Introduction

Osteosarcoma is a malignant bone tumor with high morbidity rates in adolescents and children, with a 5-year survival rate of less than 54%.1 Current clinical treatment strategies for this malignant tumor primarily include chemotherapy, radiotherapy, and surgically removing the tumors.2 In most cases, surgical intervention only removes detectable cancer, with the incomplete elimination of tumor cells potentially leading to an increased risk of recurrence. Additionally, bone defects caused by extensive surgical excision may exceed the bone’s self-healing capacity, making bone damage repair a major challenge. Adjunctive chemotherapy or radiotherapy can reduce the risk of cancer recurring in osteosarcoma patients to some extent; however, these therapies disrupt the microenvironment of bone regeneration and hinder the natural healing and regeneration of bone tissue.3 To prevent this issue, various hydrogel materials have been designed to prevent osteosarcoma recurrence and to promote bone regeneration.4,5 Nevertheless, many dual-function strategies rely on chemotherapeutic agents or energy-based therapies such as photothermal or photodynamic therapy.6 The cytotoxicity of chemotherapeutic agents, as well as the heat, light, or reactive oxygen species generated during energy-based treatments, can adversely affect surrounding bone tissue and ultimately impede osteogenesis. Therefore, materials designed for single-step, simultaneous application under identical post-surgical conditions are highly desirable.

Injectable hydrogels based on natural polymers are among the most popular and representative biomaterials for bone repair.7,8 Among the various natural polymer-based hydrogels, chitosan (CS)-based systems have attracted much attention in the field of bone tissue engineering due to their good biocompatibility, bacteriostatic properties, and biodegradability.9–11 However, conventional CS hydrogels are brittle, mechanically weak, and lack elasticity, limiting their application.12 Hydrogels constructed with dual or multiple network architecture have been shown to possess markedly improved mechanical properties and dimensional stability compared with their single-network counterparts.13 In our earlier work, dual-network hydrogels composed of thiolated chitosan (CS-NAC) and silk fibroin (SF) have demonstrated enhanced strength and elasticity.14 Nevertheless, the absence of intrinsic osteoinductive activity in these systems restricts their therapeutic efficacy.

Previous studies have found that incorporating bioactive glass (BG) into hydrogels can promote bone regeneration by releasing silicon and calcium ions, which in turn enhance the expression of genes associated with osteogenic differentiation.15 Furthermore, incorporating metallic ions into bioactive glass allows for the development of biomaterials with enhanced biological properties, that can be tailored for specific clinical applications. Gallium ions, which inhibit cancer cell growth and metastasis through various mechanisms, including interference with iron metabolism, suppression of DNA synthesis, and induction of oxidative stress, have been approved by the FDA for preventing bone resorption, inhibiting biofilm formation, and treating bone, colon, and prostate cancers.16,17 Gallium-doped bioactive glass (GaBG) can promote the early differentiation of osteoblasts and inhibit osteoclast formation,18 while selectively inducing cytotoxicity in osteosarcoma (Saos-2) cells with minimal toxicity to normal cells at appropriate gallium ion concentrations.19 Previous studies primarily incorporated gallium ions into polymer matrices or bulk bioactive glass scaffolds, resulting in uncontrolled ion release and restricted multifunctionality.20–22 Gallium-doped bioactive glasses and injectable hydrogels have been individually explored for bone-related applications, the integration of GaBG into an injectable hydrogel system with systematic in vivo validation remains limited.

Here, we develop a gallium-doped bioactive glass hydrogel composite material that addresses these limitations by combining the osteogenic, antitumor, and bioactive properties of GaBG with the injectability and tissue-adaptive mechanical properties of CS-NAC/SF hydrogel. Gallium was structurally doped into a bioactive glass network at a controlled concentration of 5 to 15 mol%. The resulting CS-NAC/SF/GaBG hydrogel exhibits well-controlled degradation, enhanced mechanical strength, excellent in vivo biocompatibility, and notable apatite-forming ability. Compared to previously reported injectable hydrogels based on CS or SF, this system achieves a more balanced combination of injectability, mechanical stability, and biological function. We hypothesize that sustained release of Si and Ca ions supports osteogenic activity, while Ga ions effectively suppress osteosarcoma growth without additional chemotherapeutic agents. To the best of our knowledge, GaBG-based injectable hydrogels with independent in vivo validation of both osteogenic and antitumor performance have rarely been reported. The study reports its results regarding the preparation, characterization, and in vivo performance of CS-NAC/SF/GaBG gels.

Materials and Methods Materials

Chitosan (CS, 2.2×106 Da, degree of deacetylation: ca. 95%), N-acetyl-l-cysteine (NAC), N-hydroxysuccinimide (NHS), 1-ethy-3-(3-dimethylaminopropyl carbodiimide) hydrochloride (EDAC), HRP (horse radish peroxidase, 300 IU/mg) and tetraethyl orthosilicate (TEOS) were purchased from Aladdin Inc, China. Cetyl trimethyl ammonium bromide (CTAB), Ca(NO3)2·4H2O, Ga(NO)3·xH2O were obtained from Sinopharm, China. The Ga precursor was used to control the molar fraction of Ga3⁺ in the glass composition. All reagents used were of analytical grade and had not been purified before use.

Synthesis of Bioactive Glasses

Bioactive glasses containing different doped amounts of Ga ions (Ga3+) were synthesized according to a previously reported sol–gel method with minor modifications.23 0.204 g of CTAB was dissolved in 83 mL of deionized water and 39 mL of ethanol. Subsequently, 1.5 mL of NH3·H2O, 1.5 mL of TEOS, and 0.297 g of Ca(NO3)2·4H2O were sequentially added at 30 min intervals. After stirring for 12 h, the suspensions were purified with water and ethanol by centrifugation (8000 rpm, 5 min) and then calcined at 650°C for 3h. The designed (85-x) Si-15Ca-xGa bioactive glasses with different mol% (x=0, 5, 10, or 15) were produced using the same method, and designated as BG, 5GaBG, 10GaBG, and 15GaBG, respectively. Parameters for these bioactive glasses are provided in Table 1.

Table 1 Parameters for Bioactive Glasses Nanoparticles

Fabrication of BG Hydrogels or GaBG Hydrogels

BG hydrogels or GaBG hydrogels were fabricated as detailed in earlier reports.14 A series of CS-NAC solutions (1.5wt%) were prepared containing 0.25wt% (0.25BG/CNS), 0.5wt% (0.5BG/CNS), 1wt% (BG/CNS), 1.5wt% (1.5 BG/CNS) or 1wt% xGaBG (xGaBG/CNS, x=5, 10, 15). Their pH was adjusted to approximately 7 using a 5% NaHCO3 solution. A 10% SF solution was added to the CS-NAC/BG or CS-NAC/xGaBG mixtures to reach a final SF concentration of 3%. To each of them (1mL), 10 μL of H2O2 solution (500 mmol/L) and 10 μL of HRP solution (1000 U/mL) were introduced with stirring. The resulting precursor solutions were incubated at 37 °C to allow gelation. The vials were inverted every 20s to examine the flowability of the solutions, and gelation time was recorded when the solutions stopped flowing. Measurements were repeated three times for each formulation (n = 3). Bioactive glass–free hydrogels (CNS) were prepared using the same procedure and served as controls. The corresponding parameters of these hydrogels are summarized in Table 2 and Supplementary Table 1.

Table 2 Parameters for Hydrogel(a)

Characterization

The microstructures of the BGs and the GaBGs of three different types were visualized by TEM (HT7700). ζ-potential and PDI parameters characterized by a Zetasizer Nano-ZS90 (Malvern Panalytical). The surface morphology and porous structure of freeze-dried hydrogels were observed by scanning electron microscopy (SEM, Quanta 200 or Vega3, Tescan) after sputter-coating with gold. Pore size distributions were quantified by analyzing 200 randomly selected pores per sample using imageJ software. Elemental composition was performed using EDS (X-MaxN Oxford Instruments). In the swelling tests, 800 mg of dry gels (Wd) were immersed in PBS with shaking until equilibrium, then weighed (Ws) after removing surface water. Each test was performed in triplicate (n = 3) for each hydrogel formulation. The swelling index (SI) was calculated as follows:

(1)

Rheological Testing

Rheological properties were measured using a rheometer (Kinexus Pro KNX2100 UK) equipped with a parallel-plate sample holder. During the frequency sweep tests, elastic modulus (G′) and viscous modulus (G′′) were measured over 0.1–100 Hz at 1% strain and 37 °C. Strain sweep measurements were conducted at 37 °C and 1 Hz to determine the gel yield strain. Shear viscosity of the samples was evaluated across 0.1–300 s−1 at both 25 °C and 37 °C.

Cell Viability and Proliferation

MC3T3 rat osteoblasts and UMR-106 cells were obtained from Cell Resource Center (Shang hai). MC3T3 cells were maintained in α-MEM supplemented with 10% fetal bovine serum and 1% penicillin–streptomycin, while UMR-106 cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM) containing 15% fetal bovine serum and 1% penicillin-streptomycin. All cells were incubated at 37°C in a humidified 5% CO2 atmosphere, with medium replacement every other day.

CS-NAC solution (or a mixture of CS-NAC with bioactive glass nanoparticles) and SF solution were introduced into different glass vials to form thin-layer-gel and sterilized under UV light for 4 hours. Subsequently, the components were combined with HRP and H2O2 to obtain five types of hydrogel precursors, as summarized in Table 2. For cell encapsulation, 1.8 mL of each precursor solution was mixed with 200 μL of culture medium containing 2 × 106 cells (MC3T3 or UMR-106). The cell-loaded mixtures were allowed to undergo gelation at 37 °C before further evaluation.

The cell proliferation of MC3T3 and UMR-106 cells within the various hydrogels was assessed using the CCK-8 assay (Dojindo, Japan) at 24 h, 48 h, and 72 h following a previously described protocol.24 Cell viability was further evaluated with a Live/Dead staining kit (Aladdin Inc, China) at 3 and 7 days. All assays were conducted with at least three technical replicates per group. Detailed procedures are provided in the Supplementary Information.

Animