Robust optical metasurface heterostructures for prevention and treatment of peri-implantitis

Abstract

Peri-implantitis poses significant challenges in the field of oral implantology, while the inherent antibacterial properties of implant abutments and their tight sealing with surrounding soft tissues play a vital role in the prevent and treatment of peri-implantitis. Herein, a robust optical metasurface CeO_x/TiO₂ heterostructure with gradient-distributed nanocavity structure was prepared on titanium surface by innovatively integrating plasma immersion ion implantation and thermal treatment technologies. The CeO_x/TiO₂ heterostructure with an optimized bandgap enhances light absorption in the second near-infrared region and a robust optical metasurface with gradient-distributed nanocavities improves the retention of NIR light by the light trap effect and increases the heat preservation capacity, leading to superior photothermal antibacterial activity in vitro and in vivo. Besides, with the existence of ceria, metasurface CeO_x/TiO₂ heterostructure shows antioxidant properties and can alleviate the local inflammatory microenvironment, promoting soft tissue sealing in vitro and in vivo by sequentially activating the “Neuroactive ligand-receptor interaction, Rap1 signaling pathway, and ECM-receptor interaction” cascade network. This study paves the way for developing novel dental implant abutments with antibacterial activities and soft tissue sealing properties based on optical metasurface heterostructures for prevention and treatment of peri-implantitis.

Graphical abstract

1. Introduction

With the widespread applications of implant techniques, postoperative complications such as peri-implantitis have also been increasingly emerged [1], manifesting as loss of attachment of the gingival tissues surrounding the implant, alveolar bone resorption, and ultimately leading to implant loosening and failure of implant restoration [2]. Currently, clinical treatments for peri-implantitis include non-surgical and surgical therapies. Non-surgical treatment can only temporarily halt the progression of inflammation and partially restore the function of the implant in the short term, but it is ineffective in preventing secondary infection of the implant. Moreover, the overuse of antibiotics in non-surgical treatment can disrupt the body's beneficial flora and may lead to the development of drug resistance, contributing to the global spread of multidrug-resistant bacteria [3]. Surgical treatment has significant limitations in aesthetic areas, is economically costly, and offers uncertain postoperative outcomes [2]. Therefore, there is an urgent need for developing a novel strategy for the prevention and treatment of peri-implantitis.

Titanium, owing to its excellent biocompatibility, is widely used in oral implantations [4], [5]. The titanium implant abutment, as the transmucosal part of the implant, forms the first line of defense against the invasion of pathogenic microorganisms by sealing with the surrounding soft tissues [6]. However, owing to the bio-inertness of titanium, its ability to bond with soft tissue is relatively poor. Bacteria, such as porphyromonas gingivalis and fusobacterium nucleatum, in the oral cavity can adhere to the implant abutment and invade the soft tissue through micro-gaps between the abutment and the surrounding connective tissues, causing progressive infection and ultimately leading to implant failure [7]. Although the success rate of osseointegration for dental implants currently reaches as high as 95%, the reported incidence of peri-implantitis ranges from 15% to 40% [8]. Therefore, surface modification of titanium-based dental implant abutments in the transmucosal area to regulate the interaction between cells/bacteria and materials is crucial for preventing the occurrence of peri-implantitis. The surface modification aims to promote the sealing of soft tissues around the abutment while inhibiting bacterial colonization. However, the antibacterial surfaces, imparting antibacterial functionality, also compromise the biocompatibility, which can hinder tissue integration of the implant and affect the implantation outcome [3], [9], [10]. In recent years, smart-responsive antibacterial coatings have attracted widespread attention [11]. They show antibacterial activity only under specific conditions and environments, thereby avoiding damage to normal tissues or inducing cellular drug resistance [10]. Therefore, the development of an intelligent-responsive titanium implant abutment that can not only facilitate early soft tissue sealing to prevent the occurrence of peri-implantitis, but also actively kill bacteria during the period of bacterial infection, and further eliminate ROS to inhibit inflammation and promote repair, is crucial for ensuring the long-term success of the implant.

TiO₂ is a material with photothermal activity, which can generate a thermal effect when exposed to light, thereby damaging the structure of bacteria. However, TiO₂ exhibits low photothermal conversion efficiency owing to its wide bandgap structure, and its photogenerated antibacterial capability necessitates activation by ultraviolet light, which limits its extensive clinical applications [12]. The two fundamental factors for improving the photothermal properties of TiO₂ include reduction of the bandgap and enhancement of electron interfacial transport [13], [14]. Heterostructures are effective means of modulating electron transfer [15], [16], [17]. Cerium oxide (CeO₂), as a rare earth oxide, possesses highly active redox pairs of Ce⁴⁺/Ce³⁺, which can facilitate electron transfer in photothermal conversion [18]. When CeO₂ forms a heterojunction with TiO₂, it may create heterostructure with a narrow band gap, altering the light absorption range to enable surface absorption of NIR light, thereby enhancing the efficiency of photothermal conversion, and achieving precise targeted antibacterial effects in a short period of time through the photothermal effect, while avoiding tissue damage, secondary surgical trauma, and the development of drug resistance. In addition, CeO₂ serves as an ideal antioxidant, capable of eliminating ROS during redox cycling, thereby inhibiting inflammation, protecting damaged tissues, and promoting tissue regeneration [19]. Furthermore, studies have reported that Ce ions can also upregulate the expression of soft tissue sealing-related genes and proteins in fibroblasts, thus facilitating soft tissue sealing [20]. Despite significant progress has been achieved on the TiO₂-CeO₂ composite system in terms of photocatalysis, these discoveries have primarily focused on applications at the nanoparticle level [21], [22]. Their potential as surface coatings for titanium-based implants has yet to be fully explored.

Except for heterostructure, metasurface can also enhance the optical performance by modulating the phase, amplitude and reflection of electromagnetic waves based on subwavelength-scale nanostructures [23], [24]. However, in general, metasurfaces are fabricated by using direct-writing lithography systems including electron-beam lithography and focus-ion-beam lithography, which show low patterning resolution and material limitations. Although metasurfaces with periodic arrays show remarkable achievement, the periodicity presents strong spatial dispersion and optical anisotropy. Disordered metasurfaces, containing positional or dimensional randomness of the constituent structures, provide much higher degrees of freedom, enabling light control in subwavelength scales and multifunctional optics [25], [26], [27].

Therefore, in this study, a disordered metasurface CeO_x/TiO₂ heterostructure was prepared on a titanium abutment with a gradient nanocavity structure using plasma immersion ion implantation and thermal treatment (Scheme 1). Through experimental characterization and density functional theory (DFT) calculations, mechanisms underlying Ce-doping-induced bandgap narrowing, excitation light red-shift, and photothermal performance were investigated. Cell and animal experiments further revealed multiple biological effects of the metasurface CeO_x/TiO₂ heterostructure, including soft tissue barrier function, photothermal antibacterial activity, and CeO_x-mediated antioxidant activity through ROS scavenging. The molecular mechanism involving cascading signal pathway regulation for cell behavior modulation and tissue remodeling was also elucidated. This strategy, targeting the disease progression of peri-implantitis, specifically activates three functional modules: prevention, treatment, and tissue regeneration, thereby significantly improving long-term implant stability and clinical success rates. The findings provide a theoretical foundation for developing next-generation multifunctional smart-responsive dental implant abutments based on optical metasurface heterostructures.

Scheme 1.

Schematic diagram illustrates the mechanism of optical metasurface CeO_x/TiO₂ heterostructure for prevention and treatment of peri-implantitis. CeO_x/TiO₂ heterostructure with optimized bandgap enhances the light absorption in the second NIR region and robust optical metasurface with gradient-distributed nanocavities improves the retention of near-infrared (NIR) light by light trap effect and increases the heat preservation capacity, leading to superior photothermal antibacterial activity. Besides, with the existence of ceria, CeO_x/TiO₂ heterostructure shows antioxidant properties and can alleviate the local inflammatory microenvironment, promoting soft tissue sealing.

2. Results and discussion

2.1. Sample characterization

Titanium-based implants with bioactive coatings still face a series of challenges in clinical applications, such as good bonding stability between the coatings and the substrates, as well as the uniformity of the coatings [28]. The coatings created using plasma immersion ion implantation technique adhere strongly and uniformly to the substrates, which is ascribed to the high-energy ions collision and reaction with the substrates, forming new chemical bonds and alloy layers, ensuring coating stability and uniformity [29]. Based on this, in this work, a CeO_x/TiO₂ heterostructure coating was prepared by combining plasma immersion ion implantation and high-temperature heat treatment techniques (Fig. 1a). Briefly, a thin and uniform Ce-doped layer was formed on the surface of Ti substrate via plasma immersion ion implantation. Subsequent high-temperature heat treatment led to the surface oxidation, forming the CeO_x/TiO₂ heterostructure. As shown in Fig. 1b, Ti with acid etching exhibits a slightly undulating ridge-like surface morphology. A similar surface structure can be found from Ce-PIII. Under heat treatment at 550 °C, TiO₂ presents nanoparticle morphology. Compared with TiO₂, nanoparticle morphologies with smaller size appear on CeO_x/TiO₂. Fig. 1c, d presents the XRD results of Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂. Compared with Ti, no new characteristic peaks be observed from Ce-PIII. After heat treatment at 550 °C, both TiO₂ and CeO_x/TiO₂ exhibit new characteristic peaks at 27.6° corresponding to the (110) plane of rutile phase of titanium oxide. This indicates that titanium oxide was formed during heat treatment. Fig. 1e-h show the XPS results of Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂. The characteristic peaks of Ti, O, and C elements can be observed from all the samples. Quantitative atomic percentages derived from XPS provide more precise information on surface chemical composition. Based on the survey spectra, we calculated the atomic percentages of each element on the surfaces of Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂ samples (Table S3). Ce elements were detected in Ce-PIII and CeO_x/TiO₂. From the XPS high-resolution spectra of Ce 3d, Ce 3d_3/2 with the binding energy of 916.6, 906.4, and 901.1 eV, and Ce 3d_5/2 with the binding energy of 898.4, 888.6, and 882.4 eV correspond to the Ce⁴⁺ in cerium oxide, while Ce 3d_3/2 of 904.0 eV and Ce 3d_5/2 of 885.6 and 880.8 eV correspond to the Ce³⁺ in cerium oxide. The relative percent contents of Ce⁴⁺ from Ce-PIII and CeO_x/TiO₂ are 49.77% and 72.91%, respectively. CeO_x/TiO₂ shows a higher content of Ce⁴⁺ than that of Ce-PIII, which is attributed to the oxidation during the high-temperature heat treatment process.

Fig. 1.

Surface characterization of various samples. Schematic illustration of CeO_x/TiO₂ preparation (a). SEM images (b), XRD patterns (c, d), and XPS full spectra (e) of Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂. (f) XPS spectra of Ce 3d from Ce-PIII and CeO_x/TiO₂ with corresponding binding energies of Ce⁴⁺ (solid lines) and Ce³⁺ (dashed lines). XPS high-resolution spectra of Ce 3d from Ce-PIII (g) and CeO_x/TiO₂ (h). Water contact angles (i), surface hardness curves (j), average hardness (k), and Tafel curves (l) of Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂. Photoluminescence spectra of PTA solutions treated with Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂ without (m) or with (n) 500 μM H₂O₂ for three days, original PTA solution without addition of H₂O₂ was used as control group. *p < 0.05, **p < 0.01, ***p < 0.001.

Fig. 1i presents the average water contact angles of various samples. The average contact angles of Ti and Ce-PIII are 83.8° and 61.5°, respectively, which suggests that Ce ion implantation effectively enhances the hydrophilicity of titanium substrate. TiO₂ and CeO_x/TiO₂ underwent heat treatment exhibit average contact angles of 50.3° and 58.8°, respectively. This indicates that oxide layer formed on the sample surface after heat treatment contribute to the hydrophilicity of the titanium substrate, thereby enhancing its biocompatibility and application potential [30]. Besides, Ce ion implantation can also increase the surface hardness of titanium substrate (Fig. 1j and k). Moreover, the heat treatment process forms dense oxide layers on the surfaces of TiO₂ and CeO_x/TiO₂ which have excellent physical and chemical stability and can effectively resist external mechanical wear and chemical erosion, further enhancing the surface hardness [31]. Corrosion resistance is one of the key factors affecting the service life of titanium implant abutments. Comparing the self-corrosion currents (Fig. 1l) and various corrosion parameters (Table S4) of the different samples, Ce-PIII shows reduced self-corrosion current, indicating that Ce ion implantation effectively improves the corrosion resistance of titanium. Compared with Ce-PIII, CeO_x/TiO₂ has a lower corrosion current. This may be ascribed to the protective oxide layer formed on the titanium surface, which blocks direct contact between corrosive media and titanium matrix, thereby slowing down the corrosion process [32]. In addition, the formation of CeO_x facilitates the reversible conversion between Ce³⁺ and Ce⁴⁺. The conversion mechanism may help slow down the corrosion tendency of Ti substrate, thereby the CeO_x/TiO₂ surface exhibiting stronger corrosion resistance [33]. In addition, the release of Ti (Table S5) and Ce ions (Table S6) from CeO_x/TiO₂ under different conditions was further characterized. The results showed that elemental release was strongly affected by pH. Under neutral conditions (pH = 7), Ti remained below the detection limit throughout the 14-day immersion period, while Ce release remained relatively low. A pH of 2.5 was used to represent strongly acidic beverages such as cola. Both Ti and Ce release increased markedly under this condition. NIR irradiation caused only a limited additional increase. After 14 days, Ti concentrations reached 2.44 ± 0.22 ppb (NIR off) and 2.63 ± 0.22 ppb (NIR on) under acidic conditions, respectively, while the corresponding Ce concentrations reached 39.4 ± 0.06 ppb (NIR off) and 45.1 ± 0.14 ppb (NIR on). The greater Ce release may be related to the higher acid sensitivity of surface CeO_x species. In comparison, Ti is protected by a relatively stable and compact TiO₂ passive layer, which limits its dissolution.

Peri-implantitis is an inflammatory disease with complex pathogenesis [34]. In recent years, many studies have shown that reactive oxygen species (ROS) play a key role in the pathogenesis of peri-implantitis. The overproduction of ROS can not only directly damage the soft tissue cells around the implant but also interfere with the body's immune response, thereby exacerbating the inflammatory response [35]. To evaluate the potential of CeO_x/TiO₂ heterostructure to scavenge ROS, PTA assay was used to observe the fluorescence intensity changes of different samples under hydrogen peroxide (H₂O₂) treatment. As shown in Fig. 1(m and n). In the absence of H₂O₂, the various samples have little effect on the fluorescence intensity of PTA, indicating that the materials themselves do not produce significant interference effects. However, when 500 μM H₂O₂ was introduced, the fluorescence intensities of Ti and TiO₂ increase significantly. In contrast, Ce-PIII and CeO_x/TiO₂ show a similar level to that of pure PTA solution. This indicates that in the presence of H₂O₂, the Ce-containing coating can effectively scavenge ROS, which is crucial for reducing oxidative stress damage to the soft tissues around the implants and regulating the body's immune response.

The HAADF-STEM image of the cross-section of CeO_x/TiO₂ reveals a distinct layered structure in its surface region (Fig. S1a). The upper layer is an oxide-rich zone with a high oxygen content, while the lower layer is a Ce element diffusion zone that closely resembles the original Ti substrate (Fig. S1b). The formation of this layered structure may stem from the combined effects of atomic diffusion and migration during the ion implantation and heat treatment processes. During ion implantation, the collision cascades displace atoms and generate defects near the surface, while the implanted ions typically exhibit a Gaussian-like depth distribution. [36], [37]. During the heat treatment process, O₂ from the atmospheric environment diffuses and permeates into the interior of the Ti substrate. Due to the hindrance posed by the metal lattice, as well as the preferential oxidation reactions occurring in the regions near the surface, an oxygen-rich zone forms in the surface layer of the Ti substrate, while the interior becomes a relatively oxygen-deficient zone. The cross-sectional mapping results reveal that the distribution depth of the Ce element exceeds 3500 nm, which is significantly greater than the typically reported ion implantation depths achieved by PIII processes (Fig. S2) [38], [39], which may be attributed to the synergistic effect resulting from the combined action of lattice defects introduced by ion implantation providing ion migration pathways and vacancy sources and subsequent heat treatment at specific temperatures supplying bulk diffusion driving forces and thermal activation. The modification depth of conventional ion implantation is typically limited to a few hundred nanometers. However, studies have demonstrated that high temperatures and high current densities can significantly enhance diffusion depth. For instance, Vlcak et al. found that nitrogen implantation depth could reach 1 μm at 500 °C [40]. Similarly, Ryabchikov et al. achieved a dramatic increase in aluminum implantation depth to 53 μm at 1170 K using a high current density of 200 mA/cm² [36]. In this study, by combining 550 °C heat treatment with an implantation current of 9.4∼11.0 mA/cm², we successfully extended the Ce implantation depth to the micrometer scale. Certainly, as for the deep diffusion mechanism of Ce beyond 3500 nm, especially the specific details regarding the synergistic effect of lattice defects and heat treatment, further verification through more sophisticated characterization and simulation techniques remains to be conducted in the future. Notably, at the interface of the CeO_x/TiO₂ heterostructure, a large number of nanocavity structures (with diameters approximately ranging from 10 to 30 nm) are observed and exhibit a gradient distribution along the direction normal to the interface that nanocavities with high density and small size are near the interface between the oxide layer and the Ti substrate, and then gradually increase in size and become sparser as approaches the surface (Fig. 2a–g and Fig. S3). Element mapping analysis confirms the existence of Ti, O, and Ce, whereas these three elements within the nanocavity region all significantly attenuate (Fig. 2b). This phenomenon may be attributed to the stress accumulation caused by the combined effects of defect supersaturation induced by high-dose ion implantation and lattice mismatch at the interface, which subsequently induces defect aggregation, collapse of defect clusters, and formation of nanocavities [41]. The high-temperature heat treatment process provides the driving force for migration akin to the Kirkendall effect (due to the different diffusion rates of Ti and O). Vacancies continuously aggregate at the interface, much like bubbles, and move toward the surface to relieve stress [42], [43], [44].

Fig. 2.

TEM characterization of CeO_x/TiO₂ heterostructure. (a) FIB image of CeO_x/TiO₂, and (b) the corresponding element mapping images of Ce, O, and Ti of the red frame region in (a). (c) HAADF image of CeO_x/TiO₂ with nanocavity structures. (d) FFT image of the red frame region in (c). (e) iFFT image of the red frame region in (c). (f) Local enlarged image of the red frame region in (e). (g) Schematic illustration of the formation process of metasurface CeO_x/TiO₂ heterostructure with gradient-distributed nanocavities. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

HRTEM and Fourier analysis further unveil the significant impact of Ce doping on the lattice structure of titanium oxide. The fast Fourier transform (FFT) patterns exhibit distinct and symmetrical diffraction spots, which suggests that the oxide layer retains high crystallinity. However, the emergence of weak reflection spots may be attributed to defects or local symmetric structural changes induced by Ce doping and stress (Fig. 2c and d) [45], [46]. The FFT pattern can be indexed to the {020} and {011} families of planes, which is consistent with the tetragonal rutile structure of TiO₂ (P4₂/mnm). The overall diffraction geometry corresponds to a zone axis close to [100]. Based on the measured reciprocal spacings (∼3.90–4.18 nm⁻¹), the calculated interplanar distances are approximately d ≈ 0.24–0.26 nm, which agree well with the standard d-spacings of the (020) and (011) planes in rutile TiO₂. The obtained inverse fast Fourier transform (iFFT) images highlight a highly ordered atomic arrangement and periodic lattice fringes. However, contrast variations and fringe shifts in certain regions suggest the presence of local strain fields or defect clusters (Fig. 2e and f) [47]. The measured average interplanar spacing is 0.332 nm, which is higher than the theoretical value of ∼0.325 nm for the (110) plane of rutile titanium oxide (Fig. S4) [48]. This systematic lattice expansion can be attributed to local lattice distortions caused by the substitution of Ti⁴⁺ (0.061 nm) with Ce³⁺/Ce⁴⁺ (with ionic radii of 0.107 nm and 0.097 nm, respectively), accompanied by the formation of oxygen vacancies due to charge imbalance [49].

The high-resolution O 1s XPS spectra of Ce-PIII and CeO_x/TiO₂ predominantly display two characteristic peaks (Fig. S5). The primary peak at 529.7 eV corresponds to lattice oxygen, mainly originating from O²⁻ within the titanium oxide or cerium oxide. The secondary peak near 531.4 eV is commonly attributed to adsorbed oxygen, hydroxyl groups (–OH), and weakly bound oxygen associated with oxygen vacancies [50]. Furthermore, Ce implantation process induces significant lattice distortion and generation of oxygen vacancies. The surface defects also facilitate the adsorption of water molecules from the environment, leading to the formation of Ti-OH groups. Following thermal treatment at 550 °C, partial repair of surface defects occurred in CeO_x/TiO₂, restoring lattice ordering. This structural recovery narrowed the primary O 1s peak at 529.7 eV. The high-temperature environment further promotes the desorption of surface hydroxyl groups but cannot completely eliminate the oxygen vacancies introduced by implantation, resulting in decreased intensity of the 531.4 eV peak.

While compensating for charges, oxygen vacancies also serve as vacancy traps that enhance the stability of the cavities [51]. Notably, the finally formed metasurface with nanocavities exhibits significant advantages in optical functionality. The characteristic size of this nanocavity network ranges from 10 to 30 nm, which is substantially smaller than the wavelengths of NIR light (780–1100 nm). This disordered array of nanocavities, embedded within a high-refractive-index matrix (TiO₂, n ≈ 2.6) with low-refractive-index voids (n ≈ 1), induces dense local refractive index variations. These variations cause photons to undergo multiple scattering and retention within the coating, significantly extending the light path length and increasing the absorption probability, thereby forming a “light-trapping”. Svensson et al. [52], [53] reported that void structures in TiO₂ (with a diameter of 79 nm and a void density of 42.4%) could enhance the optical path length of incident NIR light by up to 135 times. Additionally, the extremely low thermal conductivity of the nanocavity structures (approximating that of a vacuum, 0 W m⁻¹ K⁻¹) facilitates local heat accumulation. Besides, CeO_x/TiO₂ demonstrates a notable heat preservation effect (Fig. S6), which indicates that the robust metasurface with nanocavity structures in CeO_x/TiO₂ further enhance the overall photothermal conversion efficiency by acting as “thermal-trapping” [54].

2.2. Photothermal performance

TiO₂, with its wide bandgap structure, typically requires ultraviolet light to activate its photo-induced antibacterial capabilities [55]. However, ultraviolet light faces certain limitations in biological applications, such as low penetration depth and potential safety concerns [56], [57]. On the other hand, NIR light offers higher tissue penetration and better biological safety [58]. Therefore, exploring new ways to modify TiO₂ to enable its excitation by NIR light has become a focal point of research. In this study, we employed the construction of a CeO_x/TiO₂ heterostructure coating as a modification approach, and then measured the light absorption of various modified samples and evaluated their photothermal performance under NIR illumination. The results reveal that the introduction of CeO_x/TiO₂ heterostructure coating significantly enhances the light absorption in the second NIR region (Fig. 3a). This enhancement confirms that the heterostructure design effectively broadens the spectral response range of TiO₂, significantly boosting its ability to absorb NIR light.

Fig. 3.

Photothermal performance of various samples. (a) UV-Vis-NIR absorbance spectra of Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂. (b) Heating curves of Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂ under NIR irradiation at the density of 0.50 W/cm². (c) Heating curve of CeO_x/TiO₂ under NIR irradiation at the density of 0.50 W/cm² with three on-off cycles. (d) Heating curves of CeO_x/TiO₂ under NIR irradiation at different power densities. Heating curves (e₁) and corresponding thermal images (e₂) of the different samples immersed in ultrapure water under NIR irradiation. Heating curves (f₁) and corresponding thermal images (f₂) of different samples in mouse subcutaneous tissues under NIR irradiation.

The 1064 nm wavelength is a standard output of Nd:YAG lasers used in clinical dentistry, providing an established basis for intraoral material applications. Except for significantly enhancing light absorption performance, photothermal performance of CeO_x/TiO₂ heterostructure coating under NIR light irradiation was investigated. The results show differentiated photothermal responses among different samples under the same conditions of 1064 nm NIR light irradiation. Specifically, surface temperatures of Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂ reach 64.1 °C, 77.1 °C, 96.2 °C, and 120.9 °C, respectively (Fig. 3b), which indicates that both the defects introduced by Ce ion implantation and the TiO₂ oxide layer formed by heat treatment contribute to improving the photothermal conversion efficiency of Ti. The construction of CeO_x/TiO₂ heterostructure exhibits a more significant photothermal effect. It is worth noting that CeO_x/TiO₂ group not only demonstrates excellent photothermal conversion efficiency but also shows good photothermal stability. In continuous three on-off cycle tests, its surface temperature can be stably maintained at ∼120 °C (Fig. 3c), providing a reliable performance guarantee for its practical application. To more comprehensively evaluate the photothermal performance of CeO_x/TiO₂, the influence of NIR light with different power densities on surface temperature was conducted. Results show that surface temperatures of CeO_x/TiO₂ reach 68.9 °C, 120.6 °C, and 155.7 °C under power densities of 0.25, 0.50, and 0.75 W/cm², respectively (Fig. 3d), which further confirms that CeO_x/TiO₂ has excellent NIR photothermal response characteristics, and its performance can be significantly improved with increasing light intensity.

To better simulate the practical clinical application scenarios, photothermal performance of CeO_x/TiO₂ in ultrapure water was further investigated using 1064 nm NIR light at various power densities. The results showed that at a power density of 0.50 W/cm², surface temperatures of Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂ increase to 42.1 °C, 45.0 °C, 49.8 °C, and 56.2 °C, respectively, within 10 min (Fig. 3e₁, e₂). This trend is consistent with the previous results obtained in air, confirming the positive effect of CeO_x/TiO₂ heterostructure on enhancing the photothermal performance. Additionally, subcutaneous implantation experiments in mice were conducted to evaluate the photothermal performance within a biological context. The results indicate that at a power density of 0.75 W/cm² (Fig. 3f₁, f₂), surface temperatures of Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂ rapidly increase to 38.4 °C, 41.2 °C, 47.0 °C, and 49.9 °C, respectively. Among them, CeO_x/TiO₂ maintains the best photothermal performance, reaching a higher temperature level within the same time, which further emphasizes the superiority of CeO_x/TiO₂ in terms of efficient photothermal conversion.

2.3. Band structure and charge transfer mechanism

The Ce implantation and thermal treatment processes jointly determine the surface composition and structure. The surface modification processes not only reshape the crystal structure and defect state distribution but also provide a physical basis for altering carrier transport pathways and the energy band structure. In particular, the formation of multivalent cerium oxide and oxygen vacancies not only participates in constructing heterostructure but also plays a crucial role in carrier transport processes. Therefore, further simulation calculations were conducted to systematically elucidate the profound impacts of heterostructure on photoelectron properties from the perspectives of energy band structure, interfacial charge behavior, and defect state distribution. To explore the electronic effects of contact between Ce-containing oxide and TiO₂, an idealized crystalline CeO₂(101)/TiO₂(110) interface was constructed. The rutile phase is supported by XRD, while crystalline CeO₂ provides a structurally defined reference for examining interfacial charge redistribution. The bandgap widths of TiO₂ and CeO_x/TiO₂ were determined using the Tauc plot method, yielding values of 3.16 eV and 2.84 eV, respectively (Fig. 4a) [59]. Additionally, conduction band (CB) and valence band (VB) positions of TiO₂ and CeO_x/TiO₂ were determined using ultraviolet photoelectron spectroscopy (UPS). The results were calibrated based on the Fermi edge measured with a gold reference sample, using a photon energy of 21.22 eV. The UPS measurements reveal cutoff edge values of 16.19 eV and 16.51 eV, and Fermi edge values of 2.32 eV and 2.41 eV for TiO₂ and CeO_x/TiO₂, respectively (Fig. 4b and c). The calculated work functions are 5.03 eV and 4.71 eV, and valence band positions are 2.91 eV and 2.68 eV for TiO₂ and CeO_x/TiO₂, respectively. The conduction band positions were subsequently derived by combining these results with the measured bandgap widths [60]. Based on first-principles calculations using the Vienna Ab initio Simulation Package (VASP) (Fig. 4(d₁-d₃), (e₁-e₃)), the bandgap of TiO₂ is determined to be 3.4 eV, whereas after forming heterostructure with cerium oxide, the bandgap decreases to 3.06 eV, matching the experimentally measured bandgap width. This change is primarily attributed to the introduction of 4f orbital states from cerium oxide. Based on this, a CeO_x/TiO₂ heterojunction model was constructed. The differential charge density results reveal that the formation of an internal electric field at the CeO_x/TiO₂ interface accelerates the directional flow of electrons (Fig. 4f). An examination along the z-axis of the model demonstrates that the CeO_x side exhibits electron efflux, indicating its electron-donating nature. This facilitates efficient interfacial charge separation. During the thermal treatment process, the increase of temperature may lead to the diffusion and rearrangement of defects introduced by ion implantation, resulting in an increase in system entropy and a decrease in free energy, which further promotes the electron transfer process between TiO₂ and CeO_x, thereby causing the increase of Ce chemical valence state in CeO_x/TiO₂ compared to that in the Ce-PIII sample. The increase in the proportion of tetravalent cerium in CeO_x/TiO₂ enhances the ability to consume H₂O₂ and O₂•^- [61]. Notably, the DFT results offer a possible interpretation of the experimental observations. Clarifying the local structural form of CeO_x would allow more accurate theoretical modeling.

Fig. 4.

Band structure and charge transfer mechanism in the CeO_x/TiO₂ heterostructure. (a) Tauc plot of TiO₂ and CeO_x/TiO₂. UPS spectra of TiO₂ (b) and CeO_x/TiO₂ (c). PDOS diagrams of TiO₂ (d₁), CeO₂/TiO₂ (d₂) and CeO_x/TiO₂ (d₃). Band structure of TiO₂ (e₁), CeO₂/TiO₂ (e₂) and CeO_x/TiO₂ (e₃). (f) The differential charge density of CeO_x/TiO₂. (g) Schematic illustration of the charge transfer mechanism in CeO_x/TiO₂.

As discussed above, under NIR light irradiation, the light-trapping effect of nanocavities in metasurface CeO_x/TiO₂ heterostructure enhances the retention of NIR light. The defect chemistry induced by Ce implantation strengthens the absorption of NIR light. The electrons in CeO_x/TiO₂ heterostructure can be excited from the valence band and transferred to the conduction band via defect-mediated intermediate band (Fig. 4g). The heterostructure causes band bending due to Fermi level alignment, which accelerates the interfacial electron transfer. When the electrons in the excited state fall back to the ground state, they release heat, thereby increasing the local temperature [62]. Besides, the metasurface with nanocavity structure further enhances the heat preservation capacity by thermal trapping effect.

2.4. Photothermal antibacterial effect in vitro

Bacterial infection is the main cause of peri-implantitis. It is crucial to control bacterial infection quickly, effectively, and safely to treat peri-implantitis. In the present work, photothermal antibacterial effects of CeO_x/TiO₂ was investigated. Fig. 5a₁ shows the agar plate colony images of S. aureus from various samples with or without NIR irradiation. Without NIR treatment, a large number of bacteria colonies are seen from various samples. However, when NIR irradiation was performed, the number of bacteria colonies from TiO₂ is significantly reduced, while almost no bacteria colonies are observed from CeO_x/TiO₂. Based on plate colony count method, the antibacterial rates of various samples against S. aureus were obtained. As presented in Fig. 5a₂, TiO₂ and CeO_x/TiO₂ show antibacterial rates of 90.78% and 99.96%, respectively. It indicates that under NIR irradiation, TiO₂ and CeO_x/TiO₂ with excellent photothermal properties can effectively kill bacteria, especially for CeO_x/TiO₂. A similar antibacterial trend is observed for E. coli (Fig. 5b₁, b₂) where the antibacterial rates of TiO₂ and CeO_x/TiO₂ against E. coli are 98.04% and 100.00%, respectively. To further verify the photothermal antibacterial properties, a live/dead staining was performed. As displayed in Fig. S7a, without NIR treatment, a large number of live S. aureus with green fluorescence be seen from all the samples, while the number of dead bacteria with red fluorescence are negligible. However, when NIR treatment was introduced, obvious live bacteria be observed from Ti and Ce-PIII. In contrast, the number of live bacteria from TiO₂ and CeO_x/TiO₂ are significantly reduced, and the amount of dead bacteria significantly increase. Similar trends are observed in E. coli from various samples (Fig. S7b). These findings further confirm the superior photothermal antibacterial activity of CeO_x/TiO₂.

Fig. 5.

The CeO_x/TiO₂ heterostructure exhibits NIR-responsive in-situ antibacterial effects. Bacterial colony plate images (a₁) and corresponding antibacterial rates (a₂) against S. aureus cultured on different samples with or without NIR irradiation. Bacterial colony plate images (b₁) and corresponding antibacterial rates (b₂) against E. coli cultured on different samples with or without NIR irradiation. SEM images of S. aureus (c₁) and E. coli (d₁) cultured on different samples with or without NIR irradiation conditions. Bacterial protein leakage results for S. aureus (c₂) and E. coli (d₂) cultured on different samples with or without NIR irradiation conditions. (e) Schematic illustration of the photothermal antibacterial properties of CeO_x/TiO₂. *p < 0.05; **p < 0.01; ***p < 0.001.

To gain a deep understanding of the antibacterial mechanism, SEM was employed to observe morphological changes on bacteria from various samples (Fig. 5c₁, d₁). Without NIR treatment, S. aureus and E. coli exhibit typical spherical and rod-shaped morphologies, respectively, on various samples, indicating good growth conditions. However, upon NIR treatment, bacteria on TiO₂ and CeO_x/TiO₂ exhibit significant membrane damage. This suggests that the photothermal effect induced by NIR irradiation can disrupt the integrity of bacterial cell membranes, thereby leading to bacteria death. Additionally, a bacterial protein leakage test was conducted to assess the extent of bacterial membrane damage (Fig. 5c₂, d₂). Protein leakages in S. aureus and E. coli from TiO₂ and CeO_x/TiO₂ both increase after NIR treatment, especially for CeO_x/TiO₂. It indicates that CeO_x/TiO₂ shows good photothermal antibacterial activity under NIR irradiation by increasing local temperature, destroying membrane permeability, leading to bacterial proteins leakage, ultimately resulting in bacterial death (Fig. 5e).

To evaluate the antibacterial potential of the CeO_x/TiO₂ coating in the peri-implantitis microenvironment, the anaerobic pathogen Porphyromonas gingivalis (P. gingivalis) was employed (Fig. S8). The coating demonstrated potent bactericidal activity against P. gingivalis upon NIR irradiation. Then the effects of NIR treatment on preformed S. aureus biofilms were assessed by crystal violet staining. Pronounced purple staining covered much of the non-irradiated surfaces, whereas irradiated TiO₂ and CeO_x/TiO₂ displayed weaker staining and discontinuous coverage with localized gaps. Quantification of crystal violet staining showed significantly lower absorbance in the irradiated TiO₂ and CeO_x/TiO₂ groups than in the control group (Fig. S9). These findings suggest that the coating can effectively address the challenge of anaerobic infections and biofilm formation in deep periodontal pockets through photothermal effect.

Peri-implantitis is essentially a complex dysbiotic disease involving multiple oral microorganisms. Single-species planktonic or monolayer biofilm models cannot fully simulate the complex interactions between pathogens and host immune cells under clinical pathological conditions. Therefore, future studies will focus on constructing multispecies biofilm models containing key genera such as P. gingivalis and Fusobacterium nucleatum (F. nucleatum) to deeply investigate the antibacterial selectivity and stability of the coating within complex micro-ecological competition. Furthermore, in vivo experiments will be conducted to comprehensively evaluate the long-term efficacy of the coating in treating peri-implantitis inflammation within a dynamic microbial physiological environment.

2.5. Soft tissue sealing properties in vitro

Lowering the local bacterial load may facilitate soft-tissue repair, and the improved tissue interface may further strengthen resistance to subsequent bacterial invasion. Soft tissue sealing refers to the formation of a tight and stable bond between soft tissues and implant abutment, which can effectively resist harmful bacteria in the oral environment, significantly reduce the risk of infection, and thus provide a strong guarantee for the long-term stability of the implant [63]. In the initial stage of implant abutment placement, biocompatibility of implant abutment and its ability to induce gingival fibroblasts to spread and adhere to its surface have a crucial impact on the formation of soft tissue sealing. Based on this, this study initially investigated the effects of the CeO_x/TiO₂ heterostructure on the viability, adhesion, and migration of gingival fibroblasts in a normal physiological microenvironment. As presented in Fig. 6a and Fig. S10, there are no significant differences in cell viability between Ti and CeO_x/TiO₂, indicating good biocompatibility of CeO_x/TiO₂. Furthermore, cell adhesion and spreading behaviors on various samples were investigated. As exhibited in Fig. 6b, after culturing for 1 h, the cells on all the samples remain spherical. At 4 h, cells on TiO₂ and CeO_x/TiO₂ exhibit a higher proportion of spreading and a larger spreading area, suggesting superior cell adhesion properties. When the culture time extended to 24 h, the cells on all the samples spread further and express abundant actin. It is worth noting that compared to Ti, HGFs on TiO₂ and CeO_x/TiO₂ show more filopodia and lamellipodia, which are key components of cell adhesion and spreading [64]. It suggests that TiO₂ and CeO_x/TiO₂ can induce better adhesion and spreading effects of HGFs. To establish a robust soft tissue sealing, cell migration ability of gingival fibroblasts on implant abutment is also crucial [65]. Thus, a scratch test was used to evaluate the cell migration ability of HGFs on various samples. As presented in Fig. 6c and d, after culturing for 24 h, HGFs on all the samples migrate towards the scratch center to some extent. Among them, CeO_x/TiO₂ shows the greatest cell migration distance, which suggests that CeO_x/TiO₂ heterostructure can significantly promote fibroblast migration, thereby more effectively participating in the construction of the epithelial barrier.

Fig. 6.

The CeO_x/TiO₂ heterostructure facilitates soft tissue sealing by enhancing the adhesion, spreading, migration of HGFs, and protecting them from damage caused by H₂O₂, NIR, and hyperthermia. (a) Cell viability of HGFs cultured on Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂ for 1, 4, and 7 days. (b) Cytoskeleton staining images of HGFs cultured on Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂ for 1, 4, and 24 h. Scratch assay results (c) and corresponding fluorescence images (d) of HGFs cultured on Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂ for 0, 4, and 24 h (e₁-e₄) Relative ROS levels of HGFs cultured on Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂ without or with stimulation with NIR, H₂O₂, and hyperthermia at 7 day, and (f₁-f₄) corresponding cell viability of HGFs from Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂. (g) Schematic illustration of the antioxidant properties of CeO_x/TiO₂. *p < 0.05; **p < 0.01; ***p < 0.001.

The occurrence of peri-implantitis triggers the migration of immune cells to the inflammatory site, resulting in the production of a significant amount of ROS. Excessive ROS disrupts the intracellular redox balance, triggering oxidative stress and damaging biological macromolecules such as proteins, lipids, and DNA, which exacerbates the inflammatory response. It can damage gingival and alveolar bone tissues, affecting the integration of soft tissues with implants, and ultimately result in loosening or failure of implants [66]. Although TiO₂-based heterojunctions typically exhibit the characteristic of photocatalytically generating ROS under illumination [67], in the inflammatory microenvironment investigated in this study, we observed that CeO_x/TiO₂ exhibits a completely opposite effect. As shown in Fig. 6(e₁-e₄), the intracellular ROS level of cells on CeO_x/TiO₂ is significantly reduced under H₂O₂ stimulation. Moreover, CeO_x/TiO₂ exhibits excellent ROS scavenging ability under NIR irradiation and hyperthermia conditions, which further indicates that the antioxidant mechanism of CeO_x predominates in this system, effectively counteracting any potential photocatalytic oxidative effects. Previous studies have shown that nanoceria, through a combination of superoxide dismutase (SOD)-like and catalase (CAT)-like activities, eliminates both superoxide and hydrogen peroxide simultaneously through a series of consecutive reactions. When nanoceria reduces superoxide, hydrogen peroxide is formed, and Ce³⁺ is oxidized to Ce⁴⁺; subsequently, Ce⁴⁺ can react with hydrogen peroxide, regenerating Ce³⁺ and oxidizing hydrogen peroxide to oxygen [68]. Based on this, CeO_x/TiO₂ utilizes the reversible conversion between Ce⁴⁺ and Ce³⁺ to endow the coating with SOD and CAT enzymatic activities, enabling it to continuously and effectively eliminate ROS. At the same time, the coupling of the two reversible redox steps effectively avoids the toxicity of hydrogen peroxide generated by the SOD-like enzyme, making it an ideal antioxidant with self-regenerating capabilities [69]. Therefore, HGFs on CeO_x/TiO₂ exhibit good cell viability under H₂O₂, NIR, and hyperthermia conditions, which suggests that CeO_x/TiO₂ heterostructure can not only resist oxidative stress but also effectively reduce cell damage caused by NIR irradiation and hyperthermia (Fig. 6(f₁-f₄)). The results provide important support for the application of modified implant abutments in photothermal therapy. Thus, it can be concluded that CeO_x/TiO₂ heterostructure not only effectively promotes the adhesion, spreading, and migration of gingival fibroblasts, thereby establishing a robust soft tissue sealing, but also continuously and effectively eliminates intracellular ROS through the reversible conversion between Ce⁴⁺ and Ce³⁺ (Fig. 6g), significantly protecting gingival fibroblasts from damage caused by various adverse factors such as oxidative stress, NIR irradiation, and hyperthermia environments.

In the process of constructing the epithelial barrier, the extracellular matrix components secreted by fibroblasts, particularly collagen and fibronectin, play a pivotal role. These matrix components not only provide necessary structural support for the barrier but also regulate cellular behavior through various mechanisms, thus ensuring the integrity and functionality of the epithelial barrier and providing a protective layer against oral pathogenic bacteria [70]. To preliminarily investigate the regulatory potential of CeO_x/TiO₂ coating on the expression of soft tissue sealing-related genes, qRT-PCR assays were conducted under normal culture conditions to evaluate the expression of fibronectin (FN), type I collagen (COL I), and integrin β1. As displayed in Fig. 7(a₁-a₃), CeO_x/TiO₂ significantly upregulates the mRNA expression levels of FN, COL I, and integrin β1 in HGFs. The regulatory effect of CeO_x/TiO₂ heterostructure on COL I and FN protein expression was further elucidated through immunofluorescence staining, accompanied by quantitative analysis of fluorescence intensity using ImageJ software. As presented in Fig. 7(b₁, b₂, c₁, c₂), cells on CeO_x/TiO₂ exhibit significantly higher red (representing COL I) and green (representing FN) fluorescence intensities compared to Ti, Ce-PIII, and TiO₂. Concurrently, western blot data demonstrate that CeO_x/TiO₂ heterostructures raise the protein levels of FN, integrin β1 and COL I in HGFs (Fig. 7(d₁-d₄), which indicates that CeO_x/TiO₂ significantly enhances the extracellular matrix production and cell adhesion capabilities of HGFs by upregulating the expression of key genes and promoting protein secretion. These findings provide a strong scientific basis for optimizing the design of implant abutments and improving their integration with soft tissues. The results obtained under normal physiological conditions collectively reveal the dual advantages of the CeO_x/TiO₂ heterostructure in soft tissue sealing. On the one hand, it ensures cell survival in the inflammatory environment by scavenging ROS, which serves as a prerequisite for the soft tissue sealing function. On the other hand, by upregulating the expression of key ECM genes and promoting the secretion of related proteins, it enhances the ECM generation, adhesion, spreading, and migration capabilities of HGFs, indicating its potential to rapidly initiate barrier repair after abutment placement, thereby preventing bacterial invasion. Combined with its excellent antioxidant properties, these results suggest that the heterostructure is expected to maintain soft tissue sealing functions in inflammatory or oxidative stress environments, although this inference requires further direct verification in future oxidative stress models. These findings provide a scientific basis for optimizing the design of implant abutments and improving their integration with soft tissues.

Fig. 7.

The CeO_x/TiO₂ heterostructure promotes soft tissue sealing by enhancing the expression of extracellular matrix synthesis and cell adhesion-related genes and proteins in HGFs. Relative gene expression levels of Fibronectin (a₁), COL Ⅰ (a₂), and Integrin β1(a₃) from HGFs cultured on different samples. COL I (b₁) and fibronectin (c₁) staining images and corresponding fluorescence intensities (b₂, c₂) of COL I and fibronectin from HGFs cultured on different samples. Representative western blot results (d₁) and quantitative results of Fibronectin (d₂), Integrin β1 (d₃) and COL Ⅰ (d₄) from HGFs cultured on different samples. *p < 0.05, **p < 0.01, ***p < 0.001 compared with Ti. &p < 0.05, &&p < 0.01, &&&p < 0.001 compared with TiO₂. #p < 0.05, ##p < 0.01, ###p < 0.001 compared with Ce-PIII.

2.6. Treatment of peri-implantitis and the effect of soft tissue sealing in vivo

To evaluate the therapeutic effect of CeO_x/TiO₂ abutments on peri-implantitis in vivo, the peri-implantitis model was constructed using a silk ligature method around the second and third molars of beagle dogs [71]. In the absence of oral care, significant plaque accumulation was observed around the healing abutments on 4th week after silk ligature (Fig. S11a), which is a crucial characteristic of the onset and progression of peri-implantitis. Periodontal probing reveals that the mean probing pocket depth (PPD) across all groups ranges between 4 and 6 mm (Fig. S11b and c), further confirming the inflammation and destruction of the bone tissues surrounding the implants. X-ray radiographic analysis visually demonstrates the progressive loss of alveolar bone during the silk ligature process, with a typical V-shaped defect formed, suggesting the successful establishment of peri-implantitis models in beagle dogs.

Once the peri-implantitis model had been established, the healing abutments without surface treatment were removed, and therapeutic abutments coated with CeO_x/TiO₂ heterostructure were installed (Fig. 8a). Simultaneously, NIR light irradiation was applied once a week maintaining a temperature of 50 °C for 5 min to treat the peri-implantitis (Fig. 8b). Thermal imaging reveals that surface temperature of the abutments stabilizes at 50 °C after irradiation for 1 min. This temperature setting is based on the dual considerations of biosafety and in vivo photothermal response. Previous studies have identified 50 °C as the critical safety threshold to avoid bone thermal necrosis [72], [73]. Moreover, constrained by the optical attenuation of biological tissues and the heat dissipation effect of blood perfusion, 50 °C also represents the steady-state thermal equilibrium temperature achievable for this material in a physiological environment (Fig. S12). No obvious redness or other adverse reactions were observed in the treated areas after irradiation. After 4 weeks of treatment, a slight decrease in PPD is observed from Ti group, while a significant reduction in PPD can be found from CeO_x/TiO₂ group (Fig. S13). Meanwhile, the plaque adhering to the abutment surface also decreases significantly (Fig. S14), which suggests that abutments with CeO_x/TiO₂ heterostructure exhibit a significant therapeutic effect on peri-implantitis under NIR light irradiation. These results compellingly demonstrate that, although 50 °C is lower than the steady-state temperature achieved in the in vitro air environment (∼120 °C), this temperature is sufficient to achieve significant antibacterial effects through the photothermal effect in the complex oral physiological environment without causing thermal damage to the surrounding tissues. ELISA analysis (Fig. 8c₁, c₂) of gingival crevicular fluid (GCF) from the beagle dogs shows that the initial levels of inflammatory markers TNF-α and IL-6 are similar with no significant difference after establishment of the peri-implantitis model. After treatment for 4 weeks, the levels of TNF-α and IL-6 from Ti and CeO_x/TiO₂ groups decrease, especially for CeO_x/TiO₂ group, indicating the effectiveness of CeO_x/TiO₂ abutments on reducing the inflammatory response caused by peri-implantitis. HE staining results reveal a large number of inflammatory cells infiltrating in the surrounding gingival tissues from Ti group, whereas no significant infiltration of inflammatory cells is observed from CeO_x/TiO₂ group (Fig. S15). Qualitative and quantitative results from immunofluorescence staining show that increased expression of CD206 and decreased expression of iNOS can be found in the gingival soft tissues from CeO_x/TiO₂ group (Fig. 8(d₁, d₂, e₁, e₂)). As a key enzyme in the inflammatory process, reduced expression of iNOS implies decreased production of NO during the inflammatory response, which helps to alleviate tissue damage and inflammation [74]. Meanwhile, the increased expression of CD206 indicates an enhancement of the anti-inflammatory response and polarization of macrophages towards the M2 phenotype, which promotes tissue repair and regeneration [75]. Overall, the changes in iNOS and CD206 expression suggest that CeO_x/TiO₂ abutments may modulate the local macrophage response toward a more reparative phenotype in peri-implantitis.

Fig. 8.

The CeO_x/TiO₂ heterostructure alleviates the inflammatory response of peri-implantitis and facilitates the soft tissue sealing in vivo. (a) The optical photo of the replacement of modified abutment. (b) Treatment with NIR irradiation on beagle dog. ELISA analysis of IL-6 (c₁) and TNF-α (c₂) in gingival crevicular fluid (GCF). Immunohistofluorescence staining results of CD206 (d₁), iNOS (e₁), BP180 (f) and VEGF (g₁) from the tissues around the samples and corresponding quantitative analysis of CD206 (d₂), iNOS (e₂), and VEGF (g₂). (h₁) Images of immunohistochemical staining of COL І from the tissues around the samples and (h₂) corresponding quantitative analysis. (i) Representative photographs of micro-CT reconstruction from Ti and CeO_x/TiO₂. The quantified data of the bone volume/total volume (BV/TV, j₁), trabecular thickness (Tb. Th, j₂) bone-to-implant contact (BIC, j₃) and trabecular number (Tb. N, j₄). *p < 0.05; **p < 0.01; ***p < 0.001.

Moreover, the qualitative and quantitative results of immunohistochemical and immunofluorescence staining reveal that the expressions of BP180, VEGF, and COL Ⅰ in the surrounding gingival tissues are significantly increased after treatment for 4 weeks (Fig. 8(f, g₁, g₂, h₁, h₂₎). BP180, a component of hemidesmosomes, is involved in the adhesion of basal keratinocytes to the extracellular matrix. Increased expression of BP180 in the cell membrane indicates the formation of hemidesmosomes, which is a crucial marker for the stable connection between gingival tissue and the implant abutment surface, contributing to enhancing the soft tissue sealing [76]. VEGF, a key factor promoting angiogenesis, exhibits increased expression that aids in stimulating blood vessel formation in the gingival tissues, improving local blood circulation, and providing sufficient nutrients and oxygen for the repair and regeneration of the gingival tissue [77]. COL Ⅰ, the primary collagen component in gingival tissues, plays a vital role in maintaining the structure and function of the gingival tissues. Its increased expression suggests enhanced repair and regeneration capabilities of the gingival tissues, facilitating the formation of a tighter soft tissue sealing [78]. Therefore, the increased expressions of BP180, VEGF, and COL Ⅰ collectively contribute to the formation and maintenance of the soft tissue sealing, enhancing the stable connection between the gingival tissues and the implant abutment surface, helping prevent bacterial invasion and the occurrence of inflammatory reactions, thereby improving the long-term stability and success rate of the implants.

To further explore the therapeutic effects of CeO_x/TiO₂ group, micro-CT tomography analysis was conducted (Fig. 8i). The results show that, compared with the Ti group, the CeO_x/TiO₂ group exhibits a certain degree of increase in bone volume/total volume (BV/TV) value, bone-implant contact (BIC), trabecular thickness (Tb. Th), and trabecular number (Tb.N) (Fig. 8(j₁-j₄)). Therefore, this finding further supports the positive impact of CeO_x/TiO₂ on bone tissue regeneration and remodeling. To observe the therapeutic effect of the CeO_x/TiO₂ group more intuitively, Van Gieson staining of bone tissues was also performed (Fig. S16). The results demonstrate that after treatment, denser bone tissues are formed around the implants in the CeO_x/TiO₂ group, and the original V-shaped defects do not progress further, which fully proves the significant effects of CeO_x/TiO₂ in promoting osseointegration. Despite the promising results, this study has certain limitations. First, although the paired design helped control for inter-animal variability, the sample size was relatively small (n = 4 per group). This limited the statistical power and made it difficult to robustly verify the assumption of normal distribution. Second, regarding the microbiological analysis, we did not perform metagenomic sequencing or qPCR for specific anaerobes, lacking direct quantitative data on pathogenic community shifts; plaque staining served only as a qualitative reference. Third, while the frequency and parameters of the repeated NIR photothermal intervention were effective in the animal model, they require further optimization to align with real-world clinical scenarios for supportive peri-implantitis therapy (SPT). Future studies will incorporate larger sample sizes, dynamic monitoring of pathogenic flora, and optimized treatment protocols to rigorously establish the clinical efficacy of this strategy. And combining temperature monitoring with irradiation time control can further reduce unnecessary thermal exposure of surrounding tissues.

2.7. Transcriptome-based mechanism exploration of CeO_x/TiO₂ heterostructure on promoting soft tissue sealing

To further investigate the mechanism of CeO_x/TiO₂ heterostructure on promoting soft tissue sealing, transcriptome sequencing of HGFs co-cultured with various samples were performed. The results demonstrate good reproducibility among samples within each group (Fig. 9a), laying a foundation for the reliability of subsequent data analysis. Differential expression analysis reveals that, compared with Ti group, 1144 genes are significantly upregulated and 86 genes are significantly downregulated from CeO_x/TiO₂ group (Fig. 9b), indicating a predominantly activated transcriptional response that may contribute to enhanced cellular functions and soft tissue sealing.

Fig. 9.

Underlying mechanism of CeO_x/TiO₂ heterostructure in inducing soft tissue sealing. (a) Correlation analysis of HGFs co-cultured with Ti and CeOx/TiO₂. (b) Volcano plot of HGFs co-cultured with Ti and CeO_x/TiO₂. GO functional enrichment analysis (c) and KEGG enrichment analysis (d) of HGFs co-cultured with CeO_x/TiO₂ vs. Ti. (e) Heatmap of representative genes expression of HGFs co-cultured with Ti and CeO_x/TiO₂. (f) Representative fluorescence images of intracellular Ca²⁺ in HGFs from different groups.

GO enrichment analysis classified the differentially expressed genes into biological process, cellular component, and molecular function categories (Fig. 9c). The enrichment of antioxidant activity-related terms is consistent with the intrinsic antioxidant capacity of the CeO_x/TiO₂ heterostructure. In addition, Extracellular matrix organization and wound healing were among the prominently enriched biological processes, reflecting active matrix remodeling and tissue repair. Changes in cell–substrate junction and focal adhesion further suggested strengthened cellular attachment at the material–tissue interface. Notably, the enrichment of the collagen-containing extracellular matrix suggested that the CeO_x/TiO₂ surface may promote collagen deposition and strengthen the structural basis of the soft tissue interface. KEGG enrichment analysis identified the Rap1 signaling pathway as a major regulatory pathway in the CeO_x/TiO₂ group (Fig. 9d). Further analysis revealed 27 differentially expressed genes associated with repair process (Fig. 9e). As a central signaling hub, Rap1 could integrate extracellular cues and coordinates intracellular responses related to HGF adhesion and extracellular matrix formation.

Transcriptomic analysis revealed enrichment of the neuroactive ligand–receptor interaction (NLRI) pathway, which serve as an important upstream input for Rap1 signaling [79]. The observed upregulation of the NLRI-related genes F2RL1 and FPR1 further indicates enhanced GPCR-mediated PLC signaling, which could promote PIP₂ hydrolysis into IP₃ and subsequently increase Ca²⁺ release (Fig. 9e). The upregulation of ADCY family members could enhance upstream cAMP signaling and thereby facilitate Rap1 activation [80]. The downstream PARD6A polarity complex further support directional cell migration and the organization of a functional soft tissue sealing layer. ECM–receptor interaction pathways were also enriched. Rap1-associated signaling may subsequently enhance ECM–receptor interactions and promote the synthesis and organization of matrix components such as COL1A1, FN1, VTN, and LAMC1(Fig. 9, Fig. 10).

Fig. 10.

Schematic diagram of the potential mechanism of CeO_x/TiO₂ heterojunction structure promoting HGFs-mediated soft tissue sealing. Created in https://BioRender.com.

To validate this mechanism, co-staining with Fluo-4 AM fluorescent probe and Hoechst 33342 to quantitatively compare intracellular Ca²⁺ concentrations in HGFs between Ti and CeO_x/TiO₂ groups was carried out (Fig. 9f and Fig. S17). The results reveal that the calcium fluorescence intensity in HGFs from CeO_x/TiO₂ group is significantly higher than that from Ti group. Further analysis of the expression levels of key signaling proteins (including ADCY3, RAP1A, MMP2, and TGF-β1) in the studied pathway was conducted via Western blot. The CeO_x/TiO₂ group exhibited a consistent upregulation of these proteins relative to the control, in strong agreement with the bioinformatic predictions. This concordance indicates that the CeO_x/TiO₂ heterostructure effectively triggers downstream signaling cascades, particularly those associated with Ca²⁺-dependent pathways, leading to elevated intracellular Ca²⁺ dynamics. Collectively, these molecular events contribute to the enhancement of cellular redox homeostasis and functional activity (Fig. S18). This finding provides critical molecular evidence for the regulation of cellular signal transduction and downstream functions (e.g., ECM remodeling) by CeO_x/TiO₂ heterostructure, elucidating the mechanistic link by which the material's surface chemistry orchestrates soft tissue sealing through the receptor sensing, signal transduction, and ECM remodeling cascade. It should be noted that the proposed “neuroactive ligand-receptor interaction → Rap1 signaling → ECM-receptor interaction” pathway is mainly supported by bioinformatic predictions, protein expression correlations, and Ca²⁺ fluorescence comparisons, but lacks causal verification via perturbation experiments (e.g., receptor blockade, gene knockdown, or specific inhibitors). Thus, its actual role in CeO_x/TiO₂-regulated HGFs functions requires further experimental validation. Overall, the system combines photothermal control of local bacterial burden with antioxidant protection and regulation of cellular functions, integrating infection intervention and soft-tissue sealing within a single material design (Fig. 11).

Fig. 11.

Schematic illustration of the antibacterial and soft-tissue sealing mechanisms of the CeO_x/TiO₂ coating. NIR-induced photothermal effects reduce bacterial viability, while CeO_x-mediated ROS scavenging helps protect gingival fibroblasts from oxidative stress. The activation of the Rap1-associated signaling network may promote cell adhesion and extracellular matrix formation, supporting soft-tissue integration.

3. Conclusion

In summary, an optical metasurface CeO_x/TiO₂ heterostructure was constructed on titanium abutment surfaces using plasma immersion ion implantation and thermal treatment technology. CeO_x/TiO₂ heterostructure with optimized bandgap enhances the light absorption in the second NIR region and robust optical metasurface with gradient-distributed nanocavities improves the retention of NIR light by light trap effect and increases the heat preservation capacity, leading to superior photothermal antibacterial activity. The metasurface CeO_x/TiO₂ heterostructure exerts dual therapeutic effects: At the preventive stage, it can enhance soft tissue barrier function by promoting cell migration, adhesion, spreading, and ECM synthesis of HGFs. At the therapeutic stage, it can achieve precise targeted antibacterial action through photothermal effects, scavenge ROS via the antioxidant activity of CeO_x, suppress inflammation, and create a regenerative tissue microenvironment. Mechanism investigations demonstrate that metasurface CeO_x/TiO₂ heterostructure synergistically regulates cellular behavior and matrix remodeling by sequentially activating the “Neuroactive ligand-receptor interaction, Rap1 signaling pathway, and ECM-receptor interaction” cascade network. The metasurface CeO_x/TiO₂ heterostructure not only establishes a novel paradigm for optimizing photothermal-bioactivity coupling through interfacial electron modulation but also pioneers new directions for developing smart-responsive multifunctional implant abutments based on metasurface heterostructures. It holds promise for significantly improving long-term implant stability and clinical success rates. Although this study has demonstrated the excellent biological performance of the CeO_x/TiO₂ heterostructure coating, its long-term stability under cyclic mechanical loads, such as brushing or chewing, has not been evaluated. Future studies will introduce mechanical loading models simulating oral functions to further verify the durability and bonding strength of the coating in complex mechanical environments, thereby comprehensively assessing its clinical application potential.

4. Materials and methods

4.1. Sample preparation

Titanium plates (10 × 10 × 1 mm³, purchased from Xi'an Saite New Material Science and Technology Co., LTD, China) were ultrasonically cleaned using a mixed acid solution (HF: HNO₃: deionized water = 1:5:34). Then they were ultrasonically cleaned with ethanol and deionized water, and the corresponding samples were designated as Ti. Cerium (Ce) ions were implanted into titanium surfaces using a plasma immersion ion implantation (PIII) technique, and the corresponding samples were designated as Ce-PIII. The specific parameters are listed in Table 1. Briefly, Ti substrates were mounted on the sample holder of the ion implantation system, and the chamber pressure was reduced to 5 × 10⁻³ Pa with Ar gas maintained. The Ce cathode (Table S2) was ionized under pulsed high voltage, and Ce ions were accelerated and implanted into the substrate under a negative high-voltage bias. The implantation voltage, frequency, and pulse width were 30 kV, 5 Hz, and 500 μs, respectively, with the negative bias current fluctuating between 9.4 and 11.0 mA. Subsequently, Ti and Ce-PIII samples were heated to 550 °C in a muffle furnace with a heating rate of 5 °C/min and remained for 1 h. The resulting samples were designated as TiO₂ and CeO_x/TiO₂, respectively. In the animal experiment, the Ti group refers to commercially pure titanium healing abutments, while the CeO_x/TiO₂ group refers to commercially pure titanium healing abutments subjected to the same treatment protocol as the CeO_x/TiO₂ plate group.

Table 1.

The parameters used during plasma ion immersion implantation process.

Parameters	values
Target Voltage pulse duration (μs)	500
Cathodic arc Voltage pulse duration (μs)	500
Pulsing frequency (Hz)	5
Voltage (kV)	−30
Time (min)	90
Pressure (Pa)	5.0 × 10−3
The negative bias current (mA)	9.4∼11.0

4.2. Surface characterization

Prior to observation, samples were cleaned with ethanol to remove surface contaminants and air-dried in a clean bench. The surface morphologies of the samples were observed using scanning electron microscopy (SEM, accelerating voltage = 10 kV, SU8220, Hitachi, Japan). The phase compositions of the samples were analyzed using X-ray diffraction (XRD, Cu Kα, 30 kV, step size = 0.02°, D2 phaser, Bruker, Germany). X-ray photoelectron spectroscopy (XPS, Al Kα, AXIS, Japan) was used to examine the surface elemental compositions (C 1s peak = 284.8 eV). The water contact angles were measured using a contact angle tester (SL200B, Solon, China). The hardness variation within a 100 nm depth of the samples was measured using a nanoindenter (G200, Agilent, Santa Clara, USA). A Berkovich diamond indenter was applied with a maximum load of 500 mN, a loading rate of 200 μN/s, and a dwell time of 10 s. Five different locations were randomly selected on each sample for measurement. Aberration-corrected scanning transmission electron microscope (STEM) was performed on a probe corrected Thermo Scientific™ Spectra 300 S/TEM equipped with a Super-X Detector and operated at an accelerating voltage of 300 kV. A semi-angle of convergence of 30 mrad was utilized. The atomic-scale high angle annular dark-field (HAADF) images were collected with a detector semi-angle range of 64–200 mrad and the integrated differential phase contrast (iDPC) images were acquired by a segmented DF4 detector with 4 quadrants. The cross-sectional high-resolution TEM samples were prepared by in situ sectioning using focused ion beam (FIB) milling. Prior to FIB processing, a protective Pt layer was deposited on the sample surface to minimize ion-beam-induced damage to the nanostructure.

4.3. Corrosion resistance

An electrochemical workstation (PGSTAT302N, AUTOLAB, Netherlands) was used to obtain the Tafel curves of Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂. Saline (Commercial 0.9% NaCl saline, baseline pH ≈ 7.2) was used as an electrolyte. A three-electrode system was employed during the tests: a saturated calomel electrode served as the reference electrode, a graphite rod was used as the counter electrode, and the sample was the working electrode. To obtain a stable open-circuit potential, samples were immersed in saline for 10 min prior to testing. The tests were conducted at room temperature with a scan rate of 10 mV/s over a range of −1.0 to 1.0 V. All the experimental data were recorded relative to the reference electrode.

The release of Ti and Ce from CeO_x/TiO₂ was evaluated in artificial saliva (Solarbio, China) under different pH and NIR conditions. The pH of the acidic artificial saliva was adjusted to 2.5 using acetic acid, while neutral artificial saliva was maintained at pH 7.0. For the NIR groups, samples were irradiated with a 1064 nm laser at a power density of 0.75 W/cm² for 5 min per day. The immersion solutions (5 mL) were collected after 1, 7, and 14 days. Ti and Ce concentrations were quantified by ICP-MS (Agilent 7850, USA). Values below the detection limit were recorded as ND.

4.4. PTA detection

Hydroxyl radicals were quantified using purified terephthalic acid (PTA) detection. Hydroxy radicals will react with PTA to form hydroxyterephthalic acid, whose fluorescence intensity is directly proportional to the concentration of the hydroxyl radicals in the solution. A PTA solution was prepared by dissolving 5 mM PTA and 10 mM sodium hydroxide in PBS. Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂ were then immersed in 4 mL of PTA solution. One set of samples was treated with 500 μM H₂O₂, while the other set remained untreated. The samples were kept in the dark at 37 °C for three days. The fluorescence emission spectra of the PTA solutions were measured using a microplate reader (E_x = 310 nm, BioTek, Cytation5, USA).

4.5. Photothermal performance

To evaluate the photothermal performance of the samples, a 1064 nm NIR light source was employed with a thermal imaging camera (Fotric 285s, USA) used to monitor the sample temperature in real time. The CeO_x/TiO₂ samples were irradiated at different power densities (0.25, 0.50, and 0.75 W/cm²). The photothermal cycling stability of CeO_x/TiO₂ was assessed under 0.50 W/cm² NIR light irradiation, temperature changes were monitored and recorded for Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂ in air or immersed in 1 mL of ultrapure water for 10 min. Additionally, Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂ were implanted subcutaneously into mice and their photothermal performance in vivo was also investigated by irradiation with 0.75 W/cm² NIR light for 10 min.

4.6. Computational details

All density functional theory (DFT) calculations were performed using the Vienna Ab Initio Simulation Package (VASP). The generalized gradient approximation method with the Perdew-Burke-Ernzerhof (PBE) exchange-correlation function was used to manage the electron exchange and correlation energy. The plane wave basis (kinetic energy cut-off value 450 eV) was used to describe the valence electrons. U value was set to 5.0 eV for the Ce 4f orbitals to better describe the localized electronic states. The heterostructure model was built based on TiO₂ (110) and CeO₂ (101) with the mesh of 4*3*1 for the k-point sampling obtained from the gamma center. The thickness of the vacuum layer is 15 Å. The atomic positions were fully optimized until the energy and forces converged to 1*10⁻⁵ eV and 0.03 eV Å⁻¹, respectively.

4.7. Cooling curve measurement

Ti and CeO_x/TiO₂ samples were heated on a heating stage (room temperature: 25.5 °C). Upon reaching thermal equilibrium, the samples were removed, and their cooling curves was recorded using an infrared thermal camera (Fotric 285s, USA).

4.8. Bacterial culture

The antibacterial performance of the samples was evaluated using the standard strains of Gram-positive Staphylococcus aureus (S. aureus, ATCC 25923), Gram-negative Escherichia coli (E. coli, ATCC 25922) and Porphyromonas gingivalis (BNCC337441). The samples were sterilized with 75% ethanol and placed in 24-well plates. The bacteria in logarithmic growth phase were diluted with saline to a concentration of 10⁷ CFU/mL. Then, 60 μL bacterial suspension was added on the surface of each sample and incubated at 37 °C.

4.9. Bacterial plate counting

After bacteria incubation for 24 h, Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂ were divided into two groups, each in 24-well plates with 1 mL of saline. One group was irradiated with 1064 nm NIR light at a density of 0.5 W/cm² for 10 min, while the other group was left untreated. The samples were then transferred to tubes containing 4 mL of saline and vigorously shaken for 1 min. Bacterial suspensions were serially diluted 10, 100, and 1000 times. Each diluted suspension (100 μL) was spread on an agar plate and incubated for 18 h. The number of colonies on the agar plates was counted (according to GB/T 4789.2) and the antibacterial rate was calculated as follows,

$$\text{Antibacterial}\text{rate}(\%)=(\mathrm{A}-\mathrm{B})/\mathrm{A}\times 100\%$$ (1)

where A is the average number of colonies from the Ti sample and B is the average number of colonies on the experimental groups.

4.10. Live/dead staining

After bacteria incubation for 24 h, Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂ were divided into two groups, each in 24-well plates with 1 mL saline. One group was irradiated with 1064 nm NIR light at a density of 0.5 W/cm² for 10 min, while the other group was left untreated. The samples were then transferred to new 24-well plates, rinsed with saline, and stained with a bacteria live/dead staining kit (L13152, Molecular Probes, USA). The stained bacteria were observed and photographed using a fluorescence microscope (Olympus, Japan).

4.11. Bacteria morphology and biofilm observation

After bacteria incubation for 24 h, Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂ were divided into two groups. One group was irradiated with 1064 nm NIR light at a density of 0.5 W/cm² for 10 min, while the other group was left untreated. The samples were rinsed twice with saline and fixed with 2.5% glutaraldehyde solution for 4 h. The samples were then dehydrated with a series of ethanol solutions (30%, 50%, 75%, 90%, and 100% dehydrated alcohol). The dehydrated samples were dried using a series of ethanol and hexamethyldisilazane solutions (volume ratios of 2:1, 1:1, 1:2, and 0:1). Then the bacteria morphology was observed using a scanning electron microscope (Hitachi, Japan).

Crystal violet staining was used to quantify S. aureus biofilm formation. Bacterial suspensions (10⁸ CFU/mL) were cultured directly on the different sample surfaces. After incubation (immersed in 1 mL TSB medium for 2 days at 37 °C), the samples were gently washed with PBS to remove non-adherent bacteria and stained with crystal violet (0.2%, Adamas, China). Excess dye was removed by washing with 75% alcohol, and the retained crystal violet was subsequently solubilized. The absorbance at 560 nm was measured to quantify the biofilm biomass.

4.12. Protein leakage

The BCA protein assay kit (Thermo Fisher Scientific, USA) was used to measure protein leakage from S. aureus and E. coli cultured on various samples with NIR irradiation. To prevent protein interference in the bacterial culture medium, the bacterial suspension was centrifuged (8000 rpm, 5 min) and resuspended in saline. The absorbance of the S. aureus and E. coli suspensions was adjusted to 0.1 (OD₆₀₀) using saline solution. Sterilized samples were placed in a 24-well plate and 500 μL of the bacterial suspension was added to each well and incubated at 37 °C for 2 h. Then, Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂ were divided into two groups, one group was irradiated with 1064 nm NIR light at a power density of 0.5 W/cm² for 10 min, while the other group was left untreated. The bacterial suspensions were then centrifuged (10000 rpm, 5 min), and 20 μL of the supernatant was added to 200 μL of the BCA working solution. The mixtures were then incubated at 37 °C for 30 min. The absorbance at 562 nm was measured using a microplate reader (BioTek, USA) and the protein concentration was calculated based on a standard curve.

4.13. Cell culture

Human gingival fibroblasts (HGFs) were extracted from orthodontic patients with informed consent. The patients had no history of smoking or obvious inflammatory symptoms. HGFs were cultured in DMEM (high glucose, Gibco, USA) supplemented with 15% fetal bovine serum (FBS, Gibco, USA) and 1% penicillin/streptomycin solution (100 U/mL, Gibco, USA). The cells were cultured at 37 °C with 5% CO₂. HGFs were passaged every two days, and cells of 3-5 passages were used for subsequent experiments.

4.14. Cell compatibility

HGFs were seeded on Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂ at a density of 2.0 × 10⁴ cells/well and cultured for 1, 4, and 7 days. At each time point, 500 μL DMEM containing 10% AlamarBlue™ (Thermo Fisher Scientific, USA) was introduced and cultured for 2 h in the dark. Then, the fluorescence intensity of the aforementioned solution was measured using a spectrophotometer (E_x = 560 nm, E_m = 590 nm, Thermo Scientific, USA). Besides, Live/Dead staining of HGFs cultured with various samples for one day were performed using a Live/Dead staining kit (Thermo Fisher Scientific, USA) and the fluorescence images were taken using a fluorescence microscope (Ts2-FL, Nikon, Japan).

To assess the protective effect on cells, different culture conditions were set up to measure the changes in HGFs’ cell viability and reactive oxygen species (ROS) production. For oxidative stress assessment, after culturing for 7 days, 500 μM H₂O₂ was added and incubated for another 24 h. For heat stress evaluation, after culturing for 7 days, the cells were transferred to a 50 °C water bath and cultured for another 10 min, and another group was irradiated using 1064 nm NIR light with 0.50 W/cm² for 10 min. The control group received no treatment. All groups were then cultured for an additional 6 h. Finally, cell viability and ROS levels were then measured using AlamarBlue™ (Thermo Fisher Scientific, USA) and DCFH-DA reagent (Beyotime-Biotech Co., China), respectively.

4.15. Cell adhesion and spreading

HGFs with a density of 2.0 × 10⁴ cells/well were seeded on Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂ and cultured for 1, 4, and 24 h. Then the samples were gently washed twice with PBS to remove non-adherent cells, and the cells on the samples were fixed overnight with 4% paraformaldehyde in the dark, permeabilized with permeabilization solution (Beyotime-Biotech Co., China) for 10 min, and blocked with 5% bovine serum albumin (BSA, Aladdin, China) in PBS for 1 h to reduce non-specific binding. To visualize adhered fibroblasts, samples were incubated with Phalloidin FITC Reagent (1:1000 dilution, Abcam, UK) at room temperature for 1 h in the dark. After washing with PBS, cell nuclei were counterstained with 4’,6-diamidino-2-phenylindole (DAPI, Merck, Germany) for 5 min. Cells were then observed using a fluorescence microscope (Ts2-FL, Nikon, Japan).

4.16. Immunofluorescence staining

HGFs with a density of 2.0 × 10⁴ cells/well were seeded on Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂ and cultured for 2 days. Then, the cells were fixed and blocked, as described in section 4.15. Subsequently, the samples were then incubated with rabbit anti-human COL-I or rabbit anti-human fibronectin primary antibodies (1:200 dilution; ImmunoWay, USA) at 4 °C overnight in the dark. After washing with PBS, the samples were incubated with Alexa Fluor 488 goat anti-rabbit or Alexa Fluor 555 donkey anti-rabbit secondary antibody (1:500 dilution, Beyotime-Biotech Co., China) at 4 °C for 2 h in the dark. Cell nuclei were counterstained with DAPI (Merck, Germany) for 5 min. The target protein expression was observed and photographed using a fluorescence microscope (Ts2-FL, Nikon, Japan).

4.17. Cell scratch assay

HGFs were seeded on Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂ at a density of 5.0 × 10⁴ cells/well and cultured for 24 h. Then, the culture medium was replaced with serum-free DMEM (Gibco, USA), and a vertical line was scratched at the center of each sample surface using a 200 μL pipette tip. The samples were collected at 0, 4, and 24 h after scratching, and fixed and blocked as described in section 4.15, and stained with Phalloidin FITC Reagent (1:1000 dilution, Abcam, UK) for 1 h in the dark. Cell nuclei were counterstained with DAPI (Merck, Germany) for 5 min. Cell migration was observed and photographed using a fluorescence microscope (Ts2-FL, Nikon, Japan). ImageJ 1.54d software (NIH, USA) was used to calculate the percentage of the cell migration area relative to the initial scratch area at 0 h.

4.18. qRT-PCR

HGFs were seeded on Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂ in 24-well plates at a density of 2.0 × 10⁴ cells/well and cultured for 1, 2, and 7 days. Then, the samples were gently washed twice with PBS to remove non-adherent cells and transferred to 6-well plates. The cells were then lysed with 1 mL of TRIzol (Invitrogen, USA). RNA was extracted using chloroform and isopropanol and then reverse-transcribed to cDNA using a Takara kit (RR036A, Takara, Japan). Gene primers (Table S1) were purchased from Sangon Biotech (Sangon Biotech, China), and the relative gene expression was measured using a Takara kit (RR820A, Takara, Japan) on Quantstudio 1 (Thermofisher, USA).

4.19. Western blotting

HGFs were seeded on Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂ with a density of 2.0 × 10⁴ cells/well and cultured for 3 days. Then the cells were rinsed three times with phosphate-buffered saline (PBS) and lysed using RIPA lysis buffer (G2002, Servicebio, China) supplemented with protease inhibitors. After incubation at room temperature for 5 min, the cells were gently scraped off with a cell scraper and transferred into 1.5 mL tubes. Lysis continued with ice bath for 30 min. The lysates were then centrifuged at 12,000 rpm for 10 min at 4 °C, and the supernatants were collected. Protein concentrations were determined using a BCA protein assay kit (P0009, Beyotime, China). Equal amounts of protein were mixed with 5× loading buffer (Yoche, Shanghai, China) and denatured in boiling water for 10 min. Proteins were separated by Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis (SDS-PAGE) and subsequently transferred onto PVDF membranes. After blocking, membranes were incubated with primary antibodies: anti-ACTIN (1:1000, Servicebio, China), anti-Fibronectin (1:1000, Servicebio, China), anti-Collagen I (1:1000, Servicebio, China), anti-Integrin β1 (1:1000, Abcam, UK), anti-ADCY3, anti-RAP1A, anti-TGF-β1(1:500, Affinity Biosciences, China) and anti-MMP2 (1:500, Proteintech, China). After washing, membranes were incubated with HRP-conjugated goat anti-rabbit IgG (H + L) secondary antibody (1:3000, Servicebio, China). Then ECL reagents A and B (G2014, Servicebio, China) were mixed at a 1:1 ratio, applied to the membranes for 1 min, and chemiluminescent signals were captured using an imaging system (ChemiDoc 1061 XRS + Imaging System, Bio-Rad, USA). Band intensities were quantified using AIWBwell™ analysis software.

4.20. Transcriptome analysis

HGFs were seeded on sterilized Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂ with a density of 2 × 10⁴ cells per sheet in 24-well plates (a total of 40 sheets per group, N = 3). Total RNA was extracted using TRIzol Reagent (Ambion, Thermo), and RNA purity was assessed using a NanoDrop spectrophotometer (ND-ONE-W, Thermo). RNA concentration was measured with the Qubit RNA BR Assay Kit (Q10210, Invitrogen), while RNA integrity was evaluated using the RNA Screen Tape system (5067-5576, Agilent). Depending on the total RNA yield, 0.1-1 μg of total RNA was used for mRNA isolation with the NEBNext® Poly(A) mRNA Magnetic Isolation Module (E7490L, New England Biolabs). The NEBNext® Ultra™ II mRNA Library Prep Kit for Illumina® (E7770L, New England Biolabs) was employed to construct the mRNA library. Library concentration was determined using the Qubit™ dsDNA HS Assay Kit (Q32851, Invitrogen), and the fragment size distribution of the library was analyzed with the D1000 Screen Tape system (5067-5582, Agilent). The molar concentration of the library was accurately quantified using the KAPA Library Quant Kit (Illumina) with universal qPCR Mix (G3326, Servicebio). Finally, high-throughput sequencing was performed on the NovaSeq 6000 platform (Illumina) using the NovaSeq S4 Reagent Kit (20012866, Illumina). For in-depth bioinformatics analysis, differentially expressed genes (DEGs) were identified, followed by Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis, Gene Ontology (GO) enrichment analysis, DEG clustering analysis, and sample correlation analysis. The relevant dataset (OMIX016311) has been deposited and shared in the China National Center for Bioinformation (www.cncb.ac.cn).

4.21. Intracellular calcium ion concentration

HGFs were seeded onto sterilized Ti, Ce-PIII, TiO₂, and CeO_x/TiO₂ with a density of 1 × 10⁴ cells/well in 24-well plates. After 2 days of culture, the samples were gently washed twice with PBS. A staining solution was prepared using serum-free DMEM containing Hoechst 33342 (5 μg/mL, Adamas, China) and Fluo-4 AM (1 μM, Beyotime-Biotech Co., China). Each well received 500 μL of the staining solution and was incubated for 20 min at room temperature in the dark. Following PBS washes, the samples were observed and imaged using a fluorescence microscope (Ts2-FL, Nikon, Japan). After imaging, 500 μL of PBS was added to each well. The samples were then stored at 4 °C in the dark for 30 min. Surface cells were subsequently scraped off using a cell scraper. From each well, 100 μL of PBS was collected and analyzed using a spectrophotometer (Thermo Scientific, USA).

The relative Ca²⁺ concentration was calculated as follows:

$$\text{Relative}{\text{Ca}}^{2+}\text{concentration}=\text{Fluorescence}\text{intensity}\left(\text{Fluo}-4-{\text{Ca}}^{2+},\text{Ex}=488\text{nm},\text{Em}=520\text{nm}\right)/\text{Fluorescence}\text{intensity}\left(\text{Hoechst}33342,\text{Ex}=350\text{nm},\text{Em}=461\text{nm}\right)\text{.}$$

4.22. Establishment of peri-implantitis mode in beagle dogs

All the animal experiments were approved by the Ethics Committee of Shanghai Xuhui District Stomatological Hospital (approval No. [2024]6). Four 18-month-old male beagle dogs were selected as model animals. In this study, the implantation sites were used as the unit of analysis (a total of 4 implants and abutments were placed in each animal after the extraction of the bilateral mandibular second and third molars, resulting in 16 valid observational samples in total). To control for individual differences, a self-controlled design was employed: pure Ti abutments and CeO_x/TiO₂-modified abutments were implanted at the second and third molar positions, respectively, within the same beagle dog. To avoid positional bias, a randomized crossover design was used for abutment allocation, Ti abutments were placed at the second molar positions and CeO_x/TiO₂-modified abutments at the third molar positions in 2 animals, with the opposite arrangement in the other 2 animals. All surgical procedures were performed by the same experienced operator under sterile conditions using anesthesia. The second and third molars were extracted and the implants were immediately placed. Healing abutments were installed after implantation for three months. A peri-implantitis model was established using a ligature-induced method. Clinical indicators were regularly recorded and radiographs were obtained using an X-ray machine (WoodPecker, China). Upon successful establishment of peri-implantitis model, gingival crevicular fluid (GCF) was collected using 1 × 5 mm filter paper strips. The strips were gently inserted into the buccal, lingual, mesial, and distal areas of the experimental teeth for 30 s. The samples were stored at −80 °C until analysis. Subsequently, Ti and CeO_x/TiO₂-modified abutments were installed and the dogs are fed for another four weeks. The modified abutments were irradiated with a 1064 nm near-infrared laser at a power density of 0.75 W/cm². The irradiation was maintained for 5 min. This protocol was administered once a week for four weeks. During this period, the clinical periodontal index PPD (probing pocket depth) was recorded using a periodontal probe every two weeks. After treatment, Gingival Crevicular Fluid (GCF) was collected, and TNF-α and IL-6 levels in the GCF were quantified using ELISA kits (ECC102a.96 and ECC001.96, NeoBioscience, China) according to the manufacturer's protocols. Finally, the dogs were sacrificed, and the tissues were collected and fixed in 4% paraformaldehyde. Meanwhile, the removed abutments were stained with basic fuchsin for plaque index assessment.

4.23. Radiographic evaluation and micro-CT tomography analysis

The mandibles with implants and abutments were scanned using X-ray camera (AiRay, woodpecker, China) and micro-CT (SkyScan1176, Bruker, USA). The 3D reconstructed images were obtained with NNTViewer software. Bone volume/total volume (BV/TV), trabecular thickness (Tb. Th) and trabecular number (Tb. N) were measured. Micro-CT analyses were conducted independently by investigators blinded to group allocation.

4.24. Histological evaluation

After completing radiographic evaluation and micro-CT tomography analysis, the samples were dehydrated using different concentrations of ethanol and then embedded in polymethylmethacrylate. Finally, sections with a thickness of approximately 40 μm were obtained and stained with Van Gieson's picrofuchsin. The remaining samples were decalcified with a 10% ethylenediaminetetraacetic acid (EDTA) solution and embedded in paraffin. Subsequently, 5-μm-thick sections were obtained and subjected to H&E staining, as well as immunohistochemical staining for COL Ⅰ and immunofluorescence staining for BP180, CD206, iNOS, and VEGF. Images of the stained sections were captured using a fluorescence microscope. Histological evaluations were conducted independently by investigators blinded to group allocation. The results of immunohistochemistry and immunofluorescence staining were quantified using ImageJ software.

4.25. Statistical analysis

All data were analyzed using GraphPad Prism 10.0.0 software with one-way or two-way ANOVA, and the data are presented as mean ± standard deviation (SD) from at least triplicate tests (n ≥ 3). Since the measurements from the second and third molar sites within each dog are correlated (originating from the same animal), a paired design was adopted for comparisons between Ti and CeO_x/TiO₂ modified abutments. For continuous variables (e.g., bone resorption depth, probing depth), paired t-tests were used to compare the differences within the same dog, thereby eliminating within-dog clustering effects. p < 0.05 was considered statistically significant (*p < 0.05, **p < 0.01, and ***p < 0.001).

CRediT authorship contribution statement

Min Xing: Writing – original draft, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Haifeng Zhang: Methodology, Investigation, Funding acquisition, Formal analysis. Tiantian Chen: Methodology. Kuicai Ye: Writing – review & editing, Visualization, Validation, Investigation. Jiayin Feng: Investigation. Wenhao Qian: Supervision, Project administration, Funding acquisition. Xuanyong Liu: Writing – review & editing, Supervision, Funding acquisition. Jiajun Qiu: Writing – review & editing, Supervision, Project administration, Funding acquisition.

Ethics approval and consent to participate

The animal surgical procedures were conducted in accordance with the Guidelines for Care and Use of Laboratory Animals and approved by the Ethics Committee of Shanghai Xuhui District Stomatological Hospital (approval No. [2024]6).

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Data availability

Data will be made available on request.