A Novel Strategy for Achieving Immunomodulation and Osseointegration of Titanium Alloys: Boron‐Doped Porous Coating Modification

ABSTRACT

Durable performance of metallic implants depends largely on robust osseointegration and a balanced immune microenvironment at the bone‐implant interface. Modulating macrophage polarization and osteoblast responses through surface topography and chemistry has emerged as an effective strategy for improving titanium alloy implants. In this study, boron‐containing hierarchical porous ceramic coatings were fabricated on the low‐modulus Ti‐19Zr‐10Nb‐1Fe alloy using a one‐step micro‐arc oxidation process. The pore topography, surface roughness, wettability, and boron release kinetics were regulated by adjusting the applied voltage. Among the tested coatings, the M300V exhibited stronger coating adhesion and improved corrosion resistance. In vitro, the M300V coating promoted osteoblast proliferation and differentiation and induced macrophage polarization toward an M2‐like phenotype. Macrophage‐conditioned medium from the M300V group further enhanced osteoblast differentiation, indicating a beneficial immunomodulatory effect on osteogenesis. These responses were accompanied by increased expression of osteogenesis‐related markers, including RUNX2, COL1, and OCN. In a rat femoral condyle model, the M300V coating reduced early inflammatory responses, suppressed osteoclast activity, and improved osseointegration. These findings suggest that the improved biological performance of the M300V coating may result from the combined effects of favorable boron release kinetics and hierarchical porous microstructures, providing a promising surface modification strategy for bone‐interfacing titanium alloy implants.

A boron‐doped hierarchical porous coating is engineered on low‐modulus Ti‐19Zr‐10Nb‐1Fe by one‐step micro‐arc oxidation, enabling tunable surface features and boron release. The optimized coating simultaneously attenuates inflammation, promotes M2 macrophage polarization and osteogenesis, and enhances osseointegration in vivo, highlighting a promising immunoregulatory strategy for titanium implants.

1. Introduction

The global population aging trend has increased the urgent demand for surgical procedures for bone repair and regeneration, which places substantial pressure on healthcare systems [1]. Owing to their favorable mechanical properties and biocompatibility, traditional pure titanium and its alloys have been widely used as dental and orthopedic implants [2]. However, surface bioinertness of titanium alloys remains a major limitation because they lack spontaneous biological functionalization. To address these limitations, diverse surface‐engineering strategies have been developed, such as laser texturing [3, 4], anodization [5, 6], and coatings containing bioactive molecules or drugs [7, 8]. Among these strategies, micro‐arc oxidation (MAO) generates ceramic coatings through high‐voltage microdischarges, thereby enabling simultaneous tuning of surface topography and chemistry to promote bioactivity [9, 10]. Conventional phosphate‐ or silicate‐based electrolytes typically yield P‐ or Si‐containing coatings, yet the resulting “crater‐like” pores are discretely distributed, with relatively uniform morphologies and limited tunability in pore size and interconnectivity [11, 12]. By contrast, tetraborate electrolytes can generate distinct morphologies across different substrates. For example, Li et al. reported a “cortex‐like” texture on CP‐Ti [13], while Guan et al. obtained micro‐grooves and micro‐pores on Ti‐6Al‐4 V using sodium tetraborate and sodium hydroxide [14]. These features have been attributed to the intense discharges induced by tetraborate and its dissolution effect on the oxide layer [15, 16].

Osteoimmunology holds that the protein corona formed immediately after implantation, together with the initial innate immune response, strongly influences subsequent net bone formation [17]. Interfacial pore size, connectivity, wettability, and roughness first modulate the protein‐adsorption profile and cell‐adhesion pathways [18]. Submicron–nanoscale features promote integrin clustering and focal‐adhesion assembly, whereas micron‐scale interconnected porosity stabilizes the fibrin clot and facilitates cell migration [19, 20]. Through these corona‐mediated and adhesion/mechanotransductive cues, hierarchical microstructures further govern macrophage polarization and the proliferation and differentiation of osteogenic and endothelial cells [21, 22].

As an essential trace element for skeletal homeostasis [23], boron (B) promotes osteogenesis and mineralization by upregulating COL1 and OCN expression and increasing ALP activity and calcium deposition [24, 25]. In parallel, borate suppresses NF‐κB signaling, thereby limiting M1 polarization, promoting M2 polarization, and increasing IL‐10 expression [26, 27]. Together with VEGF upregulation, these effects initiate early angiogenesis and accelerate microenvironmental repair [28, 29]. In addition, boron‐doped hydroxyapatite and boric acid inhibit osteoclast differentiation and resorptive activity and may interact with steroid hormones to reduce calcium loss, thereby mitigating inflammation‐associated bone loss [30, 31].

The recently developed Ti‐19Zr‐10Nb‐1Fe (TZNF) alloy is promising because of its low elastic modulus (60–65 GPa), adequate strength and toughness, and absence of cytotoxic elements [32, 33]. In our previous work, MAO in a tetraborate‐based electrolyte was used to fabricate coatings with a honeycomb‐like hierarchical porous structure on TZNF alloy surfaces. The growth mechanism of MAO coatings on TZNF and the correlations between processing parameters, pore size, and coating thickness were systematically investigated, while the mechanical properties of the substrate were well retained [34]. In this work, a one‐step MAO process was employed to simultaneously incorporate boron and control surface topography by tuning electrical parameters, with the aim of combining these tailored features to promote immunomodulation and osteogenesis. MC3T3‐E1 osteoblastic cells and RAW264.7 macrophages were used to evaluate the pro‐osteogenic and anti‐inflammatory performance of the coatings in vitro, and their osseointegration capacity was further assessed in vivo to clarify the underlying mechanisms (Scheme 1). This study provides a new strategy for surface modification of TZNF alloy and reveals the dual role of boron‐doped porous coatings in promoting osseointegration and modulating immune responses, offering guidance for the design and clinical translation of hard tissue implants.

SCHEME 1.

Schematic illustration of TZNF with boron‐doped hierarchical porous ceramic coating. By combining a porous morphology with the release of B₄O₇ ^{2 −}, this strategy effectively modulates the early macrophage‐mediated inflammatory microenvironment and accelerates implant osseointegration.

2. Experimental Methods

2.1. Preparation of Coatings

The MAO was performed on TZNF alloy (at%: Ti 68.75, Zr 19.71, Nb 10.22, Fe 1.31). Plate (10×10×2 mm) and cylindrical (Ø2.2 × 5 mm) coupons were mirror‐polished and ultrasonically cleaned in DI water and ethanol. According to our previous work [34], the samples were used as anodes, with stainless steel as the cathode, in 0.1 M Na₂B₄O₇·10H₂O electrolyte. Pulsed MAO was applied for 3 min at 600 Hz and 15% duty cycle at 300, 400, or 500 V to obtain M300V, M400V, and M500V, respectively. Specimens were rinsed with DI water and air‐dried.

2.2. Surface Characterization

Surface morphology was observed using field‐emission scanning electron microscopy (FE‐SEM, Zeiss G450, Germany, 15 kV). Pore sizes were measured with ImageJ v1.53 software. For cross sectional morphology, samples were embedded in epoxy resin and polished. Surface roughness and three‐dimensional topography were assessed by atomic force microscopy (AFM, Bruker Dimension Icon, Germany, contact mode, scan area 30 µm×30 µm). Chemical composition was analyzed by energy‐dispersive spectroscopy (EDS, Oxford Ultim Extreme, UK). Before SEM observation, samples were sputter‐coated with gold.

2.3. Physicochemical Properties

Wettability was measured using a Kruss DSA30 contact angle analyzer (Germany). A 3 µL droplet was dispensed at 0.5 µL/s, and images were recorded after 5 s (n = 3). Boron release was evaluated by immersing samples (1.25 cm²/mL) in α‐MEM (α‐Minimum Essential Medium, Gibco, USA) supplemented with 10% fetal bovine serum (FBS, Ausgenex PTY LTD, Australia) and 1% penicillin–streptomycin (P/S, Gibco, USA) at 37°C for 1–28 days. After digestion with nitric acid, boron concentrations were quantified by inductively coupled plasma optical emission spectrometry (ICP‐MS, Agilent 7850, USA). Protein adsorption was measured by incubating samples in FBS for 12 h, followed by desorption in 0.1% sodium dodecyl sulfate (SDS, Solarbio, China)/phosphate‐buffered saline (PBS, Phygene, China) for 15 min, and quantification with the BCA assay kit (Dingguo, China). Adhesion strength was determined with an automatic scratch tester (WS‐2005, China) at a scratching speed of 10 mm/min, a loading rate of 30 N/min, and a maximum load of 70 N (n = 3). Vickers hardness was measured with a microhardness tester (HXD‐1000TM, China) according to GB/T 4342‐91, using a 300 gf load held for 10 s (n = 15). Apatite‐forming ability was assessed by immersing the samples in 50 mL of simulated body fluid (SBF, Phygene, China) at 37°C under static conditions in conical centrifuge tubes. After 14 or 28 days, the samples were removed, rinsed twice with fresh ultrapure water, air‐dried, and analyzed by FE‐SEM and x‐ray diffraction (XRD, Rigaku, Japan).

2.4. Corrosion Performance

Electrochemical tests were conducted using a CHI660e workstation in artificial saliva (Phygene, China) at 37°C. A three‐electrode system was used: a platinum plate (10 mm×10 mm) as the counter electrode, a saturated calomel electrode (SCE) as the reference electrode, and the sample as the working electrode (exposed area 1 cm²). Open‐circuit potential (OCP) was monitored for 2 h. Electrochemical impedance spectroscopy (EIS) was performed at OCP with a 5 mV amplitude over the frequency range of 0.01 Hz–100 kHz, and data was fitted using ZView software. Potentiodynamic polarization (PDP) scans were performed from −1 to +1 V at 1 mV/s. Polarization resistance (R _p) was calculated using the Stern–Geary equation (Equation 1). Each group was tested 3 times in fresh solution.

$$R_{\mathrm{p}}=\frac{{\beta }_a{\beta }_b}{2.303I_{\mathrm{corr}}\left({\beta }_a+{\beta }_b\right)}$$ (1)

2.5. In Vitro Experiments

2.5.1. Cell Culture

All experiments were performed under biosafety level 2 (BSL‐2) conditions. The murine macrophage cell line RAW264.7 (SCSP‐5036) and pre‐osteoblastic cell line MC3T3‐E1 Subclone 14 (SCSP‐5218) were cultured in Dulbecco's modified Eagle's medium (DMEM, Genview, China) and α‐MEM, respectively, each supplemented with 10% FBS and 1% P/S. Cells were maintained at 37°C with 5% CO₂, and the medium was refreshed every 2 days. Cells of passages 3–5 were used for experiments. Samples were sterilized by autoclaving and dried in a laminar flow cabinet before use.

2.5.2. Cellular Performance Evaluation

Different seeding densities were used depending on the experimental endpoint. For adhesion and cytotoxicity assays, RAW264.7 cells were seeded at 2×10⁵ cells/mL and MC3T3‐E1 cells at 3×10⁴ cells/mL. For proliferation and differentiation assays, RAW264.7 cells were seeded at 1×10⁵ cells/mL and MC3T3‐E1 cells at 5×10³ cells/mL. Cell viability was assessed using the Calcein‐AM/PI staining kit (Beyotime, China). Proliferation was quantified with the CCK‐8 assay (Genview, China) at 450 nm (Multiskan GO, Thermo Scientific, USA) using five replicate wells per group, and the experiment was repeated 3 times. For cell adhesion morphology analysis, cells were fixed in 4% paraformaldehyde (Biosharp, China), stained with FITC‐phalloidin (Solarbio, China) for F‐actin and DAPI (Sigma‐Aldrich, USA) for nuclei, and imaged with an Axioscope 5 microscope (Zeiss, Germany). For SEM observation, samples were dehydrated in graded ethanol (30%–100%) and tert‐butanol (25%–100%), sputter‐coated with gold, and examined by FE‐SEM.

