A stalactite-inspired NIR/pH-responsive visualizable nano-platform for microenvironment-activated therapy of peri-implantitis-induced bone defects

Abstract

Peri-implantitis is an inflammatory disease characterized by progressive peri-implant bone resorption. It poses a substantial challenge to the long-term survival of dental implants and imposes a considerable therapeutic burden. As the early stage of peri-implantitis is typically asymptomatic and difficult to detect, the optimal intervention window is often missed. Current reconstructive approaches primarily rely on invasive debridement or bone grafting, underscoring the need for a multifunctional therapeutic strategy that combines diagnostic capabilities, anti-inflammatory, and osteogenic properties. Inspired by the biomineralization process of stalactites, we developed a biomimetic composite CIMA (ACC@ICG/AZM@ALD) by encapsulating indocyanine green (ICG) and azithromycin (AZM) into amorphous calcium carbonate (ACC) and coating its surface with the bone-targeting ligand alendronate (ALD), thereby imparting the nanoplatform with antibacterial, anti-inflammatory, and osteoinductive properties. The introduction of AZM reshaped the peri-implant immune microenvironment and synergized with ICG, an FDA-approved NIR dye, to enhance antibacterial efficacy. In vitro assays demonstrated that CIMA could responsively trigger near-infrared (NIR) fluorescence upon exposure to the bone defect microenvironment, providing real-time visual warning. Furthermore, CIMA effectively scavenged reactive oxygen species (ROS), alleviated oxidative stress, and exerted potent anti-inflammatory and antibacterial effects against Staphylococcus aureus (S. aureus) and Streptococcus mutans (S. mutans). Concurrently, it promoted the osteogenic differentiation of bone marrow–derived cells while suppressing osteoclastogenesis. In vivo, using a rat peri-implantitis model, CIMA facilitated bone regeneration within the inflamed peri-implant region. Collectively, this artificially engineered, stalactite-inspired CIMA system promoted bone regeneration and provided combined anti-inflammatory and antibacterial functions, offering a promising strategy for treating peri-implantitis.

Graphical abstract

Graph 1. Schematic illustration of the preparation of NIR/pH-responsive visualizable nano-platform for microenvironment-activated therapy of peri-implantitis-induced bone defects, and its therapeutic application for peri-implantitis. A) Establishment of a Sprague–Dawley (SD) rat peri-implantitis bone defect model. In response to a defective microenvironment, CIMA releases therapeutic payloads and enables lesion-confined fluorescence reporting for early visual warning. B) Upon near-infrared (NIR) irradiation, CIMA provides synergistic antibacterial/anti-biofilm decontamination at the peri-implant site. C) Concurrent anti-inflammatory immunomodulation alleviates the local inflammatory burden and supports osteogenic regeneration. D) Schematic illustration of peri-implant bone defect repair after treatment.To validate therapeutic efficacy, we established an intraoral Sprague–Dawley rat peri-implantitis bone defect model following an “immediate extraction–immediate implantation” protocol (Graph 1A). Under neutral conditions, CIMA remains stable and co-encapsulates ICG and AZM; upon peri-implantitis–associated acidification within the defect niche, the ACC component disintegrates to release both agents while switching on ICG fluorescence for lesion-confined, real-time visual early warning (Graph 1A). Subsequent 808 nm NIR irradiation activates ICG to generate singlet oxygen (¹O₂) and mild hyperthermia, thereby achieving synergistic photodynamic/photothermal antibacterial and anti-biofilm decontamination at the peri-implant site (Graph 1B). In parallel, released AZM modulates macrophage polarization to alleviate the inflammatory burden and support osteogenic regeneration (Graph 1C), ultimately facilitating peri-implant defect repair (Graph 1D) through enabling an image-guided, sequential “monitoring–therapy–regeneration” strategy.

Highlights

• •CIMA provides microenvironment-responsive antibacterial activity with early NIR fluorescence warning.
• •CIMA suppresses oxidative stress and key inflammatory pathways to stabilize the peri-implant immune microenvironment.
• •CIMA promotes osteogenesis within inflamed peri-implant bone defects.

1. Introduction

With the global increase in dental implant placement, peri-implantitis has emerged as a progressively prevalent and clinically challenging disease [1]. Although the ideal goal is to achieve long-term implant retention, early peri-implant inflammation is often insidious; radiographic techniques are not sensitive enough to detect minor bone loss and are prone to measurement errors. Moreover, patients typically seek treatment only when symptoms become pronounced, frequently missing the optimal therapeutic window [2]. Treatment at this point typically involves invasive debridement or bone augmentation, and in severe cases, implant removal, resulting in hard- and soft-tissue defects that complicate secondary implant placement [3]. Therefore, an ideal therapeutic strategy for peri-implantitis should possess “three-in-one” functionality: (I) providing visual warning signals at the initial stage of local microenvironmental abnormalities [4]; (II) modulating inflammation in a minimally invasive manner and improving the gingiva–alveolar bone interface [5,6]; and (III) promoting regeneration of peri-implant bone defects to improve long-term implant retention [7]. To address these challenges, we propose a strategy that links the material design directly to the modulation of the peri-implantitis lesion microenvironment. The nanoplatform is engineered to release its payload in response to acidic inflammatory environments associated with peri-implantitis, enabling real-time monitoring of lesion progression via fluorescence release. The acidic microenvironment (pH shift due to local inflammation) serves as the primary trigger to initiate a local material transition that activates therapeutic interventions, such as antibacterial effects and osteogenesis promotion.

This work was inspired by natural stalactites, which form through calcium carbonate precipitation and can disintegrate under acidic conditions. Since peri-implantitis is also characterized by an acidic microenvironment, we designed an acid-responsive, stalactite-inspired material platform based on this pH-dependent mineralization–dissolution behavior. Osteoclasts pump protons into the bone resorption lacuna via H⁺-ATPase and chloride channels at their ruffled border, reducing the local pH to 4.7–6.8 [8,9]. This localized and relatively stable acidic shift can therefore serve as an indicator for monitoring inflammatory alveolar bone resorption [10]. Amorphous calcium carbonate (ACC), a biocompatible and biodegradable biomineral material, is biocompatible, biodegradable, and widely used in drug delivery [11]. Its acid-sensitive property keeps it stable at physiological pH while releasing cargos under acidic conditions. Thus, ACC is a promising platform for the early monitoring and on-demand therapy of peri-implantitis [12]. To improve local bone accumulation and reduce systemic exposure, functionalized ACC with alendronate (ALD), a bisphosphonate ligand with high affinity for bone mineral components, as previously demonstrated in bone-targeting ACC systems [[13], [14], [15]]. For imaging and phototherapy, we incorporated indocyanine green (ICG), an FDA-approved NIR fluorophore. Under 808 nm near-infrared (NIR) irradiation, it can generate both photodynamic (singlet oxygen) and photothermal effects to enhance antibacterial activity [[16], [17], [18]]. For example, Zhang et al. [19] co-encapsulated ICG and rapamycin into liposomes to generate singlet oxygen (¹O₂) and hyperthermia under NIR irradiation, thereby regulating the oral microecology and preventing implant-associated infections. In addition, azithromycin (AZM), an immunomodulatory macrolide, is frequently used for perioperative prophylaxis in implant dentistry and the treatment of periodontitis [20]; compared with amoxicillin, it has shown more substantial potential to suppress inflammatory mediators at peri-implant wound sites and has been reported to inhibit oxidative stress, inflammation, and apoptosis [[21], [22], [23], [24]]. In summary, we integrated acid responsiveness (ACC), bone targeting (ALD), imaging/phototherapy (ICG), and immuno-antibacterial functionality (AZM) to construct a stalactite-inspired, acid-activated nanoplatform, ACC@ICG/AZM@ALD (CIMA) (Fig. 1A).

Fig. 1.

Schematic illustration of the preparation of NIR/pH-responsive visualizable nano-platform for microenvironment-activated therapy of peri-implantitis-induced bone defects, and its therapeutic application for peri-implantitis. A) Schematic representation of the preparation process of CIMA. B) Establishment of an SD rat peri-implantitis bone defect model. C) CIMA responding to the bone defect microenvironment by releasing agents that provide visualized early warning (I) and, upon near-infrared irradiation, achieve synergistic antibacterial (IIa) and anti-inflammatory (IIb) therapy for peri-implantitis. D) Schematic depiction of bone defect restoration before and after treatment (III).

To validate its efficacy, we established an intraoral rat model of peri-implantitis based on an “immediate extraction–immediate implantation” protocol (Fig. 1B) [25]. As illustrated in Fig. 1C, under neutral conditions, ACC stably encapsulates ICG and AZM; when local inflammatory bone destruction induces acidification, ACC disintegrates and releases ICG and AZM, while simultaneously activating ICG's NIR fluorescence to provide a real-time, visual warning. Subsequently, 808 nm NIR irradiation enables ICG to generate singlet oxygen (¹O₂) and raise the temperature, producing a synergistic photodynamic–photothermal antibacterial effect, while AZM further improves the inflammatory microenvironment by modulating macrophage polarization. During material degradation, photodynamic and photothermal antibacterial actions cooperate with AZM-mediated immunoregulation to promote bone regeneration (Fig. 1D), enabling an image-guided, sequential “monitoring–therapy–regeneration” strategy for effective defect repair.

2. Results and discussion

2.1. Synthesis and characterization of CIMA

As shown in Fig. 2A, amorphous calcium carbonate nanoparticles (ACC NPs) were first prepared by a gas-diffusion method, and ICG together with AZM was co-loaded to obtain ACC@ICG/AZM (CIM). Subsequent surface modification with the bisphosphonate ligand ALD yielded the bone-targeting final product, ACC@ICG/AZM@ALD (CIMA). ACC@ICG (CI) was prepared using the same procedure, and its ALD-modified form (CIA) served as a control. Morphological analysis by SEM showed that CIMA particles were 660–800 nm in diameter (Fig. S1). TEM image further confirmed their core–shell structure, and EDS mapping detected C, Ca, Cl, N, O, P, and S in CIMA (Fig. 2B), collectively indicating successful encapsulation of ICG/AZM and surface grafting of the phosphorus-containing ALD ligand. ζ-potential measurements (Fig. 2C) indicated that ACC-based nanoparticles were positively charged at neutral pH, which is attributed to exposed Ca²⁺. After ALD modification, the surface potential shifted negatively. This change provided electrostatic evidence for successful functionalization. FTIR spectra (Fig. 2D) of CI, CIA, CIMA, and AZM revealed the characteristic in-plane and out-of-plane deformation bands of CO₃²⁻ at 712 cm⁻¹ and 872 cm⁻¹ in all three composites (CI/CIA/CIMA) [26], together with the vinyl stretching band of ICG at 1110 cm⁻¹, confirming the successful encapsulation of ICG in ACC. CIA and CIMA showed a P–O-P stretching band at 568 cm⁻¹, indicating successful ALD loading. The characteristic AZM peaks at 1186 and 1082 cm⁻¹ (C=O and symmetric C–O–C stretching) were also present in the CIMA spectrum, suggesting that AZM was effectively incorporated and retained its structural features [27]. In addition, the C=C stretching band of ICG at 1435 cm⁻¹ in the CIMA spectrum provided further evidence for the presence of ICG in CIMA. To evaluate the phase stability of CIMA during storage, X-ray diffraction (XRD) and Fourier transform infrared spectroscopy (FTIR) analyses were performed. XRD (Fig. S2A) showed a dominant amorphous halo; compared with 7 days, the 21-day pattern exhibited unchanged peak positions but slightly sharper and more intense features, suggesting a minor crystallization tendency. FTIR spectra (Fig. S2B) showed overall consistent band positions for the key groups, while the 21-day sample presented reduced transmittance with subtle band-shape changes. Together, these results indicate a limited ACC-to-vaterite conversion during storage, whereas the amorphous phase remains predominant, supporting satisfactory phase stability of CIMA.

Fig. 2.

Preparation and characterization of CIA and CIMA. A) Schematic illustration of the preparation process of CIMA. B) TEM and elemental mapping images of CIMA. C) Zeta potential measurements of CIA and CIMA. D) FTIR spectra of AZM, CI, CIA and CIMA. E) Fluorescence (FL) spectra of CIMA under different pH conditions. F) XPS survey spectra of CIM and CIMA. G–I) High-resolution XPS spectra of CIMA for C 1s, P 2p, and O 1s. J) UV–vis spectra of CIMA at various pH values. K, L) Time-dependent UV–vis spectra of CIMA in buffer solutions at pH 4.5 and 5.5. Data are presented as mean ± standard deviation (n ≥ 3).

X-ray photoelectron spectroscopy (XPS) further validated the chemical composition and surface structure of the CIMA nanoparticles (Fig. 2F). The survey spectrum confirmed the presence of Ca and P. The high-resolution C 1s spectrum displayed peaks at 288.8 eV(O–C=O/CO₃²⁻) and 286.0 eV (C–O/C–N). In the P 2p region, 133.3 eV (P 2p_3/2) and 132.4 eV (P 2p_1/2) peaks were assigned to phosphate (PO₄³⁻) formation introduced by the ALD [28]. Peaks at 347.2 eV (Ca 2p_3/2) and 350.7 eV (Ca 2p_1/2) were attributed to Ca 2p states (Fig. 2I). Collectively, these XPS data verify the successful synthesis of CIMA and its designed “drug-loaded ACC + ALD bone-targeting” structure.

2.2. pH-responsive fluorescence and payload release

CIMA fluorescence spectra were subsequently recorded in a series of pH-buffered media. As shown in Fig. 2E, the maximum excitation wavelength of CIMA was centered at 805 nm, consistent with that of ICG. As the pH increased from 4.5 to 8.5, fluorescence intensity gradually decreased, indicating that CIMA was responsive to pH. Notably, a strong linear correlation between fluorescence intensity and pH was observed within the range of 4.5–6.5, which aligns well with the acidic milieu of bone resorption (pH 4.7–6.8), suggesting that CIMA has potential for bone-resorption imaging [29]. The photostability of CIMA, a critical parameter for in vivo imaging, was evaluated at pH 4.5 and 8.5. As shown in Fig. S3A–C, the fluorescence intensity remained nearly unchanged over 2 h of monitoring, demonstrating excellent stability. To ensure reliable detection under complex in vitro and in vivo conditions, resistance to interference was evaluated. Although ions such as K⁺, Zn²⁺ and abundant Ca²⁺ may fluctuate during bone resorption, CIMA fluorescence was insensitive to these cations but highly responsive to H⁺ (Fig. S3D), indicating superior selectivity for acidic environments and anti-interference performance [30]. UV–vis spectra further corroborated the acid responsiveness: the maximum absorption peak at 780 nm (Fig. 2J) corresponds to the characteristic absorption of ICG, with a maximum excitation wavelength of 805 nm; under acidic conditions (pH 4.5, 5.5, 6.5), the absorbance increased over time (Fig. 2K and L, Fig. S4A), consistent with the fluorescence enhancement. In contrast, under neutral and alkaline conditions (pH 7.5/8.5), ICG release was negligible (Fig. S4C). The release profiles of ALD and AZM showed a similar pH dependence. ALD release, quantified by the ninhydrin colorimetric method, increased over time at pH 5.5 but was minimal at pH 7.5 (Fig. S5A and S5B). AZM exhibited time-dependent release at pH 4.5–6.5, whereas negligible changes were observed at pH 7.5/8.5 (Fig. S6A–F). Altogether, these results demonstrate that CIMA combines acid responsiveness, fluorescence stability, and good selectivity, making it suitable for imaging and therapy in acidic bone-resorption microenvironments.

2.3. NIR responsiveness and ROS homeostasis modulation

The photothermal performance of CIMA was then examined. Under 808 nm NIR irradiation for 120 s (1.0 W/cm²), at 25 μg/mL, the temperatures of CIMA and CIA respectively reached 46.75 ± 0.44 °C and 44.77 ± 0.67 °C (Fig. 3A). To assess thermal stability within a safe range, the laser was turned off for 30 s, during which the temperature dropped to 44.17 ± 0.57 °C; upon re-irradiation for 30 s, it increased to 48.80 ± 1.03 °C [31]. By strictly controlling the laser power and exposure duration, the maximum temperature was maintained below 50 °C to minimize potential thermal injury. After three on/off cycles (30 s on/30 s off), the temperature remained within the range of 42–50 °C (Fig. 3B) [32], and no significant decay was observed after four cycles (Fig. 3C), confirming excellent photothermal stability. To evaluate the photodynamic effect, 1,3-diphenylisobenzofuran (DPBF)—a highly specific fluorescent probe for ¹O₂ that forms an endoperoxide and decomposes to 1,2-dibenzoylbenzene upon reaction—was used (Fig. 3D). UV–vis spectra showed that the DPBF absorption at 410 nm gradually decreased with increasing CIMA concentration, indicating NIR-triggered ¹O₂ generation and subsequent DPBF consumption (Fig. 3E). After 120 s of irradiation, ¹O₂ production led to near-complete DPBF consumption (Fig. 3F).

Fig. 3.

NIR responsiveness and reactive oxygen species (ROS) regulation. A) Temperature changes of CIA and CIMA under 808 nm NIR irradiation (1.0 W/cm²). B) Temperature profile of CIMA after 2 min of heating followed by three on/off laser cycles (30 s on/30 s off). C) Temperature variation of CIMA during four NIR laser on–off cycles. D) Schematic illustration of ¹O₂ generation from CIMA under NIR. E) UV–vis spectra of DPBF after treatment with different concentrations of CIMA. F) UV–vis spectra of DPBF after treatment with CIMA for different time intervals. G) Schematic diagram illustrating the ·OH scavenging mechanism of CIMA. H) ·OH scavenging efficiency of CIMA at various concentrations under NIR irradiation. I) Comparison of ·OH scavenging efficiencies among CIA, CIMA, and CIMA + NIR groups. J) Fluorescence images of CIMA adsorbed onto HAP surfaces in buffer solutions of different pH values. K) Semi-quantitative fluorescence intensity histogram. L) Representative SEM images of CIMA adsorbed on the HAP surface after different incubation times. Data are presented as mean ± standard deviation (n ≥ 3). ∗∗p < 0.01, ∗∗∗∗p < 0.0001, ns: no significance.

Given that the immunomodulatory drug AZM has reported hydroxyl radical (·OH) scavenging activity [33], we investigated the ability of CIMA to eliminate ·OH in an acidic environment mimicking peri-implantitis. UV–vis was used to establish the standard curve of AZM and determine its encapsulation efficiency (Fig. S7) [34]. The formulation with high AZM encapsulation (88.16 ± 1.13%, ICG: AZM = 1:1) was selected to prepare CIMA (Fig. S8). In the salicylic acid (SA)–Fenton assay (Fig. 3G), the ·OH scavenging rate increased with CIMA concentration and reached 78.69 ± 0.43% at 25 μg/mL (Fig. 3H). Comparison of CIA, CIMA, and CIMA + NIR showed that CIA achieved 43.86 ± 1.81% scavenging, while CIMA significantly enhanced this effect to 78.69 ± 0.43%, confirming the contribution of AZM to ·OH elimination (Fig. 3I). No significant difference was observed between CIMA and CIMA + NIR, indicating that NIR irradiation did not significantly affect ·OH scavenging in vitro [35].

Bone-targeting capability is a critical prerequisite for enhancing the therapeutic accumulation of CIMA around implant sites; thus, the in vitro bone-targeting performance was evaluated. ALD contains bisphosphonate groups that selectively chelate Ca²⁺ and bind to the surface of hydroxyapatite (HAP), the main inorganic component of bone [36]. As a result, HAP was used to simulate the in vivo bone microenvironment. CIMA was co-incubated with HAP under stirring, then the fluorescence of free CIMA in the supernatant was recorded at different time points. As shown in Fig. S9, fluorescence decreased over time to nearly the background level, suggesting time-dependent binding of CIMA onto HAP. Under acidic conditions (Fig. 3J and K), stronger fluorescence was observed, which is consistent with the acidic microenvironment of bone resorption and indicates that CIMA retains imaging responsiveness and preferentially accumulates on HAP at lower pH. SEM images of lyophilized HAP–CIMA complexes (Fig. 3L) showed irregular lamellar HAP at 0 min and an increased number of spherical CIMA NPs on the HAP surface over time, collectively verifying the excellent in vitro bone-targeting ability of CIMA. Consistently, compared with ALD-free CIM, CIMA showed more evident deposition on HAP in TEM images at 0, 30 min, and 3 h (Fig. S10). This trend is consistent with the HAP-binding role of ALD.

2.4. In vitro antibacterial performance of CIA and CIMA

Given that peri-implantitis is a bacteria-driven infection, the antibacterial performance of the material is critical for both its therapeutic and prophylactic effects. Early peri-implant biofilms are often dominated by Streptococci [37], whereas subsequent colonization by S. aureus exacerbates inflammation and increases treatment difficulty [38,39]. We therefore evaluated the antibacterial effects of the materials against these two species (S. aureus and S. mutans) using colony-forming unit (CFU) counting. At a concentration of 25 μg/mL, CIA exhibited moderate antibacterial activity against S. aureus (38.57 ± 5.22%) and S. mutans (56.57 ± 8.46%), which was markedly enhanced under NIR irradiation (98.82 ± 0.59% and 94.25 ± 2.40%, respectively). Importantly, AZM-containing CIMA achieved a bactericidal rate of nearly 100% under NIR irradiated and non-irradiated conditions (Fig. 4A and B). SEM images supported these results: control bacteria were intact and smooth; CIA-treated bacteria exhibited partial surface wrinkling (Fig. 4C and D); and CIMA and CIMA + NIR groups displayed more pronounced deformation, with wrinkling and collapse (highlighted regions). Live/dead staining further confirmed that the CIMA + NIR group had the highest bactericidal activity (Fig. 4E and F). To assess the effect on bacterial biofilms, 3-day single-species biofilms were treated with the materials. Crystal violet staining showed that CIMA substantially disrupted biofilm structure; under NIR irradiation (CIMA + NIR group), biofilm clearance rate reached 80.96 ± 0.51% for S. aureus and 74.61 ± 0.38% for S. mutans (Fig. 4G and H). In summary, these results demonstrate that CIMA outperformed CIA, indicating a synergistic enhancement of bactericidal and antibiofilm effects by ICG and AZM.

Fig. 4.

In vitro antibacterial performance of CIA and CIMA nanoparticles (NIR −, non-irradiated; NIR +, irradiated). A) Representative colony images of S. aureus and S. mutans after different treatments. B) Antibacterial rates of different groups. C) SEM images of S. aureus (yellow, intact bacteria; pink, deformed bacteria). D) SEM images of S. mutans (blue, intact bacteria; pink, deformed bacteria). E) Live/dead fluorescence images of bacteria (green, live; red, dead). F) Semi-quantitative analysis of bacterial viability based on live/dead staining. G) Crystal violet staining of single-species bacterial biofilms after treatment. H) Semi-quantitative analysis of biofilm clearance. Data are presented as mean ± standard deviation (n ≥ 3). ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001, ns: no significance.

Biofilms formed on clinically infected implants pose a greater challenge due to the intricate composition of bacteria and EPS components. To further assess the clinical applicability of the system, severely infected dental implants retrieved from the clinical setting were used to cultivate multispecies peri-implantitis biofilms ex vivo (Fig. S11A). Crystal violet staining demonstrated that CIMA under NIR irradiation significantly reduced biofilm biomass, with approximately a 90% decrease compared to other groups (Fig. S11B–C). CLSM with live/dead staining and 3D reconstruction revealed a marked reduction in viable bacteria, increased bacterial killing, and a thinner biofilm after CIMA + NIR treatment (Fig. S11D–E). These results underscore the significant anti-biofilm efficacy of CIMA under NIR irradiation, even in complex, clinically derived multispecies biofilm environments.

2.5. In vitro biocompatibility of CIMA

Good biocompatibility is a prerequisite for in vivo use. Thus, the cytocompatibility of CIMA was evaluated using MC3T3-E1 and RAW 264.7 cell lines via CCK-8 and live/dead staining. CCK-8 assays showed that cell viability remained at or above 80% after 1, 3, and 5 days of co-incubation with CIMA at concentrations up to 25 μg/mL (Fig. S12), indicating no apparent cytotoxicity within this concentration range. Live/dead staining further showed that CIMA, whether with or without NIR, did not adversely affect cell morphology or viability (Fig. S13). Hemolysis tests revealed CIMA hemolysis rates <5% at all tested concentrations (Fig. S14), confirming its high blood compatibility. These results collectively demonstrate that CIMA exhibits favorable in vitro cytocompatibility and hemocompatibility, supporting its further in vivo application.

2.6. ROS regulation and immunomodulation of CIMA

Intracellular singlet oxygen (¹O₂) generation was investigated using DPBF as a fluorescent probe in RAW 264.7 cells. As shown in Fig. 5A, under NIR irradiation, both CIMA- and CIA-treated RAW 264.7 cells showed a significant decrease in DPBF fluorescence, indicating NIR-triggered ¹O₂ generation at the cellular level. Intracellular ROS levels were then assessed with 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA) probe assay (Fig. 5B). LPS-stimulated cells (control) displayed strong green fluorescence, reflecting high oxidative stress. While CIA treatment reduced this signal, CIMA induced the most substantial decrease, demonstrating superior ROS-scavenging capacity that aligns with the in vitro ·OH assay results.

Fig. 5.

