Dual‐Anchored Clickable Peptide via SPAAC for Gelatinase‐Responsive Antibacterial and Osteogenic Functions on Titanium Implants

TL;DR

This paper introduces a titanium implant with a dual-anchored peptide that fights infection while supporting bone growth, offering a solution to a major clinical challenge in orthopedic implants.

Contribution

A novel dual-anchored peptide design that preserves osteogenic function while enabling antibacterial activity in response to infection.

Findings

The Ti-Dual implant eliminated 99.37% of P. aeruginosa within 10 minutes and 99.99% within 120 minutes.

The RGD motif remained anchored post-activation, ensuring continuous cell adhesion and tissue integration.

In vivo, Ti-Dual suppressed inflammation, mitigated bone resorption, and enhanced osteogenesis in a rat model.

Abstract

Antibacterial orthopedic implants that simultaneously promote osteointegration remain an unmet clinical challenge. Conventional enzyme‐responsive antibacterial surfaces often suffer from irreversible loss of osteogenic motifs upon activation, limiting their regenerative capacity post‐infection. Herein, we report a dual‐anchored peptide design engineered on titanium implants (Ti‐Dual) that addresses this limitation by retaining biofunctional motifs after pathogen‐triggered activation. The peptide construct integrates an antimicrobial sequence (HHC36) and a cell‐adhesive RGD motif connected via a gelatinase‐cleavable spacer (GPLGV). Terminal azide groups enable stable dual‐point grafting through SPAAC chemistry, overcoming the low grafting efficiency associated with mixed RGD grafting systems. Under physiological conditions, the constrained conformation suppresses antibacterial activity, favoring osteogenesis. Upon infection, bacterial gelatinase cleaves the linker, activating rapid and potent antibacterial effects—eliminating 99.37% of P. aeruginosa within 10 min, and 99.73% of S. aureus and 99.99% of P. aeruginosa within 120 min—while the RGD motif remains anchored, ensuring continuous cell adhesion and tissue integration. In vivo, Ti‐Dual effectively eradicated multidrug‐resistant P. aeruginosa, suppressed inflammatory responses, mitigated bone resorption, and enhanced osteogenesis in a rat femoral infection model. This design resolves the critical trade‐off between infection responsiveness and sustained pro‐regenerative function, offering a robust and adaptive strategy for infected bone defect repair.

To address the infection‐associated orthopedic implant failures, this study presents a gelatinase‐responsive antibacterial titanium implant (Ti‐Dual) featuring a dual‐anchored fusion peptide containing antibacterial (HHC36) and osteogenic (RGD) domains connected by a cleavable GPLGV linker. Under infection, bacterial gelatinase triggers antibacterial function while preserving RGD‐mediated bone‐repair activity. In vivo results demonstrate effective P. aeruginosa eradication, reduced inflammation, and enhanced bone regeneration.

Figures

Preparation of gelatinase‐responsive antibacterial surface. (a) Schematic of the gelatinase‐responsive antibacterial peptide grafted surface. (b) The densities of three fusion peptides after 6 h of reaction at varying concentrations (n = 4). (c) XPS high‐resolution N 1s spectra of the surfaces. Ti‐Mono, Ti‐Dual‐N, and Ti‐Dual had equivalent peptide densities.

Antibacterial properties of gelatinase‐responsive peptide‐grafted surfaces. (a) Schematic of peptide conformational changes on the surfaces. Bactericidal kinetics of the surfaces against (b) S. aureus and (c) P. aeruginosa (n = 6). (d) Live/dead staining of S. aureus and P. aeruginosa on the surfaces using FDA/PI (scale bar = 20 µm). (e) SEM images of S. aureus and P. aeruginosa after incubation with different surfaces (scale bar = 1 µm). Anti‐biofilm activities of the surfaces against S. aureus and P. aeruginosa measured by crystal violet staining: (f) representative photographs (scale bar = 200 µm) and (g,h) quantitative results (n = 4). * denoted p <0.05, denoted p <0.01, * denoted p <0.001.

In vitro biocompatibility of gelatinase‐responsive peptide‐grafted surfaces. (a) Schematic of peptide conformational changes on the surfaces. (b) CCK‐8 results of mBMSCs after 1 and 5 days of culture on the indicated surfaces (n = 5). (c) Representative structure of Mono, Dual‐N, and Dual peptides from all‐atom molecular dynamics simulation. (d) Solvent‐accessible surface area (SASA) of RGD and HHC36 fragments in the fusion peptides calculated by molecular dynamics simulations. (e) Cytoskeletal staining of mBMSCs after 1 and 5 days of culture on the indicated surfaces (scale bar = 250 µm). (f) Representative images showing migration of mBMSCs on different surfaces in the open wound model (scale bar = 500 µm). (g) Quantified cell area within the open wound model after 24 h of migration (n = 3). (h) Schematic diagram of the bacteria‑cell co‑culture setup. (i) Cytoskeletal staining of mBMSCs before and after infected with P. aeruginosa for 24 h (scale bar = 250 µm). Quantitative analysis of the retention ratio of (j) cell number and (k) cell area in the field of view after infected with P. aeruginosa for 24 h (n = 4). * denoted p <0.05, denoted p <0.01, * denoted p <0.001.

In vitro osteogenic properties of gelatinase‐responsive peptide‐grafted surfaces. (a) Schematic illustration of osteogenic induction in mBMSCs. (b–d) qPCR analysis of osteogenic markers in mBMSCs after 7 and 14 days of culture: (b) RUNX2, (c) ALP, and (d) OPN (n = 6). (e) Representative immunofluorescence images showing OPN expression after 14 days (scale bar = 50 µm). (f) ALP staining (scale bar = 1 mm) and (g) corresponding quantitative ALP activity measured at 7, 14, and 21 days (n = 5). (h) Alizarin Red S staining of mineralized nodules (scale bar = 1 mm) and (i) corresponding quantitative calcium deposition analysis at 14 and 21 days (n = 5). * denoted p <0.05, denoted p <0.01, * denoted p <0.001.

In vivo assay with the infection model at 7 days post‐implantation. (a) Schematic illustration of the surgical procedure for establishing the infection model. (b) The images of the infected femoral sites. Yellow arrows indicated abscess formation. (c) Microbiological assessment by agar plate imprint method following implant retrieval. Quantification of bacterial loads (d) adherent on the screw and (e) in the surrounding peri‐implant tissue (n = 4). (f) MMP activity in peri‐implant tissues (n = 4). (g) H&E staining of the implant site (upper row: 10×, scale bar = 200 µm; lower row: 40×, scale bar = 50 µm). Yellow arrows highlighted inflammatory cells. (h) Quantitative analysis of inflammatory cells from 40× H&E sections (n = 6). (i) Immunofluorescence co‐staining of tissue adjacent to the implant for iNOS⁺ M1 macrophages (red), CD206⁺ M2 macrophages (green), and nuclei (DAPI, blue) (scale bar = 50 µm). The top row showed merged images, and the bottom row displayed individual channels. * denoted p <0.05, denoted p <0.01, * denoted p <0.001.

In vivo assay with the infection model at 7 days post‐implantation. (a) The images of the infected femoral sites. The yellow dashed line indicated the edge of the femur. (b) Quantification of bacterial loads adherent on the screw and in the surrounding peri‐implant tissue (n = 10). (c) MMP activity in explanted screws and peri‐implant tissues (n = 5). (d) H&E staining of the implant site. The image below was a magnification of the black‐boxed area in the image above. Blue arrows highlighted fibrosis, green arrows highlighted mild vascular hyperplasia, black arrows highlighted inflammatory infiltration, and yellow arrows highlighted osteolytic necrosis with mineralization. The scale bars were 1000 µm (top row), 100 µm (middle row), and 50 µm (bottom row). (e) Micro‐CT analysis of representative 3D reconstructed images of bone formation around the implant. The quantitative parameters from Micro‐CT assay: (f) bone surface density (BS/BV), (g) bone volume fraction (BV/TV), (h) trabecular thickness (Tb·Th), and (i) intersection surface (n = 3). (j) Undecalcified sections stained with methylene blue/basic fuchsin. The image below was a magnification of the red‐boxed area in the image above. The blue arrow indicated the unbonded region at the implant‐bone interface. The scale bars were 1000 µm (top row), 500 µm (middle row), and 100 µm (bottom row). (k) Immunofluorescence co‐staining of tissue adjacent to the implant for IL‐10⁺ (red), OPN⁺ (green), and DAPI‐stained nuclei (blue). The left column showed merged images, and the right column displayed individual channels (scale bar = 100 µm). (l) The images of TRAP staining (scale bar = 50 µm). * denoted p <0.05, denoted p <0.01, * denoted p <0.001.

Full text

Introduction

1

Orthopedic titanium implants are indispensable devices in bone reconstruction due to their excellent biocompatibility and bone‐matching mechanical properties [1]. However, implant‐associated infections (IAIs) persist as a formidable clinical challenge, significantly compromising surgical success rates and patient outcomes. Alarmingly, even under strict aseptic conditions, infection rates reach 1–2% in closed fracture surgeries and soar to 30% in severe open fractures with extensive tissue damage [2]. Critically, IAIs are the primary cause of implant revision surgeries, imposing substantial healthcare burdens and exposing patients to exacerbated pain, functional impairment, and even life‐threatening complications [3].

To address this challenge, extensive efforts have focused on endowing implants with intrinsic antibacterial properties. Current strategies primarily employ surface modifications incorporating antibacterial agents such as metallic ions [4, 5], antibiotics [6, 7, 8], cationic biocides [9, 10, 11], or antimicrobial peptides (AMPs) [12, 13, 14]. These functionalized surfaces confer immediate antibacterial activity to implants and demonstrate potential in preventing bacterial colonization. However, a fundamental dilemma arises where robust antibacterial activity requires sufficiently high concentrations of antibacterial agents that inevitably induce significant host‐cell cytotoxicity [15], severely compromising osseointegration [5]. This inherent conflict critically limits the clinical application of these strategies.

Recently, microenvironment‐responsive antibacterial strategies have emerged as a promising alternative, leveraging unique advantages in spatiotemporal control. These approaches engineer an implant that detects and responds to specific exogenous (e.g., light [8, 16, 17], heat [18]) or endogenous (e.g., enzymes [19, 20, 21], pH [22, 23], reactive oxygen species (ROS) [24]) stimuli at the surgical site. Their core principle lies in on‐demand activation in which antibacterial activity is selectively triggered during infection‐prone or active infection states while remaining dormant, or even enhancing osteogenesis, in normal healing phases. This precise on‐demand operation minimizes off‐target cytotoxicity while maximizing therapeutic specificity. AMPs present ideal candidates for such intelligent responsiveness due to their conformation‐dependent antibacterial activity [25]. Through rational sequence design, AMPs can undergo reversible conformational changes in response to microenvironmental cues, enabling precise “on‐off” switching of antibacterial activity.

Among various antibacterial response strategies, enzyme‐responsive mechanisms have garnered significant attention due to their high efficiency and specificity [21, 26]. This approach utilizes specific enzymes secreted by pathogenic bacteria, such as gelatinase [27] and β‐lactamase [28], as intelligent triggers to achieve precise identification of infection sites and controlled drug release. Notably, highly virulent and widely infectious pathogens, including P. aeruginosa [29] and Methicillin‐Resistant Staphylococcus aureus (MRSA) [19], secrete key virulence factors like gelatinase, which play central roles in tissue invasion, biofilm formation, and immune evasion. These processes facilitate the spread of bacteria from initial infection sites into deeper tissues and even the bloodstream, leading to bacteremia, osteomyelitis and sepsis [30]. Enzyme‐responsive antibacterial systems enable precise targeting by releasing drugs only in regions activated by virulence enzymes, thereby selectively eliminating pathogens while largely preserving normal flora and host tissues [31, 32]. Moreover, they exhibit high sensitivity and rapid response even to low enzyme concentrations, rendering them especially effective for early‐stage intervention in chronic and biofilm‐based infections [33]. Notably, an integrated analysis of bacterial activity in bone infections [34] reveals that the majority of pathogenic bacteria secrete gelatinase [26, 35, 36]. This enzyme plays a particularly destructive role by degrading collagen—a fundamental component of the extracellular matrix and bone tissue [37]. Such degradation disrupts trabecular architecture, weakens biomechanical integrity, and accelerates bone resorption, ultimately compromising bone regeneration [38]. These detrimental effects highlight the critical need for therapeutic strategies capable of rapidly suppressing bacterial activity in gelatinase‐enriched infected bone environments.

Notably, enzyme‐responsive AMPs variants have been extensively investigated for various biomedical applications, including functionalizing implant surfaces, owing to their high sensitivity. For instance, researchers can precisely orient AMPs sequences such that the antibacterial sequence is initially concealed, while a specific pro‐regenerative sequence remains exposed on the surface. In the absence of infection, the exposed pro‐regenerative sequence promotes tissue repair while simultaneously shielding the antibacterial activity of the hidden sequence. Upon infection, an enzyme‐cleavable linker between the two sequences is severed, thereby exposing the antibacterial sequence and activating its function.

Although this widely‐used enzyme‐responsive AMPs enable a switch between pro‐regenerative and antibacterial states, existing designs face a critical limitation wherein enzyme‐triggered cleavage activates antibacterial activities at the expense of essential pro‐regenerative functions. Specifically, following cleavage, the pro‐regenerative sequence detaches from the surface, resulting in the irreversible loss of its function [19, 39]. This functional trade‐off compromises post‐antibacterial bone regeneration, failing to meet clinical demands for combined anti‐infection and osseointegration capabilities. Conversely, static grafting of both antibacterial and pro‐regenerative sequences directly onto the material surface renders the system non‐responsive, while the antibacterial activity cannot be shielded during the resting state, potentially hindering the pro‐regenerative process.

To address the aforementioned issue, we introduced azide groups at both ends of the peptide and constructed a dual‐anchored functional peptide surface via strain‐promoted azide‐alkyne cycloaddition (SPAAC). We integrated a gelatinase‐responsive linker (GPLGV) between an antimicrobial peptide HHC36 and a cell‐adhesive RGD motif, yielding a dual‐anchored molecular construct: N_3_‐OEG_4_‐HHC36‐GPLGV‐RGD‐K‐OEG_4_‐N_3_ (denoted as Dual). In infectious microenvironments, pathogen‐derived gelatinase cleaves the linker, restoring HHC36 conformational flexibility and activating its antibacterial function. Concurrently, the remaining RGD sequence maintains bioactivity to promote cell adhesion and tissue integration. In a rat model of infected femoral defect, the Dual‐functionalized titanium implant (Ti‐Dual) exhibited rapid bactericidal action, sustained osteogenic capacity, and attenuated local gelatinase activity. Together, this work presents Ti‐Dual as an intelligent implant surface that dynamically switches from antibacterial to pro‐regenerative mode in response to bacterial enzymatic activity, offering a modular and adaptive design strategy for bioactive orthopedic implants.

Results

2

The Characterization of Dual‐Anchored Antibacterial Peptides (AMPs)

2.1

Initially, we designed a fusion peptide construct combining an antibacterial sequence (HHC36) and a cell‐adhesive RGD motif linked via a gelatinase‐cleavable spacer (GPLGV), with azide groups modified at both ends. This enabled the design of a dual‐anchored peptide, termed Dual (Figure S1a), which integrated both antibacterial and pro‐regenerative functions. For comparative purposes, we also designed a mono‐anchored peptide by introducing an azide group at the N‐terminus of HHC36 with the same sequence, termed Mono (Figure S1b). Furthermore, by mutating the enzyme‐responsive sequence in the dual‐anchored peptide to GGPLV, we engineered a dual‐anchor peptide lacking enzyme‐responsive properties, termed Dual‐N (Figure S1c). All three peptides showed potent antibacterial activity in solution, killing nearly 100% of both S. aureus and E. coli at concentrations ranging from 10 to 20 µm (Figure S2). Meanwhile, HRMS analysis (Figure S3) demonstrated that after gelatinase treatment, Dual and Mono exhibited characteristic cleavage fragments at glycine‐adjacent sites. In contrast, Dual‐N peptide remained intact after gelatinase treatment, confirming its suitability as a non‐cleavable control. In subsequent investigations, we focused on characterizing the performance of titanium‐based implants functionalized with these three peptides.

After 6 h of grafting, the density of all three peptides increased with higher peptide concentrations. Consistent with prior experiments, we selected reaction concentrations of 30, 30, and 50 µm for Mono, Dual‐N, and Dual peptides, respectively, to achieve comparable surface grafting densities (53.12 ± 2.94, 55.02 ± 3.18, and 57.16 ± 4.56 molecules/nm^2^ in Figure 1b; Figure S4). XPS analysis confirmed this density matching through consistent enhancement of N signals (Figure 1c; Figure S5), while FTIR spectra demonstrated similar amide I peak intensities at 1640 cm^−1^ (Figure S6). Notably, antibacterial experiments on surfaces with different grafting densities revealed that a density of 57.16 molecules/nm^2^ enabled the Ti‐Dual surface to achieve a bactericidal efficiency exceeding 99% against both S. aureus and P. aeruginosa (Figure S7). This result validated the rationality of our screening strategy for determining the effective grafting density. Meanwhile, surface wettability characterization revealed that the dual‐anchored peptide functionalized surfaces (Ti‐Dual and Ti‐Dual‐N) displayed greater hydrophilicity than the mono‐anchored counterpart (Ti‐Mono), attributable to increased PEG presence on the surfaces (Figure S8).

Preparation of gelatinase‐responsive antibacterial surface. (a) Schematic of the gelatinase‐responsive antibacterial peptide grafted surface. (b) The densities of three fusion peptides after 6 h of reaction at varying concentrations (n = 4). (c) XPS high‐resolution N 1s spectra of the surfaces. Ti‐Mono, Ti‐Dual‐N, and Ti‐Dual had equivalent peptide densities.

Furthermore, we demonstrated using fluorescent dyes that the dual peptide could be constructed on the titanium surface via a dual‐grafting method (Figure S9). First, after reacting the peptide‐modified surfaces with DBCO‐Cy3 dye, both Ti‐Dual‐N and Ti‐Dual exhibited fluorescence intensities similar to that of Ti‐Mono, indicating that there were no significant exposed azide reactive groups from the peptides on the surface. On the other hand, after reacting the peptide‐modified surfaces with N_3_‐Cy3, the fluorescence intensity of Ti‐Mono was significantly higher than that of Ti‐Dual‐N and Ti‐Dual. This was because, under comparable grafting densities, Ti‐Dual‐N and Ti‐Dual consumed twice as many of the surface DBCO groups as Ti‐Mono, leaving fewer sites available for N_3_‐Cy3 to react with and graft onto.

The stability of the surface‐grafted peptides was assessed via degradation assay in PBS. While the Micro‐BCA assay confirmed stable peptide density with no significant loss within the first week, reductions of approximately 7.94%, 8.69%, and 30.34% were recorded by the end of weeks 2, 3, and 4, respectively (Figure S10a). Furthermore, XPS results verified that the surface peptides remained detectable after four weeks, although with a decrease in signal intensity compared to earlier time points (Figure S10b,c). These findings indicated that the peptides maintained a relatively stable presence on the surface.

In Vitro Antibacterial Assay

2.2

Bacteria secreting gelatinase degraded collagen in human tissues, causing significant harm. MMP fluorescence assays confirmed that S. aureus and P. aeruginosa were active secretors of gelatinase (Figure S11a), consistent with prior literature [32, 36]. Notably, these two species represented the most prevalent pathogens in orthopedic clinical infections. We therefore investigated the in vitro antibacterial performance of the functionalized surfaces against these two bacteria.

As designed, the peptides underwent conformational changes upon exposure to bacterial‐secreted gelatinase (Figure 2a). In vitro antibacterial assays confirmed this conformational shift. In gelatinase‐secreted bacteria, enzymatic cleavage of the GPLGV sequence exposed the concealed antibacterial sequence on Ti‐Mono while liberating the antibacterial sequence on Ti‐Dual (Figure 2a). Consequently, Ti‐Mono and Ti‐Dual demonstrated potent bactericidal activity against S. aureus and P. aeruginosa, achieving 99.73% and 99.73% inhibition of S. aureus respectively, and 98.22% and 99.99% inhibition of P. aeruginosa within 2 h (Figure 2b,c; Figure S12). Given that P. aeruginosa secreted higher gelatinase levels than S. aureus, both surfaces exhibited accelerated bactericidal effects against P. aeruginosa. Specifically, Ti‐Mono eliminated 94.51% of P. aeruginosa within 60 min, while Ti‐Dual achieved 99.37% elimination in just 10 min. Conversely, Ti‐Mono and Ti‐Dual showed negligible antibacterial activity against non‐gelatinase‐secreted bacteria (E. coli and L. casei) due to their inability to cleave the GPLGV sequence (Figure S11). This observation validated the surface responsiveness to gelatinase. Similarly, the control surface Ti‐Dual‐N, functionalized with uncleavable GGPLV sequence, maintained its dual‐grafted configuration, yielding minimal antibacterial efficacy against S. aureus (58.13% inhibition) and P. aeruginosa (54.47% inhibition) over 2 h.

*Antibacterial properties of gelatinase‐responsive peptide‐grafted surfaces. (a) Schematic of peptide conformational changes on the surfaces. Bactericidal kinetics of the surfaces against (b) S. aureus and (c) P. aeruginosa (n = 6). (d) Live/dead staining of S. aureus and P. aeruginosa on the surfaces using FDA/PI (scale bar = 20 µm). (e) SEM images of S. aureus and P. aeruginosa after incubation with different surfaces (scale bar = 1 µm). Anti‐biofilm activities of the surfaces against S. aureus and P. aeruginosa measured by crystal violet staining: (f) representative photographs (scale bar = 200 µm) and (g,h) quantitative results (n = 4). * denoted p <0.05, *denoted p <0.01, *** denoted p <0.001.

Furthermore, FDA/PI staining (Figure 2d) revealed that while Ti and Ti‐S surfaces maintained predominantly viable bacteria, Ti‐Mono, Ti‐Dual‐N, and Ti‐Dual showed progressively decreasing viability, correlating with plate counting results. NPN uptake assays (Figure S13) further confirmed outer membrane disruption, with Ti‐Dual and Ti‐Mono exhibiting fluorescence intensities >950 (vs. <700 for controls), indicating significant phospholipid layer disturbance. Complementarily, SEM imaging (Figure 2e) provided visual confirmation of membrane collapse and cytoplasmic leakage in bacteria exposed to Ti‐Mono and Ti‐Dual, contrasting with the intact morphology observed on Ti and Ti‐S. Biofilm formation assays were performed by crystal violet staining (Figure 2f–h). After 36‐h incubation with bacteria, Ti‐Mono and Ti‐Dual significantly reduced biofilm formation. Specifically, the OD values of the control (Ti) group were 2.40‐ and 2.36‐fold higher than those of Ti‐Mono and Ti‐Dual, respectively, against S.aureus biofilms (Figure 2g), and 3.81‐ and 4.17‐fold higher for P. aeruginosa biofilms (Figure 2h).

In Vitro Biocompatibility

2.3

We evaluated the in vitro biocompatibility of the surfaces using mouse bone marrow mesenchymal stem cells (mBMSCs). To simulate infected conditions, Ti‐Mono and Ti‐Dual surfaces were pretreated with bacterial‐derived gelatinase prior to assessment. These pretreated surfaces were referred to as Ti‐Mono‐Gel and Ti‐Dual‐Gel, respectively (Figure 3a). Control surfaces functionalized with either single peptides or mixed peptides were prepared for comparative analysis. For the single‐peptide control, Ti‐H surfaces were prepared using a 25 µm concentration of HHC36 to achieve an equivalent peptide density (51.05 molecule/nm^2^, Figure S14a,b). As grafting RGD peptide alone proved difficult even at 400 µm (Figure S14c–e), Ti‐R surfaces prepared at this maximum concentration served as the RGD control. This result underscored the superior RGD grafting efficiency enabled by the Mono or Dual peptide strategy. Additionally, mixed‐grafted surfaces (Ti‐HR) were prepared with equivalent HHC36 density but maximal achievable RGD incorporation.

*In vitro biocompatibility of gelatinase‐responsive peptide‐grafted surfaces. (a) Schematic of peptide conformational changes on the surfaces. (b) CCK‐8 results of mBMSCs after 1 and 5 days of culture on the indicated surfaces (n = 5). (c) Representative structure of Mono, Dual‐N, and Dual peptides from all‐atom molecular dynamics simulation. (d) Solvent‐accessible surface area (SASA) of RGD and HHC36 fragments in the fusion peptides calculated by molecular dynamics simulations. (e) Cytoskeletal staining of mBMSCs after 1 and 5 days of culture on the indicated surfaces (scale bar = 250 µm). (f) Representative images showing migration of mBMSCs on different surfaces in the open wound model (scale bar = 500 µm). (g) Quantified cell area within the open wound model after 24 h of migration (n = 3). (h) Schematic diagram of the bacteria‑cell co‑culture setup. (i) Cytoskeletal staining of mBMSCs before and after infected with P. aeruginosa for 24 h (scale bar = 250 µm). Quantitative analysis of the retention ratio of (j) cell number and (k) cell area in the field of view after infected with P. aeruginosa for 24 h (n = 4). * denoted p <0.05, *denoted p <0.01, *** denoted p <0.001.

CCK‐8 assay (Figure 3b) demonstrated that both antibacterial surfaces (Ti‐Mono and Ti‐Dual) showed biocompatibility comparable to pristine Ti on Day 1, while significantly outperforming single‐peptide functionalized surfaces (Ti‐H and Ti‐R) and the mix‐grafted surface (Ti‐HR) in supporting cell viability. This enhanced performance was attributed to the minimal RGD density present on Ti‐H, Ti‐R, and Ti‐HR. Notably, Ti‐Mono and Ti‐Dual also exhibited superior biocompatibility compared to Ti‐Dual‐N, despite all three surfaces incorporating both the HHC36 and RGD sequences. Specifically, Ti‐Mono and Ti‐Dual exhibited 1.14‐ and 1.30‐fold higher cell viability than Ti‐Dual‐N on Day 1, and 1.14‐ and 1.33‐fold higher on Day 5. To elucidate the underlying mechanism, we conducted molecular dynamics simulations (Figure 3c,d; Figure S15). The results revealed that compared to Ti‐Mono and Ti‐Dual, the peptide conformation on Ti‐Dual‐N resulted in a higher solvent‐accessible surface area (SASA) for the HHC36 sequence and a lower SASA for the RGD sequence. This indicated greater exposure of the antibacterial HHC36 sequence and reduced accessibility of the cell‐adhesive RGD sequence on Ti‐Dual‐N surfaces. Consequently, Ti‐Dual‐N exhibited compromised cytocompatibility relative to both Ti‐Mono and Ti‐Dual.

Following gelatinase treatment, Ti‐Mono‐Gel exhibited a marked reduction in biocompatibility compared to Ti‐Mono, with decreases of 21.35% and 13.79% on Day 1 and Day 5, respectively. This resulted from gelatinase cleavage of the GPLGV linker, which caused detachment of the RGD sequence from Ti‐Mono‐Gel (Figure 3a). In contrast, Ti‐Dual‐Gel showed only marginal biocompatibility reduction relative to Ti‐Dual after gelatinase treatment (8.97% and 6.86% on Day 1 and Day 5), with no statistically significant difference between them. The retained performance was attributable to preserved RGD surface retention post‐cleavage due to Dual's molecular architecture. Additionally, the minor reduction in Ti‐Dual‐Gel biocompatibility resulted from enhanced antibacterial activity of the HHC36 sequence.

Cytoskeletal analysis (Figure 3e; Figure S16) corroborated the above findings, with vinculin immunostaining (Figure S17) further revealing enhanced focal adhesion formation and lamellipodia extension on Ti‐Dual surfaces.

Cell migration assays revealed trends distinct from the CCK‐8 results. Pristine Ti demonstrated no significant advantage in cell migration testing, retaining an open wound area of 1.44 mm^2^ after 24 h. This likely stemmed from Ti's lack of the porous network structure generated by alkali treatment in other groups. Meanwhile, Ti‐Dual exhibited significantly enhanced cell migration capability, with open wound site cell area measuring 1.66‐, 1.92‐, 1.41‐, and 1.67‐fold higher than those on pristine Ti, Ti‐S, Ti‐Mono, and Ti‐Dual‐N, respectively, after 24 h (Figure 3f,g).

As a bone implant with intrinsic antibacterial properties, the ability to facilitate tissue regeneration following infection resolution was of paramount clinical significance. To model infection‐mediated tissue damage, we established a bacteria‐cell coculture system (Figure 3h) where MMP fluorescence surged from 13 to >300 RFU upon P. aeruginosa infection (Figure S18). Under these challenging conditions, non‐antibacterial surfaces (Ti/Ti‐S) exhibited severe mBMSCs shrinkage and detachment, with markedly reduced cell spreading area, followed by complete proliferation arrest by 24 h. In contrast, surfaces incorporating antibacterial HHC36 peptides maintained higher cell densities due to their varying antibacterial efficacy. The Ti‐Dual demonstrated particular advantages, showing 1.36‐fold and 1.75‐fold greater cell number retention ratio than Ti‐Mono and Ti‐Dual‐N, respectively after 24 h of infection, attributable to preserved RGD accessibility and enhanced chain flexibility post‐enzymatic cleavage (Figure 3i–k). Additionally, complementary experiments with L929 cells further validated these findings (Figure S19), with Ti‐Dual surfaces supporting 16.58%‐32.95% greater cell adhesion than other groups, suggesting its superior potential for promoting tissue repair in infected environments.

In Vitro Osteogenic Differentiation

2.4

To evaluate the osteogenic differentiation potential of surface‐modified materials, we analyzed the expression of osteogenic markers (RUNX2, ALP, and OPN) at days 7 and 14 post‐osteogenic induction (Figure 4a; Table S1). Early osteogenic differentiation, assessed via RUNX2 at day 7, demonstrated that the dual‐anchored peptide functionalized surfaces (Ti‐Dual‐N and Ti‐Dual) exhibited superior performance. These surfaces showed 1.18‐fold and 1.24‐fold increases over Ti, and 1.28‐fold and 1.22‐fold enhancements relative to Ti‐Mono (Figure 4b). Immunofluorescence analysis of RUNX2 (Figure S20) confirmed significantly higher nuclear accumulation in Ti‐Dual‐N and Ti‐Dual compared to Ti and Ti‐Mono, consistent with PCR data. Regarding mid‐stage osteogenic differentiation, ALP gene expression followed a similar pattern. At day 7, Ti‐Dual‐N displayed the highest ALP expression, while Ti‐Dual levels were comparable to Ti and significantly higher (1.36‐fold) than Ti‐Mono. By day 14, Ti‐Dual ALP expression increased substantially, reaching 1.54‐fold and 1.38‐fold higher levels than Ti and Ti‐Mono, respectively (Figure 4c).

*In vitro osteogenic properties of gelatinase‐responsive peptide‐grafted surfaces. (a) Schematic illustration of osteogenic induction in mBMSCs. (b–d) qPCR analysis of osteogenic markers in mBMSCs after 7 and 14 days of culture: (b) RUNX2, (c) ALP, and (d) OPN (n = 6). (e) Representative immunofluorescence images showing OPN expression after 14 days (scale bar = 50 µm). (f) ALP staining (scale bar = 1 mm) and (g) corresponding quantitative ALP activity measured at 7, 14, and 21 days (n = 5). (h) Alizarin Red S staining of mineralized nodules (scale bar = 1 mm) and (i) corresponding quantitative calcium deposition analysis at 14 and 21 days (n = 5). * denoted p <0.05, *denoted p <0.01, *** denoted p <0.001.

Late‐stage osteogenic maturation, evidenced by OPN expression, revealed higher OPN levels on all peptide‐functionalized surfaces (Ti‐Mono, Ti‐Dual‐N, Ti‐Dual) compared to Ti at day 14 (Figure 4d), with increases of 4.01‐fold, 4.30‐fold, and 3.80‐fold, respectively. Immunofluorescence analysis of OPN (Figure 4e; Figure S21) similarly showed significantly greater nuclear accumulation in Ti‐Dual‐N and Ti‐Dual versus Ti and Ti‐S, correlating with PCR results. Additionally, we observed fewer cells on Ti surfaces in OPN immunofluorescence staining. This resulted from increased cell detachment during extended culture (14 days) and subsequent staining washes, particularly affecting smooth Ti surfaces. In contrast, the porous Ti‐S substrate maintained superior cell retention due to enhanced adhesion.

Additionally, quantitative ALP staining revealed significantly higher ALP activity on Ti‐Dual compared to other groups across all time points (Figure 4f). At day 7, Ti‐Dual exhibited ALP activity levels 1.29‐fold, 1.12‐fold, and 1.14‐fold higher than those on Ti, Ti‐Mono, and Ti‐Dual‐N, respectively. By day 14, ALP activity on Ti‐Dual remained elevated, measuring 1.34‐fold, 1.32‐fold, and 1.31‐fold greater than Ti, Ti‐Mono, and Ti‐Dual‐N. Although ALP activity increased substantially on other surfaces by day 21, Ti‐Dual maintained the highest levels, exceeding Ti, Ti‐Mono, and Ti‐Dual‐N by 1.16‐fold, 1.23‐fold, and 1.05‐fold, respectively (Figure 4g). Furthermore, matrix mineralization was quantified by alizarin red S (ARS) staining (Figure 4h,i), revealing accelerated calcium nodule formation on peptide‐functionalized surfaces. Notably, Ti‐Dual exhibited mineralization levels 2.28‐fold and 2.70‐fold higher than Ti at days 14 and 21, respectively. These results collectively demonstrate the capacity of Ti‐Dual to promote the complete osteogenic differentiation cascade.

In Vivo Assay With Infected Bone Defect Model 7 Days Post‐Implantation

2.5

In this study, we selected P. aeruginosa to establish an infected bone defect model (Figure 5a), because P. aeruginosa is a common clinically drug‑resistant bacterium. It exhibits resistance to antibiotics including cefuroxime sodium, cefazolin, and vancomycin [40] (Table S2), thereby posing a serious threat in orthopedic clinical practice.

*In vivo assay with the infection model at 7 days post‐implantation. (a) Schematic illustration of the surgical procedure for establishing the infection model. (b) The images of the infected femoral sites. Yellow arrows indicated abscess formation. (c) Microbiological assessment by agar plate imprint method following implant retrieval. Quantification of bacterial loads (d) adherent on the screw and (e) in the surrounding peri‐implant tissue (n = 4). (f) MMP activity in peri‐implant tissues (n = 4). (g) H&E staining of the implant site (upper row: 10×, scale bar = 200 µm; lower row: 40×, scale bar = 50 µm). Yellow arrows highlighted inflammatory cells. (h) Quantitative analysis of inflammatory cells from 40× H&E sections (n = 6). (i) Immunofluorescence co‐staining of tissue adjacent to the implant for iNOS⁺ M1 macrophages (red), CD206⁺ M2 macrophages (green), and nuclei (DAPI, blue) (scale bar = 50 µm). The top row showed merged images, and the bottom row displayed individual channels. * denoted p <0.05, *denoted p <0.01, *** denoted p <0.001.

Following 7‐day implantation, control groups Ti and Ti‐S exhibited significant abscess formation (visible as yellow arrows in Figure 5b), indicating disseminated P. aeruginosa infection from bone marrow to peripheral tissues. In contrast, Ti‐Mono and Ti‐Dual maintained normal histological architecture. Agar plate assays demonstrated obvious bacterial clearance on Ti‐Dual surfaces versus confluent bacterial growth on controls (Figure 5c), correlating with a 94.23% reduction in surface bacterial load relative to Ti (Figure 5d). Crucially, Ti‐Dual reduced peri‐implant tissue bacterial loads by 98.28% (Figure 5e), effectively eradicating early‐stage infection. Although Ti‐Mono demonstrated antibacterial activity in vitro comparable to Ti‐Dual (Figure 2b), its in vivo efficacy was reduced. Ti‐Mono achieved only 80.77% bacterial reduction on screws and 62.07% in surrounding tissues. The superior in vivo antibacterial performance of Ti‐Dual was attributed to its retained RGD sequence following enzymatic cleavage. This preserved fragment promoted tissue repair, which synergistically enhanced infection suppression through immunomodulation [41, 42]. Conversely, enzymatic cleavage removed the RGD sequence from the Ti‐Mono surface, leaving only the antibacterial HHC36 sequence. Consequently, Ti‐Mono lacked the tissue‐repair synergy required for optimal infection control. Notably, Ti‐Dual‐N displayed minimal antibacterial activity, evidenced by 50.00% reduction on screws and 42.24% in tissues, aligning with prior in vitro results. MMP activity analysis provided a further assessment of gelatinase activity in the peri‐implant tissues. (Figure 5f). While Ti‐Dual‐N groups maintained MMP levels comparable to infected controls (Ti, Ti‐S), both Ti‐Dual and Ti‐Mono significantly suppressed peri‐implant MMP concentrations by effectively controlling infection.

Histopathological analysis identified significant inflammatory infiltration primarily composed of neutrophils and monocytes in Ti, Ti‐S, and Ti‐Dual‐N groups. In these groups, extensive inflammatory cell aggregation penetrated cortical bone and occupied the medullary cavity (Figure 5g). Ti‐Mono demonstrated substantially reduced inflammatory cell accumulation due to its intrinsic antibacterial performance. In contrast, Ti‐Dual exhibited sparse mononuclear cell infiltration, indicating effective bacterial clearance during early implantation that consequently minimized inflammatory responses. The limited inflammatory cells observed at the bone‐implant interface likely reflected either residual sterile inflammation from surgical trauma or normal postoperative responses at day 7. Most notably, Ti‐Dual significantly attenuated infection‐associated inflammation compared to control groups, confirming its antibacterial efficacy. Quantitative analysis showed an 89.79% reduction in medullary cavity inflammatory cells relative to Ti (Figure 5h), highlighting the critical role of endogenous responsive antibacterial strategies in preventing bacterial dissemination into bone marrow and soft tissues.

Immunofluorescence staining of iNOS (pro‐inflammatory, red) and CD206 (anti‐inflammatory, green) were performed on 7‐day post‐implantation sections to assess macrophage polarization (Figure 5i). As anticipated, antibacterial groups Ti‐Mono and Ti‐Dual exhibited predominant CD206⁺ anti‐inflammatory macrophage infiltration with minimal iNOS⁺ expression. It indicated effective bacterial clearance that reduced pathogen burden while preventing excessive immune activation. Importantly, these findings aligned with prior H&E observations (Figure 5g). Conversely, low‐antibacterial groups Ti, Ti‐S, and Ti‐Dual‐N displayed significantly enhanced iNOS⁺ pro‐inflammatory signaling due to uncontrolled infection. As reported, persistent microbial challenge could initiate pathological cascades where chronic macrophage and neutrophil activation caused massive pro‐inflammatory cytokine release. This cytokine surge would recruit additional inflammatory cells, ultimately causing severe impairment of bone regeneration.

In Vivo Assay With Infected Bone Defect Model 60 Days Post‐Implantation

2.6

At the 60‐day endpoint, macroscopic examination revealed pathological fractures with associated tissue abscess formation in both non‐antibacterial controls (Ti and Ti‐S) and the weakly antibacterial Ti‐Dual‐N (Figure 6a). While Ti‐Mono maintained femoral anatomical continuity, distinct mid‐shaft swelling was evident. In contrast, Ti‐Dual demonstrated optimal structural preservation, exhibiting intact cortical bone without anatomical abnormalities. Quantitative bacteriological analysis corroborated these findings, confirming the superior antibacterial efficacy of Ti‐Dual. This group achieved bacterial load reductions of 94.38% and 97.09% on the screw and surrounding tissue compared to the Ti and Ti‐S groups, respectively. Ti‐Mono exhibited moderate antibacterial efficacy with an 84.87% reduction rate. This indicated that the loss of the RGD sequence and their associated tissue‐repair function compromised synergistic bacterial suppression during tissue regeneration (Figure 6b). Notably, prolonged implantation to 60 days resulted in progressively increased bacterial counts and gelatinase concentrations within the non‐antibacterial groups compared to 7‐day results. This trend strongly suggests progression to chronic infection when early‐stage infection remains uncontrolled. Throughout the longitudinal assessment, gelatinase activity remained lower in both Ti‐Mono and Ti‐Dual groups than in other groups (Figure 6c).

*In vivo assay with the infection model at 7 days post‐implantation. (a) The images of the infected femoral sites. The yellow dashed line indicated the edge of the femur. (b) Quantification of bacterial loads adherent on the screw and in the surrounding peri‐implant tissue (n = 10). (c) MMP activity in explanted screws and peri‐implant tissues (n = 5). (d) H&E staining of the implant site. The image below was a magnification of the black‐boxed area in the image above. Blue arrows highlighted fibrosis, green arrows highlighted mild vascular hyperplasia, black arrows highlighted inflammatory infiltration, and yellow arrows highlighted osteolytic necrosis with mineralization. The scale bars were 1000 µm (top row), 100 µm (middle row), and 50 µm (bottom row). (e) Micro‐CT analysis of representative 3D reconstructed images of bone formation around the implant. The quantitative parameters from Micro‐CT assay: (f) bone surface density (BS/BV), (g) bone volume fraction (BV/TV), (h) trabecular thickness (Tb·Th), and (i) intersection surface (n = 3). (j) Undecalcified sections stained with methylene blue/basic fuchsin. The image below was a magnification of the red‐boxed area in the image above. The blue arrow indicated the unbonded region at the implant‐bone interface. The scale bars were 1000 µm (top row), 500 µm (middle row), and 100 µm (bottom row). (k) Immunofluorescence co‐staining of tissue adjacent to the implant for IL‐10⁺ (red), OPN⁺ (green), and DAPI‐stained nuclei (blue). The left column showed merged images, and the right column displayed individual channels (scale bar = 100 µm). (l) The images of TRAP staining (scale bar = 50 µm). * denoted p <0.05, *denoted p <0.01, *** denoted p <0.001.

Histopathological evaluation via H&E staining revealed profound inflammatory infiltration in non‐antibacterial groups (Ti, Ti‐S, and Ti‐Dual‐N). These specimens exhibited distinct inflammatory aggregates composed predominantly of neutrophils and monocytes that progressively disseminated into cortical bone and occupied nearly the entire medullary space (Figure 6d). Specifically, non‐antibacterial implants induced moderate fibrosis with mild vascular proliferation, along with severe inflammatory cell infiltration (including occasional giant cells) and intermediate osteolytic necrosis featuring focal mineralization. The significant tissue swelling and architectural distortion observed in these groups reflect the key pathological transition from acute to chronic osteomyelitis [43], a process fueled by bacterial biofilm formation that drives hematoma and tissue hyperplasia. In contrast, Ti‐Mono demonstrated significantly reduced inflammatory infiltration, confirming that the surface‐exposed HHC36 sequence exerted antibacterial effects to suppress inflammatory responses. Consistent with earlier observations, Ti‐Dual maintained normal histological architecture identical to 7‐day post‐implantation specimens, with marrow cavities containing primarily adipocytes and sparsely distributed mesenchymal stem cells without pathological alterations.

Micro‐CT analysis revealed substantial femoral fractures with surrounding tissue hematoma and malunion in non‐antibacterial Ti and Ti‐S groups after 60 days of infection. Comparable pathological manifestations appeared in the Ti‐Dual‐N group (Figure 6e). In contrast, both Ti‐Mono and Ti‐Dual groups exhibited minimal fracture incidence. However, Ti‐Mono displayed minor hematoma formation, consistent with its relatively lower antibacterial efficacy compared to Ti‐Dual. Further evaluation of bone microarchitecture demonstrated Ti‐Dual's superiority in peri‐implant bone density (BS/BV), bone regeneration capacity, and osseointegration (Figure 6f). Quantitative analysis confirmed significantly higher BV/TV values for Ti‐Dual compared to other groups, measuring 1.18‐, 1.31‐, 1.18‐, and 1.43‐fold greater than Ti, Ti‐S, Ti‐Mono, and Ti‐Dual‐N groups respectively (Figure 6g). Ti‐Dual also exhibited enhanced trabecular parameters including Tb. Th and Tb. N (Figure 6h; Figure S22), along with a 1.33‐, 1.64‐, 1.37‐, and 1.89‐fold larger interfacial contact area versus other groups (Figure 6i). These findings collectively demonstrate that the fracture resistance in antibacterial screw groups originated from Ti‐Dual's exceptional infection control. This dual‐functional implant effectively prevented infection‐mediated bone destruction while demonstrating superior post‐infection bone healing and regenerative capacity.

Methylene blue‐basic fuchsin staining results (Figure 6j) revealed a critical distinction that Ti‐Dual maintained intact cortical architecture with physiologically distributed yellow marrow in the medullary cavity, whereas control groups (Ti, Ti‐S, and Ti‐Dual‐N) exclusively exhibited ectopic yellow marrow deposition in neoformed malunion sites. This pathological pattern indicated clear histological evidence of heterotopic ossification. Comparative analysis further demonstrated that although both Ti‐Mono and Ti‐Dual showed bactericidal activity against P. aeruginosa, relatively higher peri‐implant bacterial loads in the Ti‐Mono group (Figure 2b) directly correlated with localized hematoma formation. Quantitative measurements at bone‐implant interfaces revealed Ti‐Dual's superior directional osteogenesis capacity. The osteogenic indices for Ti‐Dual were 37.90‐, 26.32‐, 8.09‐, and 19.89‐fold greater than those of Ti, Ti‐S, Ti‐Mono, and Ti‐Dual‐N groups, respectively (Figure S23).

Immunofluorescence analysis revealed that Ti‐Dual, Ti‐Mono, and Ti‐Dual‐N groups exhibited positive IL‐10 signals, whereas minimal expression was observed in Ti and Ti‐S groups (Figure 6k). The detected IL‐10 in the positive groups likely originated from mesenchymal stem cells [44] or M2 reparative macrophages [45, 46], consistent with established literature documenting IL‐10's dual regulatory functions. This anti‐inflammatory cytokine suppressed osteoclast activity while promoting osteogenic differentiation, playing a critical role in post‐infection bone repair. Consequently, these findings confirmed that surface‐modified peptides participated in immunomodulatory responses following infection while concurrently facilitating bone regeneration. Additionally, osteopontin (OPN) expression was significantly elevated in both Ti‐Mono and Ti‐Dual groups (Figure 6k). Previous studies indicated OPN promoted bone repair primarily through integrin αvβ3 and CD44 receptor interactions that coordinately regulate osteoblast‐osteoclast activity [47]. During late‐stage repair, OPN critically enhances osteogenic differentiation and matrix mineralization [48]. Consequently, the elevated OPN levels in Ti‐Mono and Ti‐Dual directly correlated with their superior antibacterial efficacy and preserved osteogenic capacity within defect regions. Crucially, Ti‐Dual demonstrated maximal OPN expression among all groups due to RGD retention, while Ti‐Mono exhibited moderate reduction attributable to RGD cleavage. TRAP staining results (Figure 6l) revealed that Ti and Ti‐S exhibited extensive TRAP‐positive osteoclast infiltration at bone‐implant interfaces, directly induced by bacterial virulence factors within the infected microenvironment. Ti‐Dual‐N demonstrated moderate osteoclast activity localized within hematoma regions. In contrast, both Ti‐Mono and Ti‐Dual groups showed a complete absence of TRAP staining, indicating effective suppression of osteoclast activation. These observations were consistent with MMP activity measurements, revealing a pathogenic sequence where chronic infection triggered increased gelatinase production, promoted pro‐inflammatory macrophage polarization, and ultimately led to excessive bone resorption and elevated fracture risk.

Discussion

3

Bone implant infection represents a major cause of implant failure, primarily resulting from intraoperative contamination, hematogenous dissemination, and subsequent biofilm formation. In severe cases, such infections may progress to chronic osteomyelitis or sepsis, and when uncontrolled, may ultimately require amputation to prevent life‐threatening systemic dissemination.

To address implant‐associated infections in orthopedics, this study developed a dual‐anchored peptide strategy. The designed material responds to gelatinase secreted by common clinical pathogens such as S. aureus and P. aeruginosa, inducing a conformational change in surface‐tethered peptides that switches their antibacterial function on or off. In the absence of infection, the antibacterial activity of the HHC36 sequence on the Ti‐Dual surface remains suppressed due to conformational constraint, thereby promoting stem cell proliferation (Figure 3b), differentiation (Figure 4), and migration (Figure 3f,g). Upon infection, the material rapidly responds to bacterial gelatinase, activating potent antibacterial properties. Ti‐Dual eliminated 99.37% of P. aeruginosa within 10 min (Figure 2c), and 99.73% of S. aureus along with 99.99% of P. aeruginosa within 120 min (Figure 2b). Such rapid bactericidal efficacy is critical for early‐stage infection control, as supported by previous studies [49, 50]. Timely eradication of planktonic bacteria can prevent biofilm formation and subsequent antibiotic resistance, attenuate excessive inflammatory responses, limit the spread of infection, and reduce the risk of severe complications such as osteomyelitis and sepsis.

Compared to the mixed grafting surface, Ti‐Dual offers two distinct advantages. First, the dual‐anchored enzyme‐responsive design enables selective activation of antibacterial properties through controlled conformational changes in the antimicrobial peptide. This mechanism prevents the cytotoxicity associated with strong antibacterial activity under normal physiological conditions. Second, the fusion peptide strategy facilitates the stable incorporation of peptides that are typically challenging to graft—such as RGD—while the introduction of a cationic antimicrobial peptide further enhances the grafting efficiency onto the surface [51] (Figure S14). As a result, both the peptide grafting efficiency and the overall biological functionality of the surface are significantly improved (Figure 3b).

Notably, compared to conventional enzyme‐responsive strategies such as the Ti‐Mono in this study, Ti‐Dual retains its capacity to promote cell adhesion and proliferation even after enzymatic activation and antibacterial function are triggered (Figure 3b,i–k; Figure S19). This advantage becomes more pronounced under in vivo conditions. Although Ti‐Mono and Ti‐Dual exhibit comparable antibacterial efficacy in vitro, their performance diverges significantly under the complex demands of the in vivo microenvironment, which requires a synergistic combination of antibacterial action and tissue integration. Following bacterial elimination, Ti‐Mono loses its RGD motif and thus the ability to promote tissue integration. This impairment in regenerative capacity ultimately compromises its overall antibacterial outcome in vivo. In contrast, Ti‐Dual preserves both potent antibacterial performance and the pro‐regenerative function of RGD. Numerous studies have demonstrated a close synergistic relationship between antibacterial function and tissue regeneration. Advanced biomaterials with excellent pro‐regenerative properties typically exhibit favorable immunomodulatory capabilities. By modulating macrophage polarization and other pathways, they effectively mitigate inflammation and support bacterial clearance, while simultaneously fostering a pro‐regenerative microenvironment that synergistically enhances high‐quality tissue repair. In accordance with this mechanism, Ti‐Dual induced elevated expression of CD206—an anti‐inflammatory macrophage marker—in the surrounding tissue (Figure 5i), contributing to effective inflammation control. Furthermore, in synergy with IL‐10, Ti‐Dual promoted significantly improved bone regeneration outcomes (Figure 6k). The existence of such synergy underscores the importance of maintaining the pro‐regenerative performance of Ti‐Dual.

Cumulatively, as shown by the results of in vivo experiments, Ti‐Dual demonstrated comprehensive therapeutic efficacy through multiple integrated mechanisms. First, its rapid bactericidal action achieved 94.23–98.28% bacterial reduction during early infection stages, while localized IL‐10 secretion simultaneously attenuated inflammatory responses (89.79% reduction compared to controls) and further inhibited osteoclast precursor differentiation. Moreover, the material maintained physiological bone remodeling by not only suppressing pathological resorption (as evidenced by TRAP‐negative phenotype) but also enhancing osteogenic activity (1.18‐fold increase in BV/TV). Furthermore, sustained OPN expression promoted osteoblast differentiation and matrix mineralization, ultimately resulting in superior osseointegration (1.33‐fold increase in intersection surface vs. Ti). Taken together, this multifunctional platform addresses the dual challenges of microbial eradication and tissue regeneration, showing significant promise for clinical management of infected bone defects through its ability to concurrently control infection, modulate inflammation, and promote bone repair. The coordinated action of these mechanisms establishes Ti‐Dual as a comprehensive solution for complex bone infection scenarios.

Conclusion

4

In conclusion, we developed a surface by dual‐anchoring a gelatinase‐responsive fusion peptide. Ti‐Dual exhibits superior antibacterial activity and antibiofilm formation against both S. aureus and P. aeruginosa. The engineered fusion peptides can enhance the grafting density of RGD on the surface, promoting cell adhesion, proliferation, and osteogenic differentiation under a dual‐anchored conformation. In the presence of infection, the surface can facilitate early bactericidal activity, demonstrating better post‐infection tissue repair compared to Ti‐Mono. For femoral infections caused by the multidrug‐resistant ESKAPE bacteria P. aeruginosa, Ti‐Dual can timely kill bacteria, prevent the destruction of bone homeostasis by gelatinase, restore the normal microenvironment, thereby favouring bone reconstruction and repair.

Experimental Section

5

Peptide Surface Grafting

5.1

A 10 × 10 mm^2^ titanium (Ti) plates were ultrasonically cleaned successively in absolute ethanol and water. For alkaline‐heat treatment, the plates were immersed with 5 m NaOH at 60°C for 24 h. Then, the plates were incubated with silicon coupling agent containing DBCO group (Sialane–PEG_3.4k_ ‐DBCO, 0.5 mg mL ^−1^, dissolved in 95% ethanol) at room temperature for 24 h. After solidifying at 100°C for 30 min, the plates were reacted with the peptide solution for 6 h at room temperature, and deionized water (DIW) was added for ultrasonic cleaning for 10 min, during which DIW was replaced three times. The surfaces labeled Ti‐Dual, Ti‐Mono, and Ti‐Dual‐N were prepared by reacting with 50 µm Dual, 30 µm Mono, and 30 µm Dual‐N peptide solutions, respectively. The mix surfaces are defined as follows: Ti‐H (25 µm HHC36), Ti‐R (400 µm RGD, low grafting), Ti‐HR (25 µm HHC36 + attempted 400 µm RGD).

Micro‐BCA

5.2

The samples were immersed in 150 µL of DIW and added to the Micro‐BCA working solution of equal volume. The working solution was prepared as solution A:B:C = 25:24:1. After incubation at 37°C for 2 h, 100 µL of the mixed solution was taken to detect the OD value of 562 nm. Different concentrations of the indicated peptide were set to calculate the peptide density of the surface.

XPS

5.3

Samples were characterized by XPS to detect the valence and content of elements. The acquisition parameters were set as described: source gun type was Al K Alpha, spot size was 650 µm, CAE: Pass Energy 100.0 eV, and energy step size was 1.0 eV. High‐resolution narrow scans were recorded with a pass energy of 30.0 eV and a step size of 0.10 eV. The data was standardized by the C─C (284.8 eV).

HRMS

5.4

The peptide and enzymatic hydrolysates were analyzed by electrospray ionization mass spectrometry (ESI‐MS) using a Q Exactive Plus mass spectrometer (Thermo Fisher Scientific, USA). For enzymatic digestion (marked with a suffix of “‐Gel”), peptides (0.25 mm) were treated with gelatinase (20 U mL^−1^, 37°C, 1 h), followed by heat inactivation (60°C, 30 min) prior to MS detection.

Contact Angle

5.5

The water contact angle was characterized via the sessile drop method by Drop Shape Analyzer (DSA 100E, KRUSS, Germany). 2 µL of DIW was dropped on the surfaces, and the image taken at 2.00 s was selected for contact Angle analysis.

In Vitro Antibacterial Assay

5.6

S. aureus (ATCC6538p) and P. aeruginosa (ATCC27853) were diluted to 10^6^ CFU mL^−1^ with PBS, 10 µL of which was dropped onto the surface of Ti disks and incubated at 37°C for 10, 20, 30, 60, and 120 min. Bacteria on the surface were dissociated by vortexing, and colony counts were performed by the dilution spread plate method. E. coli (ATCC25922) and L. casei (CCTCC AB 2013355) were cocultured for 120 min. The same incubation conditions were used for bacterial morphology. The bacteria were fixed, dehydrated, and sputtering Au coating before observation by scanning electron microscope (SEM, Merlin, Zeiss, Germany).

Antibacterial Mechanism

5.7

NPN was used as a fluorescent probe for the detection of outer membrane permeability of Gram‐negative bacteria. First, for bacterial stabilization, P. aeruginosa was diluted to 10^6^ CFU mL^−^ ^1^ in HEPES solution (pH 7.4, containing 20 mm glucose), incubated at 37°C in the dark for 30 min, then centrifuged at 8000 rpm for 3 min. After resuspension, 10 µL of bacteria with 10 µM NPN was added to the sample surface, sealed with plastic wrap, and incubated for 2 h. Following staining, the samples were rinsed gently with PBS to remove unbound dye. The MFI value of the sample surface was characterized by a fluorescence microscope (DM6M, Leica, Germany).

To observe the status of bacteria, the samples incubated for 2 h can be stained by fluorescein diacetate (FDA) and pyridine iodide (PI). 100 µL of 0.05%FDA solution and 100 µL of 0.01%PI was added to incubate at 37°C for 5 min, respectively. Following staining, the samples were rinsed gently with PBS to remove unbound dye.

Anti‐Biofilm Formation

5.8

Tryptone soy broth (TSB) provided richer carbon and nitrogen sources, which favored bacterial biofilm formation. S. aureus and P. aeruginosa were centrifuged and resuspended in TSB to obtain a bacterial concentration of 10^9^ CFU mL^−1^. 1 mL of TSB and 10 µL bacteria suspension were added to immerse the samples. Sterile water was added to the surrounding pores to maintain a moist environment. The wells were covered with a layer of plastic wrap and incubated at 37°C for 36 h. After 24 h of incubation, 500 µL of TSB was replaced with fresh TSB for biofilm incubation.

Methanol‐dissolved 0.5% crystal violet solution was prepared and diluted to 0.1% in sterile PBS for biofilm staining. After the incubation for 36 h, the bacterial solution was gently removed, and the samples were rinsed gently twice. 300 µL of crystal violet solution was added to each well and stained at 37°C in the dark for 10 min. The samples were washed twice with PBS to remove excess crystal violet and then transferred to a new 24‐well plate. Light field observations were performed with a Leica upright microscope. The biofilm image was observed by a metalloscope (DM6M, Leica, Germany). After filming, 300 µL of absolute ethanol was added to each well to dissolve crystal violet, and the biofilm was completely detached and dissolved by sonication for 5 min. 100 µL of which was taken for OD value measurement at 590 nm for quantitative analysis of biofilm.

Cell Adhesion and Morphology

5.9

A 10 × 10 × 0.1 mm^3^ Ti plate was flattened to avoid changes in surface curvature affecting cell proliferation and growth. The surface biocompatibility of Ti, Ti‐S, Ti‐Dual, Ti‐Mono, Ti‐Dual‐N, and surface treated with 1 U mL^−1^ gelatinase (gel‐Ti‐Dual and gel‐Ti‐Mono) were tested. All samples were sterilized by UV irradiation overnight, and rinsed with PBS and medium (α‐MEM, 10%FBS). 10000 mBMSCs were seeded in each well and incubated at 37°C, 5% CO_2_ for 1 and 5 days. After 3 days, 1 mL of fresh medium was replaced slowly for further incubation. After the incubation, the Ti plates were transferred to a new 24‐well plate, and 320 µL of the CCK8 working solution (medium: CCK8 solution = 10:1) was added to each well, 100 µL of which was tested at 450 nm with a microreader (VarioskanFlash 3001, Thermo) after 3 h of incubation at 37°C.

After being cultured with mBMSCs (1.0 × 10^4^ cells) for the indicated times, the cells were rinsed with PBS and fixed with 4% paraformaldehyde. Then, 0.1% Triton X‐100 was used to permeabilize the cells for 8 min, and rinsed three times with PBS. Finally, the cells were stained with Cell Navigator F‐Actin Labeling kit (AAT Bioquest Inc.) for 60 min at 37°C and counterstained with 4′,6‐diamidino‐2‐phenylindole hydrochloride (DAPI, Beyotime, China) for 5 min. The samples were observed via microscope (Leica DM6M).

Cell Migration

5.10

mBMSCs (4 × 10^4^ cells) were seeded on the surface of 10 × 10 × 0.1 mm^3^ sample, and the cell density was observed after 24 h of incubation. When the cell density was close to 80%, the cells were scratched (650 µm) with 10µL‐tip, washed with PBS, and incubated with 1%FBS medium. Cells were fixed after migration for 24 h, and the migration distance was observed by cytoskeleton staining described above.

Cell‐Bacteria Coculture

5.11

mBMSCs (1.0 × 10^4^ cells) were incubated with the samples for 2 days, then 10 µL of 10^9^ CFU mL^−1^ P. aeruginosa was added and incubated in a 5% CO_2_ incubator for 24 h. The cells were washed twice with PBS and fixed with 4% paraformaldehyde at 4°C overnight. After fixation, triton‐X was permeabilized for 8 min before cytoskeleton staining. Furthermore, 50 µL of cell culture medium was taken for the MMP activity assay after 24 h of incubation. Similarly, we performed the same experiment using L929 cells, with consistent cell numbers, bacterial concentration, and culture time.

PCR

5.12

mBMSCs (1.0 × 10^4^ cells, passage 4 to 6) were seeded on different surfaces (10 × 10 × 0.1 mm^3^). After 3 days of mBMSCs culture, the complete medium was changed to osteoinductive medium (containing 10% FBS, 50 mm ascorbic acid, 10 mm β‐glycerophosphoric acid, and 0.1 mm dexamethasone), which was replaced every 3 days. The cell culture procedures of ALP and ARS were consistent with the descriptions provided above. After 7, 14 days of osteogenic culture, the samples were transferred to a new 24‐well plate, and the RNA was extracted by Beyotime RNAeasy Animal RNA Extraction kit (centrifugal column). Based on the purity and appropriate concentration of extracted RNA, reverse transcription was performed using TAKARA PrimeScript RT reagent Kit with gDNA Eraser (RR047A). All‐in‐OneTM qPCR Mix (QP001 Gene Copoeia, USA) was used to determine the gene expression level, and the primer sequence are displayed in Table S1. Gene expression levels were normalized to GAPDH expression and analyzed according to the comparative Ct (2^−ΔΔCt^) method.

ALP Activity

5.13

For ALP staining, after incubation for 7, 14, and 21 days, the medium was removed and rinsed twice with PBS, and then fixed in the dark with 350 µL 4% paraformaldehyde (4°C, overnight). Then, the samples were washed twice with PBS, and 350 µL ALP staining solution (BCIP/NBT) was added to each well for 60 min. After washing three times with PBS, the samples were observed by a stereomicroscope through the bright field channel.

For ALP quantification, the samples were transferred to a new well plate, rinsed twice with PBS, in which 350 µL of Western&IP cell lysate (without inhibitor) were added and lysed for 2 h at 4°C. Subsequently, the surface was blown vigorously to fully lyse the cells, and the lysate was collected in 1.5 mL EP tubes and centrifuged at 12000 rpm for 3 min at 4°C. The supernatant was collected for ALP activity detection by Alkaline Phosphatase Assay Kit (Beyotime, China).

Alizarin red staning

5.14

For alizarin red staining, after incubation for 14 and 21 days, the samples were transferred to a new well plate, rinsed twice with PBS, and fixed with 4% paraformaldehyde for 30 min. Then, 300 µL of alizarin red staining solution was added and incubated at room temperature in the dark for 15 min. The samples were washed with ultrapure water to remove excess dye solution, and repeated multiple times. The samples were photographed using a stereomicroscope. To quantify the amount of calcium nodules formed on the different functionalized surfaces, 300 µL 10 wt.% chlordodecyl pyridine was added to each well to dissolve the calcium nodules on the surface, and 100 µL of the solution was measured at 562 nm.

Immunofluorescence

5.15

mBMSCs (2.5 × 10^4^ cells) cultured on different surfaces (10 × 10 × 0.1 mm^3^) were gently rinsed with PBS and fixed with 4% paraformaldehyde for 10 min. Saponin was used to permeabilize the cells for 10 min. Then, the cells were rinsed three times with the PBS solution and blocked with 3% bovine serum albumin (BSA). 350 µL of primary antibody solution (1:200 in 1%BSA solution, Affinity Biosciences, USA) was added into each well and incubated at 4°C overnight. Primary antibodies against Vinculin, RUNX2, and OPN were used on 1‐, 7‐, and 14‐day samples, respectively. The 7‐day and 14‐day samples were incubated with osteogenic‐induction medium. The samples were rinsed with PBST for 5 min and repeated three times. Next, 350 µL of the conjugated secondary antibody (Goat Anti‐Rabbit IgG H&L Alexa Fluor 488, Abcam, UK). The samples were rinsed with PBST for 5 min and repeated three times. Finally, the cells were stained with YF594‐Phalloidin (UElandy Inc.) for 60 min at 37°C and counterstained with 4′,6‐diamidino‐2‐phenylindole hydrochloride (DAPI, Beyotime, China) for 5 min. The samples were observed via fluorescence microscope (DM6M, Leica, Germany). The negative control refers to the sample group that was incubated with the secondary antibody only, without the primary antibody.

In Vivo Experiments

5.16

In vivo experiments were performed in the Second Affiliated Hospital of Chongqing Medical University (IACUC‐SAHCQMU‐2025‐0120). Male Sprague‐Dawley rats (8‐10 weeks, 300–350 g) were anesthetized by intraperitoneal injection of sodium pentobarbital. After the rats were anesthetized, their left thigh was shaved and disinfected with iodophor. A skin incision of about 1.5‐2.0 cm in length was made in the middle of the femur using a surgical blade, and the muscle was bluntly separated to expose the femur. A 1.8‐mm diameter ball‐milled bone drill was used to drill through the mid‐femur, with a spacing of approximately 5 mm between the two holes (cooling by continuous saline perfusion during drilling). After cleaning with normal saline, a cotton swab (1 mm in diameter) immersed in 10^8^ CFU mL^−1^ P. aeruginosa bacteria solution was screwed into the hole and coated. A total of 3 cotton swabs were used for repeated coating three times, and a screw (2.0 mm in thread diameter, 5.0 mm in total length, 3.5 mm in thread part) was screwed into the hole. After the completion of implantation, the suture was performed layer by layer.

Statistical Analysis

5.17

Data were presented as mean ±SD (n ≥3). Statistical significance was determined using two‐tailed Student's t‐test in Excel and Origin 8.0.2, with ^^ p <0.05, ^^ p <0.01, ^^ p <0.001 considered significant, and NS indicating non‐significant results.

Conflicts of Interest

Yingjun Wang, Lin Wang, and Ru Zhong have applied for a patent pending to South China University of Technology.

Supporting information

Supporting file: advs73592‐sup‐0001‐SuppMat.docx