For immunofluorescence staining. RAW264.7 cells were seeded onto sterilized samples at a density of 2×10⁴ cells/cmofluorescence stain3 days. Cells were then fixed with 4% paraformaldehyde at room temperature for 10 min, treated with PBS containing 0.1% Tween‐20, and permeabilized with 5% Triton X‐100 for 10 min. Blocking was performed with BSA (Solarbio, China) at room temperature for 1 h. Samples were subsequently incubated overnight at 4°C with rabbit anti‐iNOS (Proteintech, China) and mouse anti‐Arg‐1 (Proteintech, China) primary antibodies. On the following day, samples were incubated with fluorescent secondary antibodies diluted 1:200, including anti‐rabbit DyLight 488 and anti‐mouse DyLight 594 (Zhongshan, China), at 37°C for 1 h in the dark. Nuclei were then stained with DAPI (Solarbio, China) for 5 min at room temperature, followed by the addition of an antifade mounting medium (Solarbio, China). Fluorescence images were acquired using an upright fluorescence microscope (Zeiss, Germany), and random fields were analyzed for fluorescence intensity.

Gene expression was analyzed by RT‐qPCR. RAW264.7 cells were cultured for 3 days, and MC3T3‐E1 cells for 7 and 14 days. RNA was extracted with TRIzol (Invitrogen, USA), reverse‐transcribed using the QuantScript kit (Tiangen, China), and amplified with RealStar SYBR (GenStar, China) premix on an ABI 7500 system. Relative expression levels were calculated using the ΔΔCt method with GAPDH as the reference gene. Target genes included inflammatory markers (TNF‐α, IL‐1β, IL‐6, IL‐10, and Arg1) and osteogenic markers (ALP, RUNX2, OPN, COL1, OSX, and OCN). Primer sequences are listed in Table S1.

ALP activity was assessed qualitatively with BCIP/NBT staining (Beyotime, China) and quantitatively with the ALP activity assay kit (Beyotime, China). Collagen secretion was determined with Sirius Red staining (Leigen, China), followed by desorption with 0.2 M NaOH/methanol (1:1) and absorbance measurement at 570 nm. Matrix mineralization was evaluated by Alizarin Red S staining (Beyotime, China), with bound dye dissolved in 10% cetylpyridinium chloride (Sigma‐Aldrich, USA) and quantified at 570 nm. All assays were performed in triplicate.

2.5.3. Inflammatory Response Analysis

RAW264.7 cells were co‐cultured with samples for 3 days. Cell viability was examined by Calcein‐AM/PI staining, and morphology was observed by SEM after graded ethanol dehydration. Culture supernatants were collected for ELISA (Boster, China) analysis of TNF‐α, IL‐6, IL‐10, and IL‐1β. RT‐qPCR was performed to assess inflammatory gene expression (TNF‐α, iNOS, IL‐10, and Arg1).

2.5.4. Boron‐Containing Medium and Conditioned Medium Assays

For the boron‐containing medium, sodium tetraborate was used as the boron source to prepare α‐MEM with varying boron concentrations (0 µg/L control, 100 µg/L B, 500 µg/L B, 1 mg/L B, 5 mg/L B, and 10 mg/L B). Reagent masses were calculated from the B mass fraction of the salt (Na₂B₄O₇·10H₂O: 11.34% B). MC3T3‐E1 cells were cultured in the medium for assays of viability (1, 4, and 7 days), ALP activity, collagen secretion, and mineralization (7 and 14 days), as described in Section 2.5.2. To prepare conditioned medium (CM), RAW264.7 cells were co‐cultured with the samples for 72 h. The supernatants were collected, clarified by centrifugation, and stored for later use. CM was mixed with fresh medium at a 1:2 ratio. The untreated control medium (UCM) was prepared using the same centrifugation protocol. MC3T3‐E1 proliferation and differentiation were then assessed as in Section 2.5.2. All assays were performed in triplicate.

2.6. In Vivo Experiments

Animal studies were conducted in compliance with the Regulations for the Administration of Laboratory Animals of the Ministry of Health of the People's Republic of China and the Guide for the Care and Use of Laboratory Animals. The study was reported in accordance with the ARRIVE guidelines. Male Sprague–Dawley (SD) rats (280–310 g) were used. All procedures were approved by the Institutional Animal Care and Use Committee (IACUC approval no. IACUC‐20240801‐01).

2.6.1. Surgical Implantation

Thirty‐two male SD rats were randomly divided into four groups (n = 8 per group). The rats were anesthetized with isoflurane and prepared under sterile conditions. A 1.5 cm longitudinal incision was made near the knee to expose the distal femoral condyle. A bone tunnel (2.2 mm diameter, 5 mm depth) was drilled under saline irrigation, and a Φ2.2×5 mm cylindrical implant was press‐fitted into the hole. After confirming stability, the site was irrigated, hemostasis was achieved, and the wound was closed with absorbable sutures. Animals were cared for postoperatively according to ethical guidelines, and specimens were collected at scheduled time points for analysis.

2.6.2. Micro‐CT Analysis

After 6 weeks, specimens were scanned with a Skyscan 1176 system (Bruker, Belgium; resolution: 18 µm). Data were reconstructed with NRecon software and analyzed with CTAn software to quantify bone mineral density (BMD), bone volume fraction (BV/TV), trabecular thickness (Tb.Th), trabecular number (Tb.N), and trabecular pattern factor (Tb.Pf).

2.6.3. Histological Analysis

Specimens were harvested at 1 and 6 weeks. The specimens were decalcified in 10% EDTA for one month and embedded in paraffin. Sections were stained with hematoxylin and eosin (H&E) to evaluate inflammation, Masson's trichrome staining to assess collagen deposition, immunohistochemistry (IHC) for osteopontin (OPN) and tartrate‐resistant acid phosphatase (TRAP; visualized with DAB substrate and blocked with 10% goat serum), and immunofluorescence (IF) for iNOS/Arg1 (M1/M2 polarization markers) and CD31 (angiogenesis marker). For IHC staining, sections were blocked with 10% goat serum and visualized with DAB substrate.

2.7. Statistical Analysis

All experiments were performed with at least three independent replicates, and data are presented as the mean±standard deviation (SD). Statistical analyses were conducted using independent samples t‐tests and one‐way analysis of variance (ANOVA) with Tukey's post hoc test. Differences were considered statistically significant at *p < 0.05, **p < 0.01, and ***p < 0.001.

3. Results and Discussion

3.1. Specimen Characterization

Inspired by the trabecular bone–like porous microstructure, boron‐doped hierarchical porous ceramic coatings were fabricated on TZNF alloys by a one‐step MAO process in a sodium tetraborate electrolyte. The morphology and chemical composition were tuned by applying voltages of 300–500 V. FE‐SEM showed a honeycomb‐like network of micron‐scale pores with pore walls densely decorated nanopores. Both length scales increased with voltage. Nanopore diameters were 75.82±10.51 nm, 125.26±15.24 nm, and 139.85±20.33 nm for M300V, M400V, and M500V, respectively, with corresponding micropore diameters of 1.03±0.15 µm, 2.80±0.25 µm, and 4.07±0.35 µm (Figure S1). Coating thickness increased monotonically with voltage (Figure 1A,B). AFM over a 30 µm×30 µm area showed parallel increases in roughness and actual surface area (Figure 1C,D). The arithmetic roughness (Sa) was 0.15±0.02 µm for TZNF, 0.70±0.19 µm for M300V, 1.24±0.11 µm for M400V, and 2.18±0.08 µm for M500V. The resulting hierarchical porosity markedly improved wettability and, together with the enlarged effective surface area, enhanced protein adsorption, with M500V exhibiting the highest adsorption capacity (Figure 1E,F). EDS mapping showed a uniform distribution of Ti, Zr, Nb, Fe, and O within the coating, with trace B detected (Figure S2). Our previous work showed that MAO of TZNF in sodium tetraborate produced coatings mainly composed of rutile and anatase, with the rutile fraction increasing with applied voltage. XPS survey and B 1s spectra identified Ti, Zr, Nb, and O as the major elements and indicated boron as B₂O₃ and B₄O₇ ^{2 −} [34].

FIGURE 1.

(A) Surface morphologies and cross sectional views of coatings. (B) Coating thickness. (C) Three‐dimensional surface topographies obtained by AFM. (D) Surface roughness (Sa) and surface area. (E) Water contact angles. (F) Protein adsorption capacity.

Boron release tests showed cumulative concentrations at 28 days of M300V (0.42±0.07 mg/L), M400V (0.62±0.02 mg/L), and M500V (1.15±0.11 mg/L). The release rate was faster during the first 7 days, and sustained release from M500V continued until day 14 (Figure 2A). The MAO coatings markedly enhanced corrosion resistance (Figure 2B–F). In electrochemical tests conducted in artificial saliva at 37°C, the corrosion potential shifted positively, the corrosion current density decreased by more than one order of magnitude, and the polarization resistance ranked M300V>M400V>M500V>TZNF (Figure 2B and Table S2). Nyquist plots exhibited the same trend, with M300V showing the largest semicircle (Figure 2C). Bode |Z| plots confirmed higher impedance for all MAO coatings than for TZNF (Figure 2D). The phase‐angle plots showed a single peak (∼80°) for TZNF but two peaks for the MAO coatings, indicative of a bilayer structure (Figure 2E). Equivalent‐circuit fitting (Figure 2F and Table S3) revealed that R1 increased by approximately an order of magnitude, particularly for M300V and M400V, with R2≫R1.

FIGURE 2.

(A) Boron release profiles in complete medium and (B–F) electrochemical tests in artificial saliva at 37°C. (B) Potentiodynamic polarization (PDP) curves, (C) Nyquist plots, (D) Bode |Z| plots, (E) Bode phase‐angle plots, and (F) equivalent circuit models.

SBF immersion assays demonstrated that boron‐doped coatings promoted hydroxyapatite (HA) formation (Figure S3 and Table S4). After 14 days of SBF immersion, phosphorus deposition was below the detection limit. Calcium adsorption increased in the sequence TZNF (0.03±0.01 at.%) < M500V (0.08±0.01 at.%) < M400V (0.20±0.06 at.%) < M300V (0.28±0.05 at.%). By 28 days, phosphorus accumulation increased markedly, consistent with the typical SBF mineralization pathway in which Ca first enriches the surface, followed by P participation and progressive maturation toward bone‐like HA [35]. The Ca/P ratio for M300V (1.45) was closest to the theoretical value for stoichiometric HA (1.67). Mechanistically, surface B–OH/tetraborate sites may lower the nucleation barrier and concentrate Ca^{2 +}/PO₄ ^{3 −} ions [36]. The smaller pores of M300V preserve more available nucleation sites and favor early nucleation. M500V releases more boron, yet its coarser pores likely reduce effective surface sites and limit subsequent phosphate deposition.

Stable adhesion at the coating–substrate interface is a key determinant of implant coating success [37]. Based on scratch‐track morphology and friction signals (Figure S4), M300V showed a straight scratch path with minimal debris, and only slight cracking at higher loads. In contrast, M400V displayed transverse spallation and delamination at lower loads, and M500V showed surface upheaval and wedge spallation from the outset (Figure S4D). The critical loads were measured as 24.61±1.15 N (M300V), 20.47±1.83 N (M400V), and 14.83±1.65 N (M500V) (Figure S4E)Vickers hardness tests indicated all MAO‐treated TZNF samples exceeded bare TZNF (332.5±7.5 HV), with M400V reaching the maximum (514.6±32.8 HV) (Figure S4F).

Collectively, MAO coatings enhanced both the mechanical stability and corrosion resistance of TZNF. The voltage‐dependent decreases in adhesion and hardness are attributable to increased porosity, cracking, and residual thermal stresses [37, 38]. The corrosion resistance is attributed to a duplex structure. A porous outer layer overlies a compact inner barrier (cross section in Figure 1A). The inner layer affords the primary protection (R2≫R1). However, higher voltages intensify microdischarges, generating microcracks that facilitate electrolyte ingress, thereby increasing i _corr and lowering R _p. Large pores also enlarge the effective reactive area and promote localized corrosion [39, 40]. A depressed CPE exponent (n < 1) further indicates aggravated surface heterogeneity with voltage due to enlarged pores and defects [41]. Overall, while voltage optimization is required to balance mechanical integrity and corrosion resistance, the protective effect of MAO on TZNF is substantial.

Tetraborate electrolytes tend to generate micropores via gas venting and melt ejection from discharges at the substrate surface [12, 13]. In addition, tetraborates readily form B–O–Ti bridges within the Ti–O network and introduce a B₂O₃ glassy phase [42]. This yields a more fluid melt with lower surface tension, which promotes subsequent spheroidization and refinement [43]. Fine‐scale phase separation may occur between TiO₂‐rich and B₂O₃‐rich regions, while subsequent thermal cycling/leaching and differential shrinkage generate nanoscale porosity/microcracks in thin pore walls, yielding a “nanopore‐decorated micropore” structure [44, 45]. Some studies also suggest that the formation of nanopores is associated with low‐energy discharge and the dissolution of titanium oxide by molten boron oxide [15, 46]. The applied voltage modulates the multiscale porosity, roughness, wettability, and boron incorporation of MAO coatings. Higher voltage (500 V) intensifies discharges, producing larger pores and thicker coatings (Figure 1A) but can also induce thermal stress and microcracking [47]. Enhanced hydrophilicity arising from the porous structure and capillarity aids protein enrichment, with larger pores providing greater volume and surface area that favor wettability and protein adhesion [48]. The boron further regulates bioactivity through its release. Notably, M300V, characterized by smaller pore sizes and less boron release, demonstrates the strongest HA‐forming ability.

3.2. Macrophage Polarization on Various Surfaces

Rapid post‐surgical macrophage accumulation on implant surfaces and subsequent cytokine/chemokine release allow macrophages to strongly influence implant outcomes. Their plasticity gives rise to distinct phenotypes, most notably M1 (pro‐inflammatory) and M2 (anti‐inflammatory/repair) [49]. In this study, RAW264.7 cells were cultured on the different surfaces to assess the effects of coating properties on macrophage polarization (Figure 3A). Live/dead staining showed good macrophage viability in all groups, with few dead cells observed. Compared with the bare TZNF surface, macrophages on the MAO‐treated TZNF surfaces exhibited better spreading and developed more elongated filopodia. These filopodia were mainly distributed at the cell edges and bridged the honeycomb‐like micropores. On all MAO surfaces, the cells did not infiltrate the pores, instead extending filopodia along the pore walls (Figure 3B). CCK‐8 results showed that cell viability increased with culture time, with no significant difference among the groups (p > 0.05, Figure 3D). Immunofluorescence staining showed that Arg1 expression was highest in the M300V group, followed by the M400V group, while iNOS expression was lower in all MAO groups than in the TZNF group (Figure 3C). Notably, macrophages in the M300V group exhibited a more elongated spindle‐like morphology, which is commonly associated with M2 polarization.

FIGURE 3.

Regulation of macrophage polarization on different samples. (A) Schematic illustration of the RAW264.7 cell culture workflow. (B) Live/dead staining and SEM images showing the morphology of RAW264.7 cells on different samples. (C) Immunofluorescence staining of macrophage polarization markers, including Arg1 for the M2 phenotype and iNOS for the M1 phenotype. (D) Viability of RAW264.7 cells on Days 1 and 3. (E) Secretion of TNF‐α, IL‐1β, IL‐10, and IL‐6 on Day 3. (F) Relative mRNA expression of inflammation‐related genes on Day 3.

ELISA measurements showed that TZNF induced the highest levels of TNF‐α and IL‐1β, whereas M300V exhibited the lowest pro‐inflammatory cytokine levels and the highest IL‐10 secretion (Figure 3E). Consistently, RT‐qPCR analysis indicated elevated expression of M1‐associated genes (TNF‐α and iNOS) in TZNF, while M2‐associated genes (IL‐10 and Arg1) were most strongly upregulated in M300V (Figure 3F). By contrast, macrophages on M500V displayed an overall less activated phenotype, with generally low expression of both pro‐ and anti‐inflammatory markers.

These findings indicate that MAO coatings promoted macrophage polarization toward the M2 phenotype. Previous studies have demonstrated that topography‐induced changes in cell morphology can mediate macrophage polarization [43, 50, 51]. Tensile cues imposed by defined micropatterns can drive macrophages toward M1 or M2 polarization. Moreover, 30‐nm TiO₂ nanotubes induce M2 activation more effectively than 80‐nm counterparts [52]. This observation is consistent with our findings that smaller pore sizes are more conducive to M2 polarization in macrophages. Smaller pores (90 nm–1 µm) inhibited M1 polarization and promoted M2 polarization, likely by constraining cytoskeletal remodeling and reducing pro‐inflammatory mechanotransduction, whereas larger pores increased surface contact and accelerated M1 activation [43]. In addition, the enhanced macrophage response on MAO surfaces may result from increased ECM protein adsorption facilitated by roughness and hydrophilicity. Hydrophilic surfaces (contact angle 40°–70°) preferentially adsorb non‐immunogenic proteins, reducing pro‐inflammatory signaling and supporting M2 polarization [49, 53]. Another study showed that, compared with the pristine smooth surface (Ra = 0.20 µm) and a high‐roughness surface (Ra = 2.60 µm), titanium surfaces with low‐to‐moderate roughness (Ra = 0.51–1.36 µm) more readily induce macrophage polarization toward the M2 phenotype [54].

Beyond topographical cues, tetraborate ions attenuate inflammation [55]. Li et al. fabricated boron‐doped calcium silicate coatings and found in vitro that a release level of 1 mg/L was more effective than 40 mg/L in promoting osteogenesis, angiogenesis, and anti‐inflammatory effects [27]. Ikedo et al. reported dose‐dependent responses in RAW264.7 cells. Proliferation increased at 1 to 5 mg/L, osteogenic genes were upregulated at 3 and 10 mg/L, and proliferation declined at 30 mg/L or higher [55]. By contrast, Demirci et al. reported that only high concentrations of boric acid (100–200 mg/L) or sodium pentaborate pentahydrate exhibited anti‐inflammatory activity in LPS‐challenged macrophages [56]. Overall, reports vary regarding the optimal concentration range across studies. For simple inorganic borate salts, a commonly cited beneficial window is roughly from the µg/L to mg/L range, while exposures above 500 mg/L may be cytotoxic. For boron‐containing nano or composite systems, the effective range should be set with regard to carrier properties and exposure duration [26, 27, 55, 56, 57]. The effects of various concentrations of boric acid and boron compounds in cells and animals are summarized in Table S5. The boron concentrations released by M300V, M400V, and M500V are approximately 1.95, 2.89, and 5.36 mg/L when converted to sodium tetraborate concentrations, and approximately 2.40, 3.55, and 6.59 mg/L when converted to boric acid equivalents. All are within the reported beneficial range. Mechanistically, borates can modulate the MAPK pathway to alter macrophage activity and secretory profiles, reducing ERK1/2 and p38 phosphorylation, and downregulating NF‐κB–controlled mediators including IL‐1β, MIP‐1α, and iNOS [27, 55]. Together with the responses of RAW264.7 cells on M300V–M500V surfaces, these results suggest that the modulation of macrophage polarization may be associated with the combined effects of coating morphology and boron release. Among the tested coatings, M300V demonstrated the most favorable immunomodulatory effects due to its balanced properties, followed by M400V.

3.3. The Immunomodulatory Effects of Macrophages on Osteoblasts

Activated macrophages can suppress excessive inflammation and release osteogenic and angiogenic mediators that promote osseointegration [20, 58]. Because macrophage‐derived secretomes influence osteoprogenitor fate, conditioned media (CM) from each group were collected to assess their paracrine effects on preosteoblasts. Unconditioned medium (UCM) served as the control. In the CCK‐8 assay, CM‐M300V significantly enhanced preosteoblast proliferation, CM‐M400V showed only a modest increase at Day 7, whereas CM‐TZNF and CM‐M500V were comparable to UCM (Figure 4A). Differentiation endpoints (ALP activity, collagen secretion, and ARS staining) were also enhanced with CM‐M300V and CM‐M400V, whereas CM‐TZNF and CM‐M500V showed no such enhancement (Figure 4B–G). RT‐qPCR showed higher expression of RUNX2, COL1, and OCN in CM‐M300V and CM‐M400V, with differences more pronounced at 14 days than at 7 days (Figure 4H–J).

FIGURE 4.

Effects of the macrophage‐conditioned osteoimmune microenvironment on osteoblasts. (A) Cell proliferation in the indicated media. (B) Quantitative analysis of ALP activity at Days 7 and 14. (C) Quantitative analysis of collagen secretion. (D) Quantitative analysis of ARS‐stained mineralization. (E) Representative images of ALP staining. (F) Representative images of collagen staining. (G) Representative images of ARS staining. (H–J) Relative mRNA expression of osteogenic genes.

Collectively, macrophage‐derived CM modulated osteoblast behavior in a surface‐dependent manner, consistent with the M2‐skewed polarization observed for M300V/M400V in Section 3.4. These findings suggest that soluble M2‐associated mediators, such as IL‐10, may activate osteogenic signaling pathways in osteoblasts and thereby enhancing their proliferation and differentiation [59, 60]. Modulation of macrophages by surface topography can further influence osteoblast‐lineage cells via paracrine signaling. Zhu et al. cultured macrophages on honeycomb‐like TiO₂ microstructures of four characteristic scales and found that the resulting macrophage‐conditioned media markedly promoted osteogenic differentiation of MSCs [43]. In addition, macrophage secretomes induced in response to boric acid significantly enhanced osteogenic differentiation and mineralization of KUSA‐A1 (a bone marrow–derived mesenchymal stem cell line) in a dose‐dependent manner over the range of 3–30 mg/L [55]. Overall, these in vitro findings indicate that MAO coatings can regulate the osteoimmune axis by tailoring surface morphology and boron release, thereby indirectly enhancing osteogenesis.

3.4. Osteoblast Differentiation on Various Surfaces

Osteogenic differentiation is a key biological process underlying osseointegration [61]. To assess osteogenic potential, MC3T3‐E1 cells were seeded on MAO‐treated TZNF specimens. All groups exhibited negligible cytotoxicity and increased cell density by Day 7, with the most evident increase observed on M300V (Figure S5A). CCK‐8 assays confirmed significantly higher metabolic activity on MAO‐treated surfaces than on bare TZNF at Day 7, with M300V exhibiting the greatest increase (Figure 5B). F‐actin staining revealed that after 1 day of culture, cells on M300V surfaces displayed a well‐spread morphology with more filopodia and lamellipodia, whereas cells on M400V and M500V were less spread and showed fewer protrusions (Figure 5A). SEM further confirmed that cells adhered preferentially adhered to pore edges, extended abundant filopodia, and they rarely penetrated into pores (Figure 5C). RT‐qPCR revealed broad upregulation of RUNX2, OSX, ALP, COL1, OCN, and OPN on MAO surfaces, with M300V showing the most pronounced changes. RUNX2 increased at early time points, consistent with the initiation of osteogenic differentiation; OSX expression was associated with extracellular matrix (ECM) formation, and COL1/ALP levels increased, indicating active matrix synthesis and early mineralization. By Day 14, OCN and OPN levels were markedly elevated while ALP declined, consistent with matrix maturation and mineralization. These results are shown as a heatmap (Figure 5D) and bar charts (Figure S5B). Across ALP, ARS, and collagen assays, MAO surfaces outperformed TZNF, with M300V showing the highest responses overall (Figure 5E–J).

FIGURE 5.

Adhesion and osteogenic differentiation of osteoblasts. (A) F‐actin/DAPI staining at Days 1 and 2. (B) Cell viability at Days 1, 4, and 7. (C) Cell adhesion morphology at Day 1. (D) Heatmap of osteogenic gene expression. (E) Representative images of ALP staining at Days 7 and 14 and (F) corresponding quantitative analysis. (G) Representative images of ARS staining and (H) corresponding quantitative analysis. (I) Representative images of type I collagen staining and (J) corresponding quantitative analysis.

MC3T3‐E1 cells were cultured in media supplemented with sodium tetraborate, with concentrations expressed as elemental B (0 µg/L, 100 µg/L, 500 µg/L, 1 mg/L, 5 mg/L, and 10 mg/L; Figure S6). Proliferation was unchanged at 100 µg/L–5 mg/L but was suppressed at 10 mg/L (Figure S6A). ALP activity increased at 100 µg/L–1 mg/L (Figure S6B,E), and these concentrations produced the greatest Day‐14 mineralization (Figure S6D,G), whereas collagen deposition was largely unaffected (Figure S6C,F). Thus, B concentrations comparable to those released from MAO coatings favored osteogenic differentiation.

Microscale features provide mechanical interlocking and appropriate roughness, whereas nanoscale cues more effectively boost osteoblast proliferation and differentiation [50]. Osteoblast adhesion is initiated by filopodia that anchor to adsorbed extracellular matrix (ECM) proteins via integrins, which in turn activate downstream signaling molecules such as focal adhesion kinase (FAK) and Src‐family kinases to support cell survival, migration, and lineage commitment [50, 52]. Increased roughness and well‐defined microtopography increase the available cell–substrate contact area and promote osteogenesis. Moderately rough surfaces with Ra around 1 to 2 µm are generally favorable [62]. Nonetheless, when overall roughness is comparable, surface patterning and feature alignment can still modulate cellular responses [63].

In addition, an appropriate boron concentration range further potentiates osteogenesis by enhancing ECM deposition and Ca–P nucleation, amplifying BMP/Wnt signaling, and upregulating RUNX2 and ALP [64]. Boron also modulates the RANKL/OPG axis and mineralization‐related proteins [36]. Tsuchiyama et al. showed that low concentrations of boron released from borate bioglass (≤121 µM, ∼1.31 mg/L) enhanced osteogenic functions of MC3T3‐E1 cells without apparent cytotoxicity [26]. Consistently, 0.1–1 mg/L boron promotes MC3T3‐E1 calcification, whereas ≥1 mM boron suppresses proliferation, with toxicity mitigated by co‐released ions [30, 64, 65, 66]. For boric acid specifically, Hakki et al. observed that 1–10 µg/L markedly increased MC3T3‐E1 mineralization [24], while Ikedo et al. found that 1–30 mg/L inhibited mineralization in KUSA‐A1 cells, suggesting concentration and cell‐line‐dependent effects [55]. Moreover, 15–200 mg/L boric acid inhibits HUVEC angiogenesis but stimulates fibroblast/keratinocyte proliferation, migration, and growth‐factor release, suggesting a potential pro‐healing role [56]. More information regarding the effects of boric acid and boron compounds on osteoblasts is summarized in Table S5.

3.5. In Vivo Early Inflammatory Response

To evaluate the in vivo early inflammatory response and osteogenic capacity of boron‐doped MAO coatings, different specimens were implanted into rat femoral condyles, and peri‐implant inflammation and new bone formation were examined at 1 and 6 weeks (Figure 6A). As shown in Figure S7, the surface morphology of implants prepared using the same MAO parameters in vivo was consistent with that of flat specimens in vitro.

FIGURE 6.

Inflammatory response around the implants at 1 week of implantation. (A) Schematic illustration of the in vivo experiment. (B) H&E staining, (C) Masson staining, and (D) TRAP staining of the inflammatory response surrounding the implants at 1 week. (E) Immunofluorescence staining of iNOS and Arg1 as markers of M1 and M2 macrophage polarization, respectively. (F) Immunofluorescence staining of CD31 as an angiogenesis marker.

At 1 week, H&E staining (Figure 6B) revealed varying degrees of neutrophil and macrophage infiltration surrounding the implants in all groups, which did not fully correspond to the in vitro observations. Masson's trichrome staining showed that the collagen fibers were loose and immature, appearing blue or cyan, indicating that tissue repair had commenced but that a mature fibrous architecture had not yet developed (Figure 6C). A pronounced fibrotic response was observed at the bare TZNF interface. TRAP staining demonstrated that osteoclast activity was highest in the TZNF group, whereas it was markedly reduced across all MAO‐treated groups (Figure 6D). Arg1/iNOS immunofluorescence staining revealed stronger Arg1 and iNOS signals in the M300V and M400V groups, suggesting enhanced macrophage activation with a repair‐associated polarization tendency in the peri‐implant microenvironment of these implants (Figure 6E). In addition, CD31 immunofluorescence surrounding the MAO implants was elevated, particularly in the M300V and M400V groups, indicating enhanced early endothelial activity (Figure 6F). Collectively, these findings suggest that MAO coatings attenuate early osteoclast activity, promote peri‐implant immunomodulation, and enhance the early angiogenic response, thereby establishing a microenvironment favorable for subsequent bone regeneration.

The in vivo and in vitro inflammatory phenotypes were not fully concordant, potentially because in vivo responses are shaped by additional determinants that are absent in simplified in vitro systems. These factors include the recruitment of multiple immune cell populations (e.g., neutrophils and lymphocytes), serum protein adsorption, dynamic fluid exchange, and tissue injury–associated signaling. Importantly, a moderate acute inflammatory response during the early in vivo phase represents a necessary stage of tissue repair. It facilitates the clearance of necrotic tissue and debris, the recruitment of reparative cells, and the initiation of angiogenic and regenerative signaling cascades.

3.6. In Vivo Osseointegration

All implants were placed in the femoral condyles. At 6 weeks, bare TZNF had displaced toward the marrow cavity, whereas MAO‐coated implants remained well fixed. Micro‐CT revealed increased peri‐implant bone formation in all MAO groups, most pronounced for M300V (Figure 7A). BV/TV reached 48.6%±2.6% in M300V, compared with 25.8%±3.8% in TZNF, 32.7%±2.4% in M400V, and 35.1%±1.3% in M500V (Figure 7B). In the M300V group, trabecular number and thickness were increased, whereas separation was reduced. Histological analysis was consistent with these findings, showing limited new bone and collagen on TZNF and substantial deposition on MAO surfaces, particularly M300V, without extensive fibrous encapsulation (Figure S8A,B). TRAP staining revealed numerous osteoclasts around TZNF, few around M300V, and intermediate levels around M400V/M500V (Figure 7C). OPN immunohistochemistry was weak in TZNF, strong in M300V/M400V, and intermediate in M500V (Figure 7D). Arg1/iNOS immunofluorescence indicated the strongest M2‐skewed profile in M300V (high Arg1, low iNOS), an M1‐biased profile in TZNF, and intermediate patterns in M400V/M500V (Figure 7E). CD31 staining showed sparse vascularization in TZNF, dense CD31‐positive microvessels in M300V, and moderate vascularization in M400V/M500V (Figure 7F). Taken together, these data indicate that MAO coatings markedly improve osseointegration of TZNF implants in vivo, with M300V providing the most favorable combination of enhanced osteogenesis and angiogenesis together with reduced osteoclast activity.

FIGURE 7.

Osseointegration around the implants after 6 weeks of implantation. (A) Three‐dimensional reconstruction images obtained by micro‐CT. (B) BV/TV, Tb.N, Tb.Th, and Tb.Sp in the newly formed bone around the implants. (C) TRAP staining, (D) OPN staining, (E) Arg1/iNOS immunofluorescence staining, and (F) CD31 immunofluorescence staining of peri‐implant new bone formation at 6 weeks.

After implantation, a protein corona forms rapidly on the implant surface and stabilizes the blood clot, followed by a short and necessary acute inflammatory response. This response normally resolves within a few days. Macrophages then shift toward reparative phenotypes, including M2‐like polarization, which supports angiogenesis and tissue repair. In parallel, MSCs and pre‐osteoblasts adhere and differentiate, deposit and mineralize matrix, and form woven bone. Over the next weeks to months, this tissue remodels into lamellar bone, achieving direct bone–implant contact and long‐term stability. However, excessive or prolonged acute inflammation may induce fibrous encapsulation and impede osseointegration [21].

Based on in vitro and in vivo results, the boron‐doped MAO porous coating modulates the early interfacial microenvironment through coupled “topography–chemistry” effects. The interconnected micro/nanopores and increased real surface area improve wettability and protein adsorption, which facilitates adhesion and spreading of immune cells and osteogenic cells. Concurrently, tetraborate released during the early postoperative phase (about the first 7 days) maintains a low‐level, tolerable alkaline milieu and participates in signaling [67]. Boron‐doped porous coatings may modulate cell behavior through complementary topographical and chemical cues (Figure 8).

FIGURE 8.

Proposed mechanism by which a boron‐doped porous coating promotes osseointegration, possibly through enhanced M2 polarization, attenuated inflammation, and upregulated osteogenic gene expression.

The hierarchical micro‐ and nanoporous topology may engage cells through contact‐mediated adhesion. In macrophages, such topographical cues may promote clustering of integrins such as α5β1 and αvβ3 and activate the FAK/Src‐related adhesion signaling, potentially strengthening M2‐polarizing transcription. In osteoblasts [21, 43], the micro‐/nanostructured surface may support α2β1 binding to adsorbed extracellular matrix proteins and activate adhesion‐associated FAK/MAPK (ERK1/2), thereby potentially augmenting RUNX2‐related osteogenic differentiation [68]. In parallel, tetraborate species may act as chemical regulators. In macrophages, they may attenuate NF‐κB signaling, thereby reducing pro‐inflammatory cytokine expression and favoring M2‐like polarization [27, 55]. In osteoblasts, borate have been proposed to modulate BMP‐ and Wnt‐related osteogenic signaling, which may be associated with the upregulation of RUNX2 and osteogenic markers, including ALP, COL1, OCN, and OPN [26, 29]. In addition, factors secreted by M2‐like macrophages, particularly IL‐10, may bind IL‐10 receptors on osteoblasts and engage JAK/STAT3 signaling. This macrophage‐osteoblast crosstalk may cooperate with RUNX2‐associated osteogenic regulation, thereby supporting osteogenic gene expression, cell survival, and differentiation [59, 60]. Consistent with this proposed immuno‐osteogenic model, the in vivo results showed reduced inflammatory responses, lower osteoclast activity, increased CD31‐positive staining, higher BV/TV, and a denser bone‐implant interface at 6 weeks. However, these findings should be interpreted as indirect and correlative evidence, because the physical and chemical cues of the coatings were not fully decoupled, and the proposed signaling pathways were not directly validated in the present study.

Among the coatings produced at different voltages, M300V appears to provide the most balanced combination of pore size, wettability, and tetraborate release, which may explain its overall superior biological performance. Nevertheless, further optimizing pore microstructure and release kinetics to match the window from acute inflammation resolution to early angiogenesis and osteogenesis. Future studies should include direct pathway readouts, such as phosphorylation and localization assays for p‐FAK/p‐Src, p‐Smad1/5/8, and β‐catenin, as well as pathway‐specific inhibitors and neutralizing antibodies, such as IL‐10 blockade, to clarify the causal contribution of topographical cues, tetraborate species, and macrophage–osteoblast crosstalk to the proposed regulatory network.

4. Conclusion

In this study, boron‐doped hierarchical porous ceramic coatings were fabricated on low elastic modulus TZNF alloys using a one‐step MAO process. By adjusting the applied voltage, both boron content and morphology of the coatings were effectively regulated. The average pore size increased from 1.03 to 4.07 µm, Sa increased from 0.70 to 2.18 µm, and boron release increased from 0.42 to 1.15 mg/L. Hydrophilicity and protein adsorption were promoted by increased surface roughness and porous microstructure. Composed mainly of rutile and anatase phases, the coatings exhibited strong adhesion to the substrate. In addition, the compact inner layer contributed to improved corrosion resistance, with M300V showing 87% lower I _corr than untreated TZNF.

The coatings showed excellent biocompatibility, and favorable immunomodulatory and osteogenic responses. Among the tested groups, M300V showed the most favorable combination of boron release kinetics and surface micromorphology. This coating reduced pro‐inflammatory cytokine secretion and enhanced extracellular matrix mineralization, further suggesting a beneficial immunomodulatory effect on osteogenesis. In vivo, the M300V coating reduced early inflammatory responses, suppressed osteoclast activity, and promoted early peri‐implant bone formation. Overall, boron‐doped MAO surfaces promote osteogenesis by modulating the bone microenvironment and offer a viable strategy for implant modification.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: adhm71440‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.