ROS regulation and immunomodulatory effects of CIA and CIMA on RAW 264.7 macrophages. A) DPBF fluorescence staining after co-culture with different samples under NIR irradiation. B) DCFH-DA staining showing intracellular ROS levels after LPS stimulation and different treatments. C) Flow cytometry analysis of M1 marker (CD86) and M2 marker (CD206) expression in LPS-pretreated RAW 264.7 cells. D) Representative immunofluorescence images of CD86 (M1) and CD206 (M2) expression after different treatments. E–H) Quantitative flow cytometry results for CD86⁺ and CD206⁺ macrophages. I–L) ELISA analysis of IL-6 and TNF-α secretion after LPS stimulation and different treatments. Data are presented as mean ± standard deviation (n ≥ 3). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

To assess macrophage polarization, RAW 264.7 cells were pretreated with PBS (Control), LPS, LPS + CIA, or LPS + CIMA. LPS-induced oxidative-stress-related inflammation, as well as the M1 macrophage marker (CD86) and M2 macrophage marker (CD206) markers, were analyzed using flow cytometry and immunofluorescence [40]. Flow cytometry revealed that LPS stimulation significantly increased the population of CD86⁺ M1 macrophages, which was reduced by both CIA and CIMA treatments, with CIMA being more effective (Fig. 5C–E–H). Specifically, in Fig. 5G, the overall proportion of CD86⁺ M1 cells was slightly lower compared to Fig. 5E, while the overall proportion of CD206⁺ M2 cells in Fig. 5H was lower than in Fig. 5F. These slight differences likely reflect minor technical variations, such as differences in cell passage or minor fluctuations in fluorescence intensity, which can influence the sensitivity of phenotypic responses. Despite these variations, the relative trends for M1 and M2 macrophages within each group remained consistent. Conversely, the proportion of CD206⁺ M2 macrophages showed an opposite trend. Conversely, the proportion of CD206⁺ M2 macrophages showed an opposite trend. Immunofluorescence confirmed that CIMA reduced CD86 and enhanced CD206 compared with LPS (Fig. 5D). ELISA (Fig. 5I–L) further showed that CIMA and CIA suppressed LPS-induced overexpression of pro-inflammatory cytokines IL-6 and TNF-α. Notably, under LPS stimulation, the NIR-irradiated groups showed a slight increase in inflammatory cytokines, which may be attributable to photothermal or photodynamic effects on inflammation-related signaling pathways (Fig. 5I–L) [41]. To provide pathway-level evidence, an NF-κB inhibitor was applied, and p65 nuclear translocation and IKKβ signaling were assessed by immunofluorescence (Fig. S15A and S15B). Inhibitor treatment reduced p65 nuclear localization and weakened IKKβ fluorescence, and CIMA produced a comparable decrease. Collectively, these results indicate that CIA can alleviate LPS-induced oxidative stress and modulate macrophage phenotype, while AZM-loaded CIMA shows more effective intracellular ROS elimination and M2 polarization, underscoring the critical role of AZM in modulating the inflammatory macrophage microenvironment.

2.7. In vitro osteogenic evaluation

Given the immunomodulatory and osteogenic potential effects of AZM, MC3T3-E1 cells were used to evaluate the impact of CIA and CIMA on adhesion and osteogenic differentiation. After 24 h of co-culture, cytoskeletal staining showed more pronounced actin organization and greater spreading in the CIA and CIMA groups than in the control group, with no significant influence from NIR irradiation (Fig. 6B). After 7 days, ALP staining (Fig. 6A), and after 21 days, Alizarin Red S staining results (Fig. 6C–S17), revealed that CIMA and CIMA + NIR induced higher ALP expression and more mineralized nodules than the control and CIA groups. Consistently, compared with ALD-free CIM, CIMA showed stronger ALP staining after 7 days (Fig. S16), indicating that ALD functionalization may contribute to osteogenic differentiation in vitro. qRT-PCR at day 7 (Fig. 6D and E) showed significant upregulation of early osteogenic genes (RUNX-2, BMP-2) in the CIMA group, indicating initiation of osteogenic differentiation [42]. Compared with CIMA alone, CIMA + NIR showed no additional increase in osteogenic gene expression and slightly lower signals in some mineralization-related assays (Fig. 6C–G). This observation aligns with prior reports indicating that, under in vitro conditions lacking inflammatory cues or bacterial challenge, NIR exposure can impose a mild physiological perturbation on cells, which may in turn modulate the expression of selected osteogenesis-related genes [43]. In contrast, data from day 21 showed upregulation of late osteogenic markers (COL-1, OCN), indicating enhanced mineralization [44]. The temporal expression of biomarkers generally aligns with the process of osteogenic differentiation, from early induction to late-stage maturation, confirming the osteogenic potential of CIMA materials.

Fig. 6.

Effects of CIA and CIMA nanoparticles on MC3T3-E1 cell adhesion/osteogenic differentiation and on osteoclast induction (NIR −, non-irradiated; NIR +, irradiated). A) ALP staining after 7 days of co-culture with MC3T3-E1 cells. B) Cytoskeleton staining of MC3T3-E1 cells after 24 h of co-culture. C) Alizarin Red S staining of mineralized nodules after 21 days of co-culture. D–G) qRT-PCR analysis of osteogenic genes (RUNX-2, BMP-2, COL-1, OCN) at different time points. H) Schematic workflow of osteoclast induction and SEM images of multinucleated osteoclasts on bovine cortical bone slices. I) Quantification of TRAP-positive cells in different groups. J) SEM images of mature osteoclasts and resorption pits on bone slices. K) Confocal fluorescence images showing NIR-range signals at CIA- or CIMA-modified bone surfaces cultured with osteoclasts, indicating pH-responsive fluorescence activation during bone resorption. Data are presented as mean ± standard deviation (n ≥ 3). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001, ns: no significance.

2.8. In vitro osteoclast induction and functional assessment

To complement the osteogenic findings, we evaluated the effect of CIA and CIMA on osteoclast. RAW 264.7 cells induced with M-CSF + RANKL for 7 days formed large multinucleated osteoclasts with vigorous TRAP activity [45]. SEM images revealed large, irregular, multinucleated cells with abundant pseudopodia and localized resorption pits on bone slices (Fig. 6H), confirming the successful establishment of an in vitro bone resorption model. TRAP staining and quantification (Fig. 6I and J) showed fewer osteoclasts in CIA and CIA + NIR groups than in the control, consistent with the known antiresorptive effect of ALD. Notably, this inhibitory effect was significantly stronger in the CIMA and CIMA + NIR groups, suggesting that AZM further suppresses osteoclastic activity. To verify pH-responsive monitoring during bone resorption, pre-incubated bone slices (CIA or CIMA) with induced osteoclasts were imaged by confocal microscopy. Strong NIR fluorescence was specifically observed at the osteoclast–bone interface (white arrows, Fig. 6K) but not in RAW 264.7 cultures alone, indicating that the material selectively released fluorescent probes at active resorption sites and emitted NIR signals under irradiation, visualizing ongoing bone resorption. Thus, CIMA can both monitor and therapeutically suppress bone resorption. In this study, CIMA treatment shifted macrophage polarization towards a repair-associated phenotype, accompanied by the downregulation of pro-inflammatory cytokines. This alteration in the local cytokine milieu facilitated osteogenic differentiation while also inhibiting osteoclastogenesis. The interplay between immune modulation and bone remodeling suggests that CIMA's immunoregulatory effects simultaneously promote bone regeneration by skewing macrophages toward a reparative phenotype and suppressing osteoclast activity, thus creating a more favorable microenvironment for osteogenesis. Taken together, the in vitro decontamination and immunomodulation assays support pairing lesion-local monitoring and decontamination with immunoregulation to improve the regenerative microenvironment in implant-associated inflammatory bone defects.

2.9. Therapeutic effects of CIMA in a rat alveolar peri-implantitis model

Based on the in vitro antibacterial, immunomodulatory, and osteogenic assays, we next evaluated CIMA in vivo using a Sprague–Dawley (SD) rat post-extraction, implant-associated peri-implantitis model (Fig. 7A, Fig. S18). Rats were divided into four groups: Control (implant only), PI (peri-implantitis), PI + CIMA, and PI + CIMA + NIR.

Fig. 7.

Therapeutic effects of CIMA nanoparticles in a rat peri-implant alveolar bone model of peri-implantitis. A) Schematic timeline of the in vivo experiment, including model establishment, drug administration, and sample collection. B) Micro-CT images of the implants and surrounding maxillary alveolar bone in each group. C) Quantitative analysis of BV/TV, BS/BV, Tb.Th, and Tb.Sp after 4 weeks of treatment. D) H&E staining of peri-implant alveolar bone in rats. E) Masson's trichrome staining of peri-implant tissues showing collagen deposition. F) Methylene blue/acid fuchsin staining of undecalcified peri-implant bone sections showing direct bone–implant contact. G) Representative agar plate images for in vivo antibacterial evaluation of implants from different groups. H) Colony-counting statistics of explanted implants. I) NIR in vivo fluorescence imaging of rats with peri-implantitis at different time points. J) Semi-quantitative analysis of fluorescence intensity/area. Data are presented as mean ± standard deviation (n ≥ 3). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001, ns: no significance.

Promoting osseointegration is crucial for the long-term stability of bone implants. Micro-CT and histology confirmed that NIR-activated CIMA effectively intervened in early peri-implantitis [46]. 3D reconstructions (Fig. 7B) showed extensive alveolar bone loss in the PI group compared with the Control, confirming successful model establishment. Bone loss was attenuated in PI + CIMA and was markedly improved in PI + CIMA + NIR. Quantitatively (Fig. 7C), PI rats showed decreased bone volume/tissue volume (BV/TV) and trabecular thickness (Tb.Th) and increased bone Surface/bone Volume (BS/BV) and trabecular separation (Tb.Sp). In contrast, the PI + CIMA + NIR group showed a notable recovery of these parameters toward normal levels, indicating effective osteogenic effects under NIR irradiation. PI + CIMA also showed some osteogenesis, attributable to the intrinsic antibacterial, immunomodulatory, and osteoinductive effects of CIMA, with NIR irradiation providing a synergistic boost.

Inflammatory reactions in the peri-implant tissues were assessed by hematoxylin and eosin (H&E) staining. H&E staining (Fig. 7D) showed vascular proliferation and dilation, erythrocyte extravasation, and dense inflammatory infiltrates (plasma cells, lymphocytes) in the PI group, together with reduced osteocyte area, whereas PI + CIMA and PI + CIMA + NIR exhibited reduced inflammation and increased fibroblasts, consistent with CIMA's antibacterial and immunomodulatory functions. Masson staining results (Fig. 7E) revealed decreased collagen in PI compared with Control, while PI + CIMA and PI + CIMA + NIR exhibited greater collagen deposition around implants, with PI + CIMA + NIR being the highest (Fig. S19). Additionally, TRAP staining confirmed fewer osteoclasts in the PI + CIMA + NIR group compared to the PI group (Fig. S20A).

Undecalcified MB–AF staining (Fig. 7F) showed that Control and PI + CIMA + NIR had continuous bone tightly surrounding the implant threads, and high-magnification views further showed direct contact between newly formed bone (red-stained) and the implant surface, indicating successful osseointegration [47]. In contrast, the PI group exhibited a discontinuous bone-implant interface, where a blue-stained fibrous tissue layer was interposed between the implant and the bone, indicating a failure of osseointegration.

To assess in vivo antibacterial efficacy, we retrieved implants at day 7, then collected and cultured bacteria. Plate counting (Fig. 7G and H) showed significantly fewer colonies in PI + CIMA + NIR than in PI, confirming effective NIR-activated eradication of colonized peri-implant bacteria. Giemsa staining (Fig. S20B and S20C) was consistent: abundant bacteria in PI, moderate in PI + CIMA, and markedly reduced in Control and PI + CIMA + NIR.

For in vivo dynamic monitoring, small-animal fluorescence imaging was performed on days 1, 7, and 28 (Fig. 7I). At days 1 and 7, PI + CIMA showed weaker fluorescence, suggesting that mild inflammation was insufficient to trigger substantial CIMA disintegration and ICG release. By day 28, PI + CIMA exhibited the strongest fluorescence, whereas the non-inflamed group, although injected with CIMA, showed weak signals. Semi-quantitative analysis confirmed significant differences (Fig. 7J). To examine whether the fluorescence response varies with bacterial burden in a complex microbial setting, mixed bacterial suspensions were incubated with CIMA for 10 min and imaged using an NIR-I system (Fig. S21A). Fluorescence intensity increased with bacterial concentration (10°–10¹⁰ CFU/mL) (Fig. S21B). Confocal 3D imaging of ex vivo maxilla at day 7 (Fig. S22) revealed more extensive fluorescence distribution in the inflamed group than in the non-inflamed group, demonstrating that CIMA can respond to the low pH microenvironment of bone defects by inducing greater ICG release, making it an ideal candidate material for peri-implantitis monitoring.

2.10. In vivo immunomodulatory and osteogenic mechanisms of CIMA

To further elucidate the mechanism, peri-implant alveolar bone tissues were collected at the 4th week after treatment for RNA-seq (Fig. 8A). A total of 1739 DEGs were identified, including 732 upregulated and 1007 downregulated genes (Fig. 8B), consistent with clustering and correlation heatmaps (Fig. 8C–S23). GO enrichment showed that DEGs were mainly involved in immune and inflammation-related processes (Fig. 8D). Encyclopedia of Genes and Genomes (KEGG) analysis indicated that CIMA significantly affected inflammation-related pathways such as NF-κB and TNF (Fig. 8E, Fig. S24). Principal Component Analysis (PCA) demonstrated distinct transcriptomic profiles between CIMA-treated and control groups (Fig. 8F). GSEA revealed that multiple inflammatory pathways, including NF-κB (ES –1.52), MAPK (−1.66), PI3K-Akt (−1.67), and TNF-β (−1.72), were significantly downregulated in the CIMA group (Fig. 8G and H, Fig. S25), suggesting that CIMA alleviates peri-implant inflammation by coordinately suppressing several key inflammatory cascades, while simultaneously creating a favorable niche for osteogenesis. Western blotting (Fig. S26) confirmed downregulation of p65, the active subunit of NF-κB [48], together with reduced IL-6 [49] and Caspase-3 [50]. Concurrently, the upregulation of RUNX-2 was observed, confirming the promotion of osteogenic differentiation. Thus, CIMA orchestrates a dual mechanism, mitigating inflammation while actively fostering bone formation.

Fig. 8.

In vivo anti-inflammatory and osteogenic molecular mechanisms of CIMA nanoparticles. A) Schematic illustration of RNA sequencing of peri-implant alveolar bone tissues. B) Volcano plot of differentially expressed genes (DEGs) showing upregulated and downregulated genes. C) Principal component analysis (PCA) of transcriptomic profiles. D) Gene Ontology (GO) enrichment analysis of DEGs. E) KEGG pathway enrichment analysis highlighting inflammation-related signaling pathways. F) Heatmap of representative DEGs. G–H) Gene set enrichment analysis (GSEA) showing downregulation of NF-κB, MAPK, PI3K-Akt, and TNF-related pathways after CIMA treatment. Data are presented as mean ± standard deviation (n ≥ 3).

To verify the in vivo immunoregulatory mechanism of CIMA, immunofluorescence staining was performed on peri-implant alveolar bone tissues to analyze the expression of P65 and macrophage phenotypic markers. Compared with the model group, the nuclear fluorescence intensity of p65 was markedly decreased in the CIMA-treated group and even weaker in the CIMA + NIR group (Fig. 9A and C), confirming effective inhibition of NF-κB pathway activation. Given the pivotal role of macrophages in immune modulation, we next evaluated the influence of CIMA on macrophage polarization in vivo. Immunofluorescence staining was used to identify CD86⁺ M1 (pro-inflammatory) and CD206⁺ M2 (anti-inflammatory) macrophages in peri-implant regions (Fig. 9B). Quantitative analysis showed that, compared with the PI group, the proportions of CD86⁺ M1 macrophages were significantly reduced to 12.30 ± 0.45% and 12.65 ± 1.26% in the CIMA and CIMA + NIR groups, respectively (Fig. 9D). Conversely, the proportions of CD206⁺ M2 macrophages increased to 37.16 ± 8.10% and 49.88 ± 6.23% in the same groups (Fig. 9E). These findings are consistent with the in vitro macrophage polarization results, demonstrating that CIMA effectively modulates immune homeostasis in vivo by suppressing NF-κB-mediated pro-inflammatory signaling and promoting M2 macrophage polarization, thereby fostering a regenerative microenvironment favorable for bone repair.

Fig. 9.

In vivo immunomodulatory capacity of CIMA nanoparticles. A) Representative immunofluorescence images of p65 expression in peri-implant alveolar bone tissues. B) Representative immunofluorescence images of CD86⁺ M1 macrophages (red) and CD206⁺ M2 macrophages (green) around the implants. C) Quantitative analysis of the ratio of p65-positive area to total tissue area. D) Quantitative analysis of the ratio of CD86⁺ M1 macrophages to total area. E) Quantitative analysis of the ratio of CD206⁺ M2 macrophages to total area. Data are presented as mean ± standard deviation (n ≥ 3). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001, ns: no significance.

2.11. In vivo biosafety evaluation

The clinical application of novel nanomedicines necessitates evaluation of their biosafety and toxicity. Therefore, a systematic assessment of CIMA's biocompatibility was conducted. Previous in vitro assays had already demonstrated its favorable cytocompatibility. In vivo, histological examination of major organs revealed no pathological changes (Fig. S27), indicating that neither CIA nor CIMA induced injury to the heart, liver, spleen, lungs, or kidneys. Serum biochemical analyses of hepatic and renal function confirmed biosafety, with no significant differences among groups (Fig. S28). All results demonstrate that CIMA possesses excellent in vivo biocompatibility and biosafety.

3. Conclusion

To address the therapeutic challenge of peri-implantitis–associated inflammatory bone defects, drawing inspiration from stalactites, we developed a stalactite-inspired, NIR/pH-responsive and visually activatable nanoplatform (CIMA). This system features three main innovations: (I) CIMA shows antibacterial and antibiofilm activity and provides lesion-associated early visualized warning signals in response to local microenvironmental abnormalities, thereby supporting lesion-localized management; (II) CIMA attenuates oxidative-stress–related responses and modulates inflammation-related signaling implicated in peri-implantitis pathogenesis; and (III) CIMA supports osteogenic recovery under inflammatory conditions, consistent with improved peri-implant repair outcomes. Nonetheless, several limitations remain: the spectrum of tested bacterial strains needs to be expanded; direct in vivo comparisons between CIMA and CIA require further exploration; and tissue-level molecular mechanisms warrant deeper investigation. In addition, the LPS-induced animal model used in this study cannot fully recapitulate the biofilm-contaminated milieu that characterizes clinical peri-implantitis. Given its integrated monitoring, therapy, and regeneration capabilities, CIMA may hold broad translational potential in oral implantology, orthopedics, and other implant-related fields. In future work, we will further explore simplified and scalable fabrication strategies to facilitate the clinical translation of such multifunctional platforms.

4. Materials and methods

4.1. Materials

Ammonium bicarbonate (NH₄HCO₃), anhydrous calcium chloride (CaCl₂), sodium alendronate trihydrate (Alendronate), and salicylic acid (C₇H₆O₃) were purchased from Macklin Biochemical Co., Ltd. (Shanghai, China). Indocyanine green (ICG) and 1,3-diphenylisobenzofuran (DPBF) were obtained from Merck (Germany). Absolute ethanol and methanol were supplied by Hengmao Chemical Reagent Co., Ltd. (Tianjin, China). NF-κB inhibitor was purchased from MedChemExpress (MCE) (Monmouth Junction, NJ, USA). The alkaline phosphatase (ALP) assay kit and bacterial live/dead staining kit (DAMO/PI) were purchased from Beyotime Biotechnology (Shanghai, China). Azithromycin (C₃₈H₇₂N₂O₁₂), crystal violet (CV), Alizarin Red S (ARS) staining solution, lipopolysaccharide (LPS), FITC-labeled phalloidin, 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA), and the tartrate-resistant acid phosphatase (TRAP) assay kit were obtained from Solarbio Life Sciences (Beijing, China). Fetal bovine serum (FBS) was purchased from PAN Biotech (Aidenbach, Germany). The Cell Counting Kit-8 (CCK-8) was obtained from APExBIO (Houston, USA). The modified Giemsa staining kit was purchased from Servicebio Biotechnology Co., Ltd. (Wuhan, China).

4.2. Synthesis of CIMA

Azithromycin (3 mg) was dissolved in 50 mL of anhydrous ethanol, followed by the addition of 3 mg of indocyanine green (ICG) and 0.3 g of anhydrous calcium chloride (CaCl₂) under stirring until complete dissolution [12]. The resulting solution was transferred into a flask covered with aluminum foil (to protect from light) and punctured with five small holes, then placed in an environment containing ammonium bicarbonate for vapor diffusion at 60 °C for 24 h. The obtained product was collected by centrifugation, washed with anhydrous ethanol, and freeze-dried to yield ACC@ICG/AZM (abbreviated as CIM) nanoparticles. The synthesized CIM particles were then dispersed in deionized water, and a sodium alendronate aqueous solution (20 mM) was prepared separately. The alendronate solution was slowly added dropwise to the CIM dispersion, after which the pH was adjusted to 8.0, and the mixture was stirred at 25 °C for 4 h [51]. The resulting product was collected by centrifugation, washed with deionized water, and freeze-dried to obtain ACC@ICG/AZM@ALD (abbreviated as CIMA). Using the same method, nanoparticles loaded with ICG alone (ACC@ICG, CIA) were also prepared.

4.3. Characterization

The surface morphology of the samples was observed using a scanning electron microscope (SEM, Sigma 300, Zeiss, Germany), and their morphology and elemental distribution were further analyzed using a transmission electron microscope (TEM, HT7800, Hitachi, Japan) equipped with an energy-dispersive X-ray spectrometer (EDS). The ζ potential was measured with a zeta potential analyzer (NanoBrook, Brookhaven, USA). The chemical structure and composition of the samples were characterized using a Fourier transform infrared spectrometer (FTIR, Nicolet 5700, Thermo Fisher Scientific, USA) and an X-ray photoelectron spectrometer (XPS, ESCALAB 250Xi, Thermo Fisher Scientific, USA). Fluorescence spectra at different pH values (4.5, 5.5, 6.5, 7.5, and 8.5) were recorded using a fluorescence spectrophotometer (FL970, Techcomp, China), with both excitation and emission slit widths set at 5 nm and an excitation wavelength of 780 nm. Fluorescence stability was evaluated by monitoring the fluorescence intensity changes at pH 4.5 and 8.5 over time. To assess anti-interference capacity, fluorescence intensities were measured in sample solutions containing NaCl (1 mM), CaCl₂ (1 mM), KCl (1 mM), FeCl₃ (200 μM), CuCl₂ (200 μM), MgCl₂ (200 μM), AlCl₃ (200 μM), and ZnCl₂ (200 μM) (n ≥ 3). UV-visible (UV-vis) spectra at different pH values (4.5, 5.5, 6.5, 7.5, and 8.5) were obtained using a UV–vis spectrophotometer (UV–2600, Shimadzu Corporation, Kyoto, Japan) to monitor changes in absorption over time.

4.4. Photothermal effect

The photothermal performance of CIMA was evaluated under an 808 nm near-infrared (NIR) laser irradiation. Suspensions of CIMA and CIA (pretreated in a pH 5.5 phosphate-buffered saline, PBS) were placed in EP tubes and irradiated with an 808 nm laser at a power density of 1.0 W/cm². The temperature changes of the samples were recorded using an infrared thermal imaging camera (FLIR A35, China).

4.5. Singlet oxygen generation and hydroxyl radical scavenging activity

To achieve optimal drug encapsulation, CIMA was synthesized with different mass ratios of ICG to AZM (1:0.5, 1:1, 1:2, and 1:3). UV–vis spectroscopy was used to obtain the absorption spectra of ICG and AZM in ethanol, and standard calibration curves were established to determine the encapsulation efficiency (EE). The EE was calculated according to the following formula:

$$EE(\%)=1-\frac{\text{Drug}\text{content}\text{in}\text{the}\text{supernatant}}{\text{Drug}\text{content}\text{of}\text{total}\text{injected}}\times 100\%$$

1,3-Diphenylisobenzofuran (DPBF) was used as a fluorescent probe to detect the generation of singlet oxygen (¹O₂). DPBF reacts selectively with ¹O₂ to form an endoperoxide, which subsequently decomposes into 1,2-dibenzoylbenzene. Samples with 12.5, 25, 50 μg/mL concentrations (pretreated in a pH 5.5 phosphate-buffered saline, PBS) were mixed with DPBF solution (10 mM) and irradiated with an 808 nm laser at 1.0 W/cm² for 5 min. After irradiation, the mixtures were centrifuged to collect the supernatants, and the absorbance of DPBF at 410 nm was recorded by UV–vis spectroscopy to quantify ¹O₂ production.

The hydroxyl radical (·OH) scavenging capability of the samples was evaluated using salicylic acid as a probe. Samples at different concentrations were mixed with salicylic acid (SA) solution (10 mM) and irradiated with an 808 nm laser at a power density of 1.0 W/cm² for 5 min. The optical density (OD) at 510 nm was then measured using a microplate reader (VICTOR Nivo 3S, PerkinElmer, USA).

4.6. In vitro bone-targeting ability of CIMA

CIMA was incubated with hydroxyapatite (HAP) under light-protected conditions at different pH values (4.5, 5.5, 6.5, 7.5, and 8.5), and the fluorescence intensity and UV–vis absorption of the supernatant were continuously recorded to monitor the adsorption process. After incubation for predetermined time intervals, the HAP samples were observed using a fluorescence microscope (DM i8, Leica Microsystems CMS GmbH, Germany), and the morphology of the HAP surface was examined by scanning electron microscopy (SEM, Sigma 300, Zeiss, Germany).

4.7. In vitro antibacterial assay

To evaluate the antibacterial activity of the materials under near-infrared (NIR) irradiation, six groups were designed: Control, CIA, CIMA, Control + NIR, CIA + NIR, and CIMA + NIR. S. aureus and S. mutans were selected as representative peri-implant pathogens to assess the antibacterial efficacy of the samples. The NIR irradiation was applied in a cyclic mode (808 nm, 1.0 W/cm²; initial heating for 2 min followed by 30 s on/30 s off cycles) for a total of 5 min. Each sample was added to 5 mL of broth medium containing S. aureus or S. mutans, immediately subjected to NIR treatment as described above, and then incubated on a shaker for 6 h. After incubation, the bacterial suspensions were serially diluted and spread onto agar plates. The antibacterial performance of each group was quantified by colony-forming unit (CFU) counting according to the following formula:

$$Antibacterialratio(\%)=\left(N_0‐N_1\right)/N_0\times 100$$

where N₀ is CFU count in the control group and N₁ is the number of bacteria in the experimental group. Bacterial suspensions from each group were washed three times with PBS, fixed with 2.5% glutaraldehyde for 2 h, and washed again three times with PBS. The samples were then dehydrated in a graded ethanol series (30%, 50%, 70%, 80%, 90%, and 100% v/v) for 15 min at each step, followed by centrifugation. Finally, the morphology of the bacteria was observed by SEM.

Bacterial viability was further evaluated using a live/dead bacterial staining kit. After staining for 15 min, the samples were observed under a fluorescence microscope (DM i8, Leica Microsystems CMS GmbH, Germany), where green fluorescence indicated all bacteria and red fluorescence indicated dead bacteria.

In vitro antibacterial biofilm assay of CIMA: Biofilms of S. aureus and S. mutans were established by culturing 1 mL of bacterial suspension (10⁶ CFU/mL) in 12-well plates. After biofilm formation, the samples were applied for treatment, and groups receiving laser irradiation were exposed to an 808 nm laser at 1.0 W/cm² in a cyclic mode. Following 6 h of treatment, the biofilms were stained with crystal violet for subsequent analysis. Specifically, the biofilms were gently washed with PBS, fixed with 2.5% glutaraldehyde for 1 h, stained with 0.1% crystal violet solution for 30 min, and carefully rinsed several times with PBS. The morphology and coverage of the biofilms were observed under an optical microscope, and the staining results were quantitatively analyzed using ImageJ software.

Ex vivo multispecies biofilm assay using clinically infected implants: Severely infected dental implants retrieved from the clinical setting were immersed in BHI and shaken for 30 min to collect a multispecies bacterial suspension. The suspension was incubated under anaerobic conditions (10.5% H₂, 10.5% CO₂, 79% N₂) for 48 h to form multispecies biofilms. After biofilm formation, samples were treated with the indicated materials, and groups receiving irradiation were exposed to an 808 nm laser (1.0 W/cm²) in cyclic mode (initial heating for 2 min followed by 30 s on/30 s off cycles; total 5 min), consistent with the in vitro antibacterial protocol. To evaluate the impact of bacterial load on the fluorescence response, mixed bacterial suspensions at different concentrations (10°–10¹⁰ CFU/mL) were incubated with CIMA for 10 min, and fluorescence intensity was measured using a NIR-I small animal imaging system. Biofilm biomass was evaluated by crystal violet staining and quantified by ImageJ. Biofilm viability was further assessed by live/dead staining and CLSM with 3D reconstruction. This study was approved by the Ethics Committee of the Affiliated Stomatological Hospital of Nanchang University (Approval No. 2025-072).

4.8. Cell culture

Murine macrophages (RAW 264.7 cells) and the murine preosteoblastic cell line MC3T3-E1 were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 1% (v/v) penicillin–streptomycin (KeyGen Biotech, China) and 10% (v/v) fetal bovine serum (Abcell, China). Cells were maintained under standard culture conditions (37 °C, 5% CO₂) in a humidified incubator. For subsequent in vitro experiments, cells were divided into six groups according to treatment and NIR exposure: control (C), CIA, CIMA, control + NIR (C + NIR), CIA + NIR, and CIMA + NIR.

4.9. Assessment of cytocompatibility

RAW 264.7 and MC3T3-E1 cells were seeded into 96-well plates at a density of 5.0 × 10³ cells per well and incubated for 24 h to allow adhesion. The cells were then washed with phosphate-buffered saline (PBS, pH 7.4) and replaced with fresh culture medium (100 μL/well) containing different concentrations of the respective materials. After incubation for 1, 3, and 5 days, the cells were washed twice with PBS and subjected to CCK-8 assays (Solarbio, China) according to the manufacturer's instructions. The absorbance at 450 nm was measured using a multimode microplate reader (VICTOR Nivo 3S, PerkinElmer, USA). Cell viability was calculated using the following formula:

$$Cellviability(\%)=\frac{A-A_{\text{blank}}}{A_{\text{control}}-A_{\text{blank}}}\times 100\%$$

A is the absorbance of the cells treated with materials, A_control is the absorbance of the cells treated with PBS, and A_blank is the absorbance of the medium containing CCK-8 without cells.

4.10. Live/dead cell staining

After co-culturing RAW 264.7 and MC3T3-E1 cells with the materials for 24 h, live/dead staining was performed using a Calcein-AM/propidium iodide (PI) double-staining kit (BIOPRIMACY, China). Cells were incubated with Calcein-AM/PI in the dark for 15 min, and fluorescence images were captured using a fluorescence microscope (DM i8, Leica Microsystems CMS GmbH, Germany).

4.11. Hemolysis assay

Fresh rats blood samples were centrifuged to obtain red blood cells (RBCs), which were then diluted by adding 1 mL of RBCs to 3.67 mL of physiological saline. The diluted RBC suspension (100 μL) was added to 1 mL of saline containing different concentrations of CIMA. For the laser irradiation group, samples were exposed to an 808 nm laser at 1.0 W/cm² in a cyclic mode. The mixtures were incubated at 37 °C for 3 h and then centrifuged at 12,000 rpm for 15 min. The absorbance of the supernatant was measured at 540 nm using a multimode microplate reader (VICTOR Nivo 3S, PerkinElmer, USA) to determine the hemolysis rate. Double-distilled water (dd H₂O) served as the positive control, and physiological saline was used as the negative control.

4.12. Intracellular singlet oxygen generation

Intracellular singlet oxygen (¹O₂) levels were evaluated using 1,3-diphenylisobenzofuran (DPBF) as a fluorescent probe. RAW 264.7 cells were seeded in 24-well plates at a density of 5.0 × 10⁴ cells per well and co-cultured with the materials for 24 h. After incubation, the cells were washed and then incubated with PBS containing DPBF (8 μM) for 30 min. Following gentle rinsing, fluorescence images were captured using a fluorescence microscope (DM i8, Leica Microsystems CMS GmbH, Germany).

4.13. Intracellular reactive oxygen species (ROS) scavenging

Intracellular ROS levels were evaluated using 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA) as a fluorescent probe. RAW 264.7 cells were seeded in 24-well plates at a density of 5.0 × 10⁴ cells per well and co-cultured with the materials for 24 h. After incubation, the cells were washed and then incubated with serum-free medium containing DCFH-DA (10 μM) for 30 min. Following gentle rinsing, fluorescence images were captured using a fluorescence microscope (DM i8, Leica Microsystems CMS GmbH, Germany).

4.14. Regulation of macrophage polarization and NF-κB signaling in RAW 264.7 cells

Six groups were established based on the presence or absence of NIR irradiation to investigate macrophage polarization: Control, CIA, CIMA, Control + NIR, CIA + NIR, and CIMA + NIR. RAW 264.7 cells were seeded in 6-well plates at a density of 1.0 × 10⁵ cells per well and incubated for 12 h. To simulate an inflammatory microenvironment, RAW 264.7 cells in the experimental groups were treated with lipopolysaccharide (LPS, 300 ng/mL) for 24 h to induce pro-inflammatory M1 polarization, while control cells were treated with an equivalent volume of PBS. Cells in the CIA group were treated with LPS (300 ng/mL) and CIA (25 μg/mL), and those in the CIMA group were treated with LPS (300 ng/mL) and CIMA (25 μg/mL). For the laser irradiation groups, cells were exposed to an 808 nm NIR laser at 1.0 W/cm² in a cyclic mode. After 24 h of co-culture with the respective materials, macrophage polarization was evaluated by immunofluorescence staining and flow cytometry. The culture supernatants were collected, and the levels of IL-6 and TNF-α were measured using mouse ELISA kits (IL-6 and TNF-α, respectively).

For NF-κB pathway inhibition, RAW 264.7 cells were stimulated with LPS (300 ng/mL) and treated with an NF-κB inhibitor (8 μM) for 24 h. After incubation, cells were fixed and subjected to immunofluorescence staining for p65 and IKKβ to evaluate p65 nuclear translocation and changes in IKKβ signal intensity.

4.15. Osteogenic differentiation of MC3T3-E1 cells

Six groups were established based on the presence or absence of NIR irradiation to evaluate osteogenic differentiation: Control, CIA, CIMA, Control + NIR, CIA + NIR, and CIMA + NIR. MC3T3-E1 cells were seeded onto placed in 12-well plates at a density of 2.0 × 10⁴ cells per well and cultured in mineralization-inducing medium containing 10% FBS, 1% penicillin–streptomycin (PS), 10 mM β-glycerophosphate (Sigma-Aldrich, USA), 0.1 μM dexamethasone (Sigma-Aldrich, USA), and 50 μg/mL ascorbic acid in α-MEM (Sigma-Aldrich, USA). After co-culturing with the samples for 24 h, the cells were fixed with 4% paraformaldehyde, stained with FITC-labeled phalloidin (Solarbio, China) to visualize the cytoskeleton, and counterstained with DAPI (Solarbio, China) to label nuclei. Cell adhesion and morphology were observed and imaged using a laser scanning confocal microscope (LSM980, Zeiss, Germany).

MC3T3-E1 cells were seeded into 12-well plates, and once the cell confluence exceeded 80%, the medium was replaced with osteogenic induction medium. The cells were co-cultured with the samples, and the medium was refreshed every 2 days. On day 7, the cells were fixed with 4% paraformaldehyde and stained using an alkaline phosphatase (ALP) staining kit (Beyotime, China) according to the manufacturer's protocol. Macroscopic images were captured using a digital camera (Nikon, Japan), and microscopic images were acquired using an optical microscope (DM i8, Leica Microsystems CMS GmbH, Germany).

On day 21, the cells were fixed with 4% paraformaldehyde and stained with Alizarin Red S solution (OriCell, China) to evaluate mineralization. Macroscopic images were captured using a digital camera (Nikon, Japan), and microscopic images were acquired using an optical microscope (DM i8, Leica Microsystems CMS GmbH, Germany). Quantitative analysis of mineral deposition was performed using ImageJ software.

On days 7 and 21, the relative mRNA expression levels of osteogenesis-related genes—Runt-related transcription factor 2 (RUNX-2), bone morphogenetic protein 2 (BMP-2), osteocalcin (OCN), and collagen type I alpha 1 chain (COL-1)—were quantified by real-time qPCR. The relative expression levels were normalized using the 2⁻ΔΔCt method. Primer sequences are listed in Table S1 (Supporting Information).

4.16. Induction of osteoclast differentiation

RAW 264.7 cells were seeded into 48-well plates containing bovine cortical bone slices (100 μm thickness) at a density of 2.0 × 10⁴ cells per well. The cells were co-cultured with the materials in the presence of Receptor Activator for Nuclear Factor-κ B Ligand (RANKL, 50 ng/mL) and Macrophage colony-stimulating factor (M-CSF, 30 ng/mL) for 5 days to induce osteoclast differentiation. The morphology of osteoclasts and bone resorption pits was observed by scanning electron microscopy (SEM, Sigma 300, Zeiss, Germany). Cells were fixed with 4% paraformaldehyde and stained with a tartrate-resistant acid phosphatase (TRAP) staining kit (Solarbio, Beijing, China) for observation.

4.17. In vitro bone resorption assay

An in vitro bone resorption model was established by culturing osteoclasts on bovine cortical bone slices (100 μm thickness). RAW 264.7 cells were seeded at 2.0 × 10⁴ cells per well and induced with RANKL (50 ng/mL) and M-CSF (30 ng/mL) for 5 days to induce osteoclasts. The morphology of osteoclasts and resorption lacunae was examined using SEM (Sigma 300, Zeiss, Germany). The bone slices were then immersed in 25 μg/mL CIA or CIMA solution at 37 °C for 1 h, washed three times with PBS, and stained with FITC-labeled phalloidin (200 nM) and Hoechst for 30 min in the dark. Fluorescence imaging was performed using a laser scanning confocal microscope (LSM980, Zeiss, Germany).

4.18. Animal model and treatment

Male Sprague–Dawley (SD) rats (6-8 weeks old) were obtained from VITAL RIVER (Hangzhou, China) and housed under specific pathogen-free (SPF) conditions with a 12 h light/dark cycle, controlled temperature and humidity, and free access to food and water. All animal procedures were approved by the Ethics Committee of Nanchang University (NCULAE-20250519002) and complied with institutional guidelines. Rats were anesthetized and maintained under 2% isoflurane. The maxillary first molars were extracted using a dental elevator, and implant sites were prepared in the extraction sockets under saline irrigation to prevent overheating. Screw-type miniature Ti–6Al–4V implants (1.6 × 3 mm, ShengdaXing Metal Materials Co., Baoji, China) were inserted. No antibiotics were administered postoperatively.

After 4 weeks of healing, the rats were randomly divided into four groups (n = 6): Control, PI (peri-implantitis), PI + CIMA, and PI + CIMA + NIR. Except for the Control group, the other groups received subgingival injections of lipopolysaccharide (LPS, 50 μL, 3 mg/mL) around the implants every 2 days for 4 weeks to induce experimental peri-implantitis.

Following induction, the PI + CIMA group received subgingival injections of 25 μg/mL (100 μL) CIMA suspension every 2 days, while the PI + CIMA + NIR group additionally received near-infrared irradiation (808 nm, 1.0 W/cm², 5 min cyclic mode: initial 2 min heating followed by 30 s on/30 s off) directed externally at the buccal region. After 4 weeks of treatment, peri-implant tissues were harvested for analysis.

4.19. Micro-CT analysis

Peri-implant specimens were fixed in 4% paraformaldehyde and scanned using a micro-CT system (MILabs B.V. U-OI/CTHR, Netherlands). Three-dimensional reconstructions of the maxilla were generated using Imalytics Preclinical v3.0 software, and quantitative bone parameters were analyzed using CT evaluation software.

4.20. Histological evaluation

Fixed peri-implant tissues were dehydrated in graded ethanol and embedded in methyl methacrylate. Sections (200 μm thick) were cut and ground to a final thickness of 20 μm. H&E, Masson's trichrome, and TRAP staining were performed for histological evaluation of decalcified samples.

For undecalcified samples, methylene blue/acid fuchsin staining was conducted to assess new bone formation around the implants.

To examine in vivo macrophage polarization, dual immunofluorescence staining was performed on peri-implant bone tissues to identify CD86⁺ M1 and CD206⁺ M2 macrophages. Semi-quantitative analysis of fluorescence intensity was carried out. P65 immunofluorescence staining was imaged using a laser scanning confocal microscope (LSM980, Zeiss, Germany), and fluorescence intensity was quantified with ImageJ (Bethesda, USA).

In addition, rat blood and serum samples were collected for hematological and biochemical analyses, including alanine aminotransferase (ALT), aspartate aminotransferase (AST), creatinine (CREA), and urea (UREA).

4.21. In vivo antibacterial activity of CIMA

After treatment, titanium implants were retrieved and vortexed in PBS to detach adherent bacteria. The antibacterial efficacy of CIMA under NIR irradiation was quantitatively evaluated by plate counting according to the standard CFU calculation formula. H&E staining was used to assess inflammation at the implant sites, and bacterial infiltration was examined using Giemsa staining (Servicebio, Wuhan, China).

4.22. In vivo fluorescence imaging

Animals were divided into three groups (Control, CIMA, and PI + CIMA) and subjected to fluorescence imaging using a small animal in vivo imaging system (IVIS Lumina III, PerkinElmer, USA) on days 1, 7, and 28 post-treatment (excitation: 760 nm; emission: 845 nm). Data were analyzed using Living Image software.

4.23. In vivo bone resorption monitoring

Seven days after treatment, the maxillary alveolar bone around the first molar was harvested as the observation region. Three-dimensional fluorescence imaging reconstruction was performed using a laser scanning confocal microscope (LSM980, Zeiss, Germany) under 760 nm excitation.

4.24. Transcriptomic analysis of in vivo immunomodulation

Peri-implant maxillary bone tissues were collected, ground, and used for RNA extraction with the MJZol Total RNA Extraction Kit. RNA sequencing was performed on the Majorbio sequencing platform (Shanghai, China).

4.25. Statistical analysis

All experiments were performed at least in triplicate. Statistical analyses were conducted using GraphPad Prism 8.0.2 software (GraphPad, USA). One-way analysis of variance (ANOVA) followed by Dunnett's post hoc test was applied to compare groups. Results are presented as mean ± standard deviation (SD). Statistical significance was defined as ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; when the p-value was greater than 0.05, the difference was considered not statistically significant and was denoted as ns > 0.05.

Funding statement

This work was funded by The National Natural Science Foundation of China (No. 32560229 to Xiaolei Wang and No. 82160194 to Lan Liao), Key Projects of the Jiangxi Provincial Key Research and Development Program (No. 20243BBI91023 to Lan Liao and No. 20212BBG73004 to Xiaolei Wang), The Jiangxi Province Key Laboratory of Bioengineering Drugs (No. 2024SSY07061 to Xiaolei Wang) and The Interdiscipline Innovation Fund Project of Nanchang University (PYJX20230001 to Xiaolei Wang), Clinical Research Cultivation Project of the First Affiliated Hospital of Nanchang University (No. YFYLCYJPY202401).

CRediT authorship contribution statement

Ziyu Zhou: Writing – original draft, Software, Methodology, Conceptualization. Qiaowen Zheng: Methodology, Data curation. Jiahui Yang: Methodology, Investigation. Yifan Cai: Formal analysis, Data curation. Zixun Lan: Visualization, Software. Yingying Xu: Validation. Xianglin Dai: Software, Methodology. Lan Liao: Validation, Project administration, Funding acquisition. Xiaolei Wang: Writing – review & editing, Validation, Project administration, Funding acquisition.

Data availability statement

All relevant data are within the manuscript and its Supporting Information files.

Ethics approval and consent to participate

All animal procedures were performed with the Guidelines for the Care and Use of Laboratory Animals of Nanchang University in China and approved by the Animal Ethics Committee of Nanchang University (Nanchang, China, (NCULAE-20250519002). This study was also approved by the Ethics Committee of the Affiliated Stomatological Hospital of Nanchang University (Approval No. 2025-072), and all procedures involving human participants/samples were performed in accordance with the principles of the Declaration of Helsinki.

Declaration of competing interest

All authors declared that no conflict of interest existed.

Appendix A. Supplementary data

The following is the Supplementary data to this article: