AM (250 μl) was added to each well.After incubation for 30 min, the calcein AM solution was discarded and diluted PI (250 μl) was added.The samples were then incubated for another 30 min prior to their observation by a confocal laser microscope.

Cell proliferation
The CCK-8 assay (Dojindo, Japan) was used to detect the proliferation of the RAW 264.7 cells.These cells were seeded on the surface of each group in 24-well plates at a density of 5 × 10 4 cells per well and cultured for 4, 24, or 48 h, whereupon CCK-8 was added, and the cells were incubated for 2 h in the dark.The absorbance was then measured at 450 nm using a multifunctional enzyme marker.
2.6.RNA sequence analysis RAW 264.7 cells were seeded onto the specimens in 24-well plates at a density of 2 × 10 5 cells per well, and after 48 h, the total RNA was extracted using TRIzol (Invitrogen, Carlsbad, CA, USA).Transcriptome sequencing and analysis were conducted by OE Biotechnology Co., Ltd We set Q < 0.05, foldchange of > 2, or foldchange of < 0.5 as the threshold for significantly differentially expressed genes (DEGs).Gene ontology and KEGG pathway enrichment analyzes were performed using R (v 3.2.0).

Secretion of cytokines
An enzyme-linked immunosorbent assay (ELISA) kit (Abclonal China) was used to determine the concentrations of tumor necrosis factor-α (TNF-α), interleukin-6 (IL-6), and interleukin-10 (IL-10) in the supernatant of RAW 264.7 cells.RAW 264.7 cells were seeded onto the specimens in 24-well plates at a density of 1 × 10 5 cells per well for 24 or 48 h.The supernatant on the surface of each group of specimens was collected and centrifuged at a rotating speed of 6000 rpm min −1 for 10 min.ELISA was performed according to the instructions, a standard solution was configured, and unknown samples were detected.The absorbance was measured at a wavelength of 450 nm with a multifunctional enzyme label instrument and a standard curve was plotted.Unknown sample protein concentrations were calculated according to the instructions.
2.8.Cellular immunofluorescence of macrophage polarization-related proteins RAW 264.7 cells were seeded onto the specimens in 24-well plates at a density of 5 × 10 4 cells per well.After 24 or 48 h, the samples were fixed with paraformaldehyde (4%) and subsequently blocked with goat serum (10%).ARG / iNOS/CD80/ CD163 monotonic solution of 200 μl was added to each well (diluted with PBS 1:100) followed by overnight storage at 4 °C.Subsequently, samples were washed three times with PBS, and 200 μl of the corresponding fluorescent secondary antioxidant solution was added to each well (diluted with PBS 1:400) for incubation at room temperature for 1 h in the dark.The pro-inflammatory biomarkers iNOS (Abclonal, China) and CD80 (Abclonal, China), and the anti-inflammatory biomarkers ARG (Abclonal, China) and CD163 (Abclonal, China) were labeled green and combined with 4', 6-diaminyl-2-phenylindoles (ZSBG-BIO, ZSDB-BIO, ABCLONE, China) to stain the nucleus.The fluorescence intensity was further analyzed quantitatively.The corrected total cellular fluorescence (CTCF) was calculated as the following: CTCF = integrated density − (area of select × mean fluorescence of background readings).

Statistical analysis
All data are expressed as mean ± standard deviation.Data were processed using SPSS software (version 26.0;IBM Corp., USA) through a one-way analysis of variance.The threshold for statistical significance was set at p < 0.05.

Characteristics
Pure Ti tablets (annotated as Ti with a size of 1 cm × 1 cm), TNT, and TNT-GO were employed as the blank control, control, and experimental groups, respectively.Visual inspection of the samples (figure 1(a)) revealed that Ti had a metallic color, whereas TNT was light yellow after anodization, and became yellowish-brown after GO loading to form TNT-GO.The SEM results (figure 1(b)) revealed that the surface of Ti was smooth with only a few scratches, whereas those of the TNT group revealed the presence of a uniform array of nanotubes with a diameter of approximately 100 nm.The deposition of the GO films on top of the uniformly arranged nanotubes was confirmed by the SEM images of the TNT-GO group.The Raman spectra (figure 1(c)) of TNT-GO revealed the presence of the two characteristic peaks of GO: the D peak attributed to the disorderly vibration of graphene, which appeared at approximately 1350 cm −1 , and the G peak, the main characteristic peak of graphene, which appeared at approximately 1580 cm −1 .These results further confirmed the successful loading of GO on the surface of the nanotubes.
The surfaces of the three groups were then subjected to further analyses (figures 2(a) and (c)).The roughnesses of Ti, TNT, and TNT-GO were approximately 9.7 ± 1.4, 48.4 ± 1.2, and 45.1 ± 2.6 nm, respectively (table 1).Furthermore, the water contact angle of pure Ti was approximately 88.2 ± 3.2°, while that of TNT was approximately 48.9 ± 2°.This increase in hydrophilicity can be attributed to anodization.Deposition of the GO on the TNTs increased the hydrophilic angle even further, and the water contact angle of TNT-GO decreased to approximately 33.1 ± 3.4°(figures 2(b) and (d)).The XRD pattern (figure 2(e)) of Ti contained only the diffraction peaks of pure Ti.The XRD pattern of TNT exhibited diffraction peaks related to the characteristic peaks of TiO 2 in the anatase phase in addition to the Ti diffraction peaks.The crystal type of TNT did not change even after the deposition of GO on the TNTs.

Cell morphology
SEM was used to observe the spreading of the RAW 264.7 cells seeded on the surface of each group for 4 and 24 h (figure 3).The cells of the Ti group had an elongated morphology, which is a characteristic of M1 macrophages, whereas those in the TNT and TNT-GO groups were round.

Viability assessment
Fluorescent staining was employed to observe the viability of the RAW 264.7 cells seeded on the surface of each group for 24 and 48 h (figure 4).Green and red fluorescent dyes were used for staining live and dead cells, respectively.The number of macrophages on the surface of each group increased with time.At each time point, the largest number of RAW 264.7 cells were observed on the surface of the Ti group, and this was accompanied by significant proliferation.Many cells were observed to have an elongated cell morphology.Compared with the Ti group, the degree of macrophage proliferation and number of dead cells was noticeably smaller for the TNT and TNT-GO groups.

Cell proliferation
CCK-8 was used to detect the proliferation of the RAW 264.7 cells seeded on the surface of each group for 4, 24, and 48 h (figure 5).The results revealed that, unlike the Ti group, the TNT-GO group inhibited the proliferation of the RAW 264.7 cells at 24 and 48 h.The proliferation of macrophages on the surfaces of the three groups after 48 h differed significantly.The highest rate at which RAW 264.7 cells proliferated was observed to occur on the surface of Ti, and the number of cells in the Ti group was clearly larger than that in the TNT group.Compared with the Ti surface, the proliferation rate of RAW 264.7 cells on the TNT-GO surface was lower at 24 h and 48 h.Data are expressed as mean ± SD. abc statistical differences in water contact angle and roughness of samples, and different letters represent statistically significant differences between groups.Abbreviations: Ti, titanium; TNT, titanium dioxide nanotube; TNT-GO, TNT-graphene oxide (GO).

RNA sequencing (RNA-Seq) analysis
The results of RNA-Seq analysis are shown in figure 6.Approximately 64.7 Gb of raw data was generated (143,252,948 sequences for Ti; 14,4714,288 sequences for TNT; and 143,307,070 sequences for TNT-GO).A comparison of the results of the TNT and Ti groups enabled 146 significantly DEGs to be identified.The number of genes significantly differentially expressed between TNT-GO and Ti and between TNT-GO and TNT were 125 and 63, respectively (figure 6(a)).
The functional changes behind these genomic data were explored by conducting gene ontology analysis (figures 6(b) and (c)).In the comparison of TNT with Ti and TNT-GO with Ti, the gene ontological terms are related to inflammatory processes, such as the inflammatory response, immune system processes, innate immune responses in the mucosa, oxygen binding, and chemokine signaling pathways.By manually scanning genetic data related to inflammatory response and immune system processes, it was found that genes related to inflammatory response and chemokine signaling pathways were significantly downregulated in the TNT and TNT-GO groups.

ELISA
The ELISA approach was used to detect the cytokine secretion of the RAW 264.7 seeded on the surface of each group for 24 and 48 h (figure 7).The secretion of M1-type polarization-related cytokine TNF-α by macrophages in the TNT-GO group was significantly decreased relative to that of Ti, and significant differences existed among the three groups.At the same time, the secretion of M1-type polarization-related cytokine IL-6 in the TNT-GO group was also significantly lower than that in the other two groups, whereas the secretion of M2-type polarization-related cytokine IL-10 in the TNT-GO group was significantly increased, with significant differences among the three groups at 48 h.

Expression of markers characteristic of M1/M2 phenotypes
The polarization-related protein expression of the RAW 264.7 cells on different samples was detected by immunofluorescence staining (figures 8 and 9).The fluorescence intensity was further analyzed quantitatively, and the results are shown in figure 10.At 24 h, the expression of M1-polarization-related proteins iNOS and CD80 was significantly higher in the Ti group than in the other two groups, and significant differences were observed among the three groups.For the TNT-GO group, the expression of M1 polarization-related proteins iNOS and CD80 was significantly lower, whereas the expression of M2 polarization-related proteins ARG and CD163 was significantly higher, and statistical differences were observed among the three groups.

Discussion
Ti implants are widely used because of their excellent mechanical properties, good corrosion resistance, and biocompatibility.However, these implants remain 'foreign objects' and can trigger immune inflammatory reactions to complicate the healing process.Owing to their phagocytic activity, macrophages are the first line of defense against infectious microorganisms and the earliest cells to come into contact with the surfaces of implants after implantation.Macrophages are activated upon migration to the site of injury and exhibit two functional polarization states: the M1 phenotype promotes the inflammatory response, and the M2 phenotype releases anti-inflammatory cytokines such as IL-10, suppresses the inflammatory response, accelerates tissue repair, and contributes to tissue remodeling [8-10, 16, 17].
In view thereof, an understanding of the exact role of macrophages in the tissue healing process, particularly the inflammatory response to the implantation of a biomaterial, is of the utmost importance.Regulating the polarization of macrophages to an anti-inflammatory prohealing phenotype (M2) can thus induce a favorable immune response after implantation.
Research has indicated that surface roughness, surface morphology, wettability, charge, chemical composition, and biological factors [18][19][20][21][22][23] can affect the polarization of macrophages and the activation and production of pro-inflammatory factors in these microorganisms [24,25].The inflammatory responses of implant materials can be regulated by controlling their surface morphologies [26,27].The shape and phenotype of macrophages are significantly affected by the nanomorphology, particularly in the case of a highly ordered morphology [28].These findings were confirmed by other studies, which showed that nanotubes with a diameter of 100 nm increased the number of protein adsorption sites, and that the gaps between the nanotubes can serve as nutrient supply pathways [29,30], which also suppress the inflammatory response [31].Therefore, the nanomorphology used in this study is a favorable factor for the M2 type polarization of macrophages.
In recent years, the surfaces of hydrophilic biomaterials have been found to promote macrophage apoptosis, increase the secretion of anti-inflammatory cytokines, and reduce the levels of pro-inflammatory factors [32,33].Wang et al coated the surface of Ti with GO and found the coating to improve the surface hydrophilicity and biocompatibility of Ti [34].Guo et al showed that the deposition of GO can improve the hydrophilicity of the PEEK surface because GO is richly endowed with oxygen-containing functional groups [35].The results of this study found that GO coating could further improve the hydrophilicity of TNT.An increase in the surface roughness, hydrophilicity, and protein adsorbability of biomaterials was expected to encourage macrophages to interact with adsorbable proteins and promote the polarization of M2-type macrophages.SEM imaging (figure 3) revealed that the macrophages in the Ti group had the shape of spindles, whereas the cells on the surfaces of the TNT-GO materials were circular.Previous studies led to the discovery that spindle-like cell morphology indicates activation and the ability to migrate, whereas a circular morphology is indicative of non- activation [36,37].In this study, quantification of the cytokine secretion and assessment of the polarization revealed reductions in the proliferation and activation of macrophages on the surface, as well as the inflammation levels, in the TNT and TNT-GO groups.
This phenomenon was investigated by conducting RNA-Seq analysis.The results revealed that the transcriptome responses of RAW 264.7 depended on the morphology of the substrate.Compared with the unmodified smooth Ti surfaces, the TNT and TNT-GO surfaces significantly impeded inflammatory responses, and downregulated the genes associated with chemokine signaling pathways (figure 6).These results demonstrated that changes in the inflammation levels are related to the surface morphology of the implant.
The effect of modification of the surface of the material on macrophage polarization was examined using ELISA and fluorescence staining to detect the phenotypes of macrophages.The TNT-GO surface suppressed the pro-inflammatory factors IL-6 and TNF-α (figure 7) and promoted the secretion of anti-inflammatory factor IL-10.The surface also promoted the expression of ARG and CD163 in macrophages (figures 8(b) and 9(b).Tu et al confirmed that the addition of GO to a hydrogel can positively contribute to immune regulation, promote the expression of anti-inflammatory factors such as ARG-1, IL-10, CD163, CD206 in macrophages, inhibit the expression of pro-inflammatory factors such as IL-1β, INOS2 and TNF-α, and induce the polarization of macrophages towards the M2 phenotype [38].Other studies have indicated that IL-10 plays important roles in tissue remodeling and the inhibition of inflammatory immune responses [24,39,40].The presence of these antiinflammatory cytokines and tissue remodeling responses can assist with the vascularization of regenerative biomaterials, improve their overall performance, and allow them to achieve their expected functions.Moreover, a higher M2:M1 ratio near the implantation of biomaterials can enhance the remodeling outcomes [41].Biomaterials that promote M2 polarization after appropriate inflammatory response can promote angiogenesis and wound healing.
The SEM, ELISA, and immunofluorescent staining results revealed that, compared with TNT, TNT-GO can further weaken the inflammatory response, strengthen the expression of anti-inflammatory factors and proteins, and promote the polarization of macrophages towards the M2 phenotype.These results suggest that the GO coating may also have certain anti-inflammatory properties in addition to its nanomorphology and hydrophilicity.Previous studies have also shown that the polarization of macrophages may additionally depend on micro-environmental factors [42].In a recent study, Han et al confirmed the regulatory effect of dispersed GO on macrophage polarization and showed that this material could be used for myocardial infarction repair [43].Their study revealed that GO can be used as an antioxidant to reduce the inflammation and inflammatory polarization of macrophages, down-regulate the polarization of M1 macrophages and the secretion of related inflammatory cytokines, and thus limit the inflammatory response.GO can also be used as a carrier of interleukin-4 plasmid DNA (IL-4 pDNA), and jointly induce the M2 polarization of macrophages.In this study, TNT-GO promoted the polarization of M2-type macrophages, which could further promote tissue healing and repair [44,45].Further studies would be necessary to explore and reveal the mechanism underlying the action of TNT-GO.
Based on these findings, it was thus concluded that macrophages are sensitive to the surfaces of implants and that the physical characteristics of the surface of biomaterials can affect the shape and phenotype of macrophages.The inflammatory response of implant materials can thus be regulated by their surface morphology [21,26,27,46].This suggests that suitable nanomodification of the implant surface and an increase in its hydrophilicity can regulate the polarization of macrophages to lower the inflammatory response and promote tissue healing.Compared with Ti and TNT, the nanomorphology of TNT-GO, combined with its enhanced hydrophilicity, was more conducive for reducing the inflammatory response of macrophages.The results of this study offer insights into designing novel 'immune guiding' materials that can regulate the reaction of the body to foreign materials.

Conclusion
TNT-GO material modified by anodic oxidation and electrodeposition has good roughness and hydrophilicity.This material can reduce the secretion of M1-type polarization-related factors IL-6 and TNF-α and the expression of related proteins iNOS and CD80 in macrophages to weaken the inflammatory response.TNT-GO also promoted the secretion of IL-10, the expression of ARG and CD163, and induced the M2 polarization of macrophages, thereby proving to be a promising material for dental implants.

Figure 3 .
Figure 3. SEM images of the mouse mononuclear macrophages (RAW 264.7) on the different surfaces.

Figure 4 .
Figure 4. Viability of the RAW 264.7 cells at 24 and 48 h.Green and red fluorescent dyes were used for live and dead cell staining, respectively.

Figure 6 .
Figure 6.RNA sequencing (RNA-Seq) analysis of the RAW 264.7 cells on the Ti, TNT, and TNT-GO samples: (a) Volcano maps showing the differentially expressed genes.Red, blue, and gray dots indicate up-regulated, down-regulated, and unchanged genes, respectively; (b) Gene ontology enrichment of the Ti, TNT, and TNT-GO samples; (c) Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment of the Ti, TNT, and TNT-GO samples.

Figure 8 .
Figure 8. Immunofluorescence staining of the RAW 264.7 cells at 24 h: (a) iNOS was labeled in green, and the nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI); (b) ARG was labeled in green, and the nuclei were stained with DAPI.The column 'Merge' shows merged images of iNOS or ARG and nuclei.

Figure 9 .
Figure 9. Immunofluorescence staining of the RAW 264.7 cells at 24 h: (a) CD80 and (b) CD163, both labeled in green, and the nuclei were stained with DAPI.

Table 1 .
Water contact angles and surface roughness of the samples.