The differences in color between the two substrates are shown in Figure 1A. It was observed that GO was uniform and well-coated in yellow on bare Ti due to its brownish color. FE-SEM observations were conducted to analyze the surface modifications of the substrates (Figure 1B). Bare Ti showed a rough grinding surface, while GO-Ti substrate showed a deposited GO surface. The chemical components of the surface were investigated using EDS to confirm the GO coating on the Ti (Figure 1C). As Ti has a surface composed of TiO and TiC, Ti, O, and C were identified in bare Ti (Jung et al., 2016; Marcin Behunova et al., 2021). Like bare Ti, GO-Ti revealed three peaks indicating Ti, O, and C. It was confirmed that the GO coating relatively increased the intensity of the C and O elements compared to the intensity of the Ti peak.

The components of the graphene-coated surfaces were accurately analyzed using Raman spectroscopy (Claramunt et al., 2015; Perumbilavil et al., 2015). GO generally displays a D-band and G-band in the Raman spectrum. In Figure 1D, the D and G bands were not apparent in bare Ti but were observed in GO-Ti. In GO-Ti, the D band associated with the vibration of sp³ carbon atoms was observed in the range 1,340–1,350 cm^-1, while a G band associated with the in-plane vibration of sp² carbon atoms was detected at 1,600 cm^⁻1 (Qin et al., 2014; Lopez-Diaz et al., 2017; Oh et al., 2021). The uniformity and quality of G” on the substrate surface were visualized by Raman mapping with an integrated intensity of the D and G bands (1,200–1700 cm^⁻1) (Figure 1D). Because the Raman spectra were not detected in the D or G bands, the integrated intensity was between 0 and 100, determined based on the highest integrated intensity of the D and G bands among the several GO flask coatings. Raman mapping is depicted as blue to red images scaled from 0 to 100. The overall blue color mapping indicates that the total intensity is zero due to the absence of GO on bare Ti. The mapping of the GO-coated surface is mainly indicated by red and green colors. The color difference indicates the difference in the integrated intensity within the range, which is related to the thickness uniformity of the coated GO. Although the quality of the coated GO, such as thickness uniformity, was slightly different on the entire surface, the existence of GO on the Ti surface was clearly confirmed. Contact angle experiments were conducted to investigate the improvement in the hydrophilicity of the surface after GO deposition (Zuo et al., 2013; Oh et al., 2021). In Figure 1F, the contact angle of Ti was 63.74° ⁻5.92°, and that of GO-Ti was 28.52° ± 1.74°. This indicates that when the GO was deposited, the surface of Ti became more hydrophilic. These findings suggest that the GO-coated layer has higher hydrophilicity and better biocompatibility to brain cells than bare Ti.

To investigate astrocyte cell area and adhesion on Poly-D lysine (PDL), Ti and GO-Ti, we stained glial fibrillary acidic protein (GFAP), an astrocyte marker, and F-actin with F-actin using Phalloidin. First, we verified that most primary culture cells were astrocytes (Supplementary Figure S1). Next, we confirmed that astrocytes were attached on PDL, Ti, and GO-Ti (Figure 2A). Also, astrocytes on GO-Ti had a wider cell area than PDL and Ti (Figure 2B). A prior study revealed that increased cell area enhanced cell adhesion and proliferation (Ruiz et al., 2011; Pan et al., 2016). Therefore, we confirmed that the mRNA expression levels of integrin (Itg) and glial cell adhesion molecule (G-CAM; Hepacam) (Figures 2C–F). There was no difference in Itgb1 mRNA expression levels after 2, 4, or 24 h seeding (Figure 2C). G-CAM mRNA expression was increased on GO-Ti than PDL after 24 h (Figure 2D). The Itgav mRNA expression was shown to be greater on GO-Ti than Ti after 2 h (Figure 2E). The Itgb3 mRNA expression was elevated not only on GO-Ti at 2 h, but also on PDL at 4 h and 24 h (Figure 2F). Therefore, we suggest that astrocytes have a faster and stronger cell adhesion on GO-Ti than PDL or Ti and that the increased astrocyte cell area could influence cell growth and proliferation.

Cell growth due to increased cellular area was associated with cell proliferation on GO-Ti (Ruiz et al., 2011). To determine whether cell proliferation was affected by GO-Ti, we stained astrocytes with cell proliferation marker, Bromodeoxyuridine (BrdU) (Figure 3A). The number of astrocytes stained with BrdU on GO-Ti was higher than PDL and Ti. Following that, we investigated cell viability using CCK-8 to determine whether GO-Ti could affect astrocyte proliferation (Figure 3B). On days 1 and 4, GO-Ti had better cell viability than PDL and Ti. Furthermore, we confirmed the mRNA expression level of GFAP, an astrocyte marker (Figure 3C). Although there was no difference at 4 h, we observed that the mRNA expression level of GFAP in astrocytes on GO-Ti increased at 24 h. Taken together, we suggest that GO-Ti can upregulate Gfap mRNA expression and improve cell proliferation in astrocytes.

In many pathological conditions in CNS such as Alzheimer’s and Parkinson’s disease, astrocytes could be induced to trigger primarily immune response and neurotoxicity (Choi et al., 2014; Phatnani and Maniatis, 2015; Colombo and Farina, 2016). Astrocytes produce proinflammatory cytokines to trigger an inflammatory response and eliminate pathogens or damaged cells. Astrocytes release anti-inflammatory cytokines to suppress chronic inflammatory responses and maintain immune system equilibrium (Jensen et al., 2013; Norden et al., 2016). Here, we incubated astrocytes with lipopolysaccharide (LPS) (1 μg/mL) for 4 h to induce an immune response. First, we determined whether the astrocyte cultured on GO-Ti could induce immune response by LPS treatment. GFAP mRNA expression was higher on GO-Ti than on PDL and Ti (Figure 4A). We also investigated the inflammatory response in astrocytes activated by LPS to GO-Ti changes (Figures 4B,C). When LPS was not treated, there was no variation in the mRNA expression levels of pro-inflammatory cytokines such as Il6, Tnf, and Il1b. In contrast, the mRNA expression levels of the three pro-inflammatory cytokines on GO-Ti increased following LPS treatment compared to Ti. However, there was no difference between PDL and GO-Ti (Figure 4B). These results showed that Ti decreased the basal immune response of astrocytes, whereas GO-Ti exhibited a basal immune reaction in astrocytes. The mRNA expression level of Il10, an anti-inflammatory cytokine, was enhanced in astrocytes on GO-Ti compared to PDL in both circumstances, with and without LPS. When LPS was treated to Ti and GO-Ti, the mRNA expression level of Il10 increased. Furthermore, we investigated the levels of protein expression of BDNF, a neurotrophic factor secreted by astrocytes, on GO-Ti (Figures 4D,E). When LPS was not treated, BDNF protein expression was increased on Ti and GO-Ti compared to PDL, but there was no difference between Ti and GO-Ti (Figure 4D). When treated with LPS, GO-Ti enhanced BDNF protein expression in astrocytes compared to PDL and Ti (Figure 4E). To establish if the three groups had distinct immunological responses due to improved LPS sensitivity, the mRNA expression level of Toll-like receptor (Tlr) 4, a receptor to which LPS binds, was studied (Figure 4F). There was no difference in the PDL, Ti, or GO-Ti after LPS treatment. Therefore, we suggest astrocytes on GO-Ti could have an efficient and balanced inflammatory response.

When astrocytes on PDL, Ti, and GO-Ti were incubated with LPS, the mRNA expression of Tlr4 exhibited no difference (Figure 4F). Therefore, we investigated other factors for this response. We confirmed mechanotransduction, a process by which mechanical stimulation of GO-Ti influences cell chemical activity as well as nuclear transcription factors. First, we studied representative mechanotransduction factors, yes-associated protein 1 (YAP), which is a major downstream effector of the Hippo signaling pathway. YAP regulates cell proliferation and remodeling of injured tissues as well as cell growth, survival, differentiation, morphological changes, and stem cell renewal (Jin et al., 2020). YAP can also be regulated by the extracellular matrix (Dupont et al., 2011). The nuclear transcription factor NF-κB is widely known to be a signaling pathway that induces an immune response. NF-κB induces immune responses by regulating the expression of cytokines, chemokines, and adhesion molecules. Previous studies have revealed that NF-κB induces both pro-inflammation and anti-inflammation (Lawrence, 2009). We confirmed the protein expression levels of YAP and NF-κB on PDL, Ti, and GO-Ti (Figures 5A–D). Both with and without LPS, the protein expression level of YAP was increased on Ti and GO-Ti compared to PDL. However, there was no difference in the protein expression levels of YAP between Ti and GO-Ti (Figures 5A,B). Likewise, the protein expression level of NF-κB was observed in the same pattern as YAP (Figures 5C, D). It is known that the nuclear transcription factors YAP and NF-κB located in the cytoplasm became active when translocated into the nucleus (Lawrence, 2009; Dupont et al., 2011; Jin et al., 2020). Here, we confirmed the ratio of YAP and NF-κB proteins located in the cytoplasm and nucleus (Figures 6A–D). When LPS was not treated, the ratio of YAP protein located in the nucleus increased on GO-Ti rather than PDL and Ti. When LPS was not treated, the ratio of YAP proteins located in the nucleus increased on GO-Ti rather than PDL and Ti (Figure 6A). However, after LPS treatment, the ratio of YAP proteins located in both the cytoplasm and nucleus showed no differences between PDL, Ti, and GO-Ti (Figure 6B). The ratio of NF-κB proteins located in both the cytoplasm and nucleus was not different on PDL, Ti, and GO-Ti when not treated with LPS (Figure 6C). However, after LPS treatment, the ratio of NF-κB proteins located in the nucleus was higher on PDL and GO-Ti than Ti (Figure 6D). Therefore, our findings imply that enhanced YAP activity in astrocytes stimulates adhesion, growth, and proliferation. Moreover, when astrocytes on GO-Ti induced an immune response, the increased activity of NF-κB protein can induce pro- and anti-inflammatory responses.

Grade 2 titanium (Titanart, Korea) was selected as the titanium plate. Prior to the experiment, it was polished with 800, 1,200, and 2000 grit silicon carbide paper (Daesung, Korea) to obtain an even surface of titanium. The residue was removed from acetone, ethanol (Duksan, Korea), and DW by ultrasonic washing. A single layer of graphene oxide (GO; Graphene Supermarket, USA) was used, and high-purity ethanol was used as the electrolyte. GO (500 ug/mL was prepared using 80% EtOH, and ultrasonic wave treatment was performed for 15 min to obtain a uniform electrolyte. The Ti substrate was placed in an electrolyte maintained at 1 cm from the anode. A Ti substrate was also used as the anode. For the EPD coating, an anode EPD process, in which electricity passes at a current of 10 mA for 1 min, was performed using a DC power source (TPM series, TOYOTECH, Korea) because GO has a negative charge. Subsequently, it was ultrasonically washed for 5 min in the order of acetone, ethanol, and water, and then dried. For the in vitro tests, the coated material was sterilized with a 70% ethanol solution.

Field-emission scanning electron microscopy (FE-SEM, Zeiss 300 Gemini, Germany) and energy-dispersive spectroscopy (EDS, UltraDry, Thermo USA) were used to confirm the coating surface and the elements of bare Ti and GO-Ti. After attaching bare Ti and GO-Ti to the stage, the surface was observed at 5,000 × magnification at 5 kV. The EDS element analysis was also fixed at 5,000 magnifications at 5 kV, and the components and intensities of Ti, C, and O were analyzed in bare Ti and GO-Ti by line scan. Raman spectroscopy was used to determine the presence of graphene. Excitation was performed at 532 nm to evaluate the bare Ti and GO-Ti, and the measurement range was set from 800 cm^-1 to 2000 cm^-1. In addition, the GO coating on the surface was imaged using Raman intensity through Raman mapping. For mapping, the laser power was set to 2 mW, the width was 50 μm × 50 μm, and the Raman shift was measured in the range of 1,200–1700 (cm^⁻1). The hydrophilicity of the bare Ti and GO-Ti surfaces was evaluated by measuring the contact angle. The hydrophilicity of the surface was compared by dropping water droplets onto the surface of each material to measure the inner angle of the water droplets.

C57BL/6 mice were used to culture primary astrocytes. All experiments in this study were conducted in accordance with the guidelines for animal accordance at Dankook University (Animal Experiment Approval No. Dankook University 19-016, Cheonan, Korea).

To culture primary astrocytes, the brains of C57BL/6 mice were isolated at postnatal days 0–2. To obtain the cerebral cortex for use in the experiment, the cerebrum was separated, and the medulla was removed. The cerebral cortex was then pulverized by pipetting and cultured in a culture dish coated with 0.1 mg/mL Poly D-lysine (PDL, #354210, Corning). The culture medium for obtaining astrocytes was high-glucose Dulbecco’s modified Eagle medium (#LM001, Welgene) with 10% fetal bovine serum (#S001-07, Welgene), 10% horse serum (#26050-088, Gibco), and 1% penicillin/streptomycin (#LS202-02, Welgene) 05, Welgene). Primary astrocytes were incubated at 37°C and 5% CO₂ for 3 days.

Primary astrocytes cultured in a culture dish were reacted with 1x trypsin-EDTA solution (Welgene) at 37°C for 3 min and suspended by pipetting. The suspended astrocytes were then placed in a 15 mL conical tube (SPL) and centrifuged at 2,000 rpm at 4°C for 10 min. The supernatant was removed, and 1 mL of astrocyte culture medium was added, followed by pipetting of single cells. Astrocytes were dispensed with a culture solution on 0.1 mg/mL poly D-lysine-coated cover glass, Ti, and GO-Ti. The astrocytes were then incubated at 37°C and 5% CO₂ for 24 h.

After removing the culture medium of astrocytes cultured on PDL-cover slips, Ti, and GO-Ti in a 24-well plate (SPL), they were washed twice with 1 × PBS for 5 min. Subsequently, astrocytes were fixed by treatment with 4% paraformaldehyde (PFA) for 30 min. The fixed astrocytes were washed three times for 5 min with 1x PBS, and PBS (blocking solution) containing 2% normal goat serum (#005-00-121, Jackson Immunoresearch) and 0.3% Triton X-100 were added at room temperature for 1 h. Blocking was performed using a shaker for 30 min. After blocking, the primary antibody was diluted in the blocking solution and incubated on a shaker at 4°C for 20 h. The primary antibodies used were as follows: rabbit monoclonal anti-yes-associated protein 1 (YAP, #14074, Cell Signaling), mouse monoclonal anti-NF- κ B p65 protein (#6956, Cell Signaling), mouse monoclonal anti-5′-Bromo-2-deoxyuridine (BrdU, #B2531, Sigma-Aldrich), chicken polyclonal anti-GFAP (#AB5541, Millipore Bioscience Research). F-actin was stained using Alexa Fluor™ 488-conjugated phalloidin (#A12379, Invitrogen™), and F-actin was stained using Alexa Fluor™ 594-conjugated phalloidin (#A12381, Invitrogen™). Then, the primary antibody solution was removed, and the cells were washed thrice for 5 min with 1x PBS. The secondary antibody was diluted in blocking solution, reacted at room temperature for 1 h, and washed thrice for 5 min with 1 × PBS. The secondary antibodies used were as follows: Alexa Fluor 594-conjugated goat polyclonal anti-rabbit (#111-585-003, Jackson ImmunoResearch Inc.), Alexa Fluor 488-conjugated goat polyclonal anti-chicken (#103-545-155, Jackson ImmunoResearch Inc.), and Alexa Fluor™ 488-conjugated goat polyclonal anti-mouse (#115-545-003, Jackson ImmunoResearch Inc.). DAPI (6-diamidino-2-phenylindole) was diluted in 1x PBS and incubated at room temperature for 5 min. After the reaction was completed, the cells were washed for 5 min with 1x PBS. Astrocytes on the PDL-cover slip, Ti, and GO-Ti were fixed using Dako Faramount Mounting Medium (#S3023, Dako) for microscopic observation. Images were acquired using a confocal microscope (Zeiss LSM700) and ImageJ software was used for image analysis. The fluorescence intensity of YAP and NFκB was analyzed as follows: YAP and NF κ B fluorescence intensity (a.u.) = cell intensity - (area × mean background fluorescence), and the nucleus/cytoplasm ratio was analyzed the nucleus positive protein intensity/total protein intensity ratio.

The Cell Counting Kit-8 (CCK-8, #ALX-850-039-KI01, ENZO) was used to measure cell viability and proliferation. The culture medium of astrocytes (2 × 10⁵cells/well) cultured in 0.1 mg/mL PDL-cover slip, Ti and GO-Ti in a 24-well plate (SPL) was removed. The cells were then washed twice with 1 × PBS. In each well, 450 μL of astrocyte culture medium and 50 μL of CCK-8 solution was added and reacted at 37°C and 5% CO₂ for 3 h. The absorbance was measured at 450 nm using a microplate reader (BIOTEK). The values calculated for cell viability and proliferation were as follows: O.D value of the sample–O.D value of the blank (culture medium + CCK-8 solution).

The extracted protein was adjusted to 15 μg for each sample and reacted with 5× sample buffer (60 mM Tris-HCl, 2% SDS, 25% glycerol, 5% β-mercaptoethanol, and 0.1% bromophenol blue) at 100°C. for 5 min. Electrophoresis was performed using 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and the resolved proteins were transferred to a polyvinylidene fluoride (PVDF) blotting membrane (#IPVH00010, Millipore Bioscience Research). To prevent non-specific protein binding, skim milk (#232100, BD DifcoTM) was added to 5% skim milk solution in Tris-buffered saline (TBS-T) containing 0.1% tween 20. Alternatively, BSA (#A2153-10G, SIGMA) was added to make a 2% bovine serum albumin solution in Tris-buffered saline (TBS-T) containing 0.1% tween 20 to make a blocking solution, and then reacted at room temperature for 1 h and 30 min. After the reaction, washing was performed thrice for 5 min with TBS-T, and the primary antibody was diluted with blocking solution and incubated at 4°C for 20 h. The primary antibody information and ratios used are as follows: 65 kDa, 1:5,000, rabbit monoclonal anti-yes associated protein 1 (YAP, #14074, Cell Signaling), 65–78 kDa, 1:5,000, mouse monoclonal anti-NF κ B p65 protein (#6956, Cell Signaling), 28–37 kDa, 1:10,000, BDNF (#ab108319, abcam) 43 kDa, 1:5,000, β -actin (#sc-47778, Santa Cruz Biotechnology). After washing thrice for 5 min with TBS-T, the secondary antibody was diluted with TBS-T and incubated at room temperature for 1 h. The secondary antibody information and ratios used were as follows: goat anti-rabbit IgG-H + I HRP-conjugated (#A120-101P, Bethyl antibodies) and goat anti-mouse IgG-H + I HRP-conjugated (#A90-116P, Bethyl antibodies). After the reaction, washing was performed thrice for 10 min each time with TBS-T. Finally, HRP was activated and visualized by chemiluminescence using an enhanced peroxidase detection kit (#EBP-1071; ELPIS-Biotech) for 1 min. Data were analyzed using ImageJ software and GraphPad Prism 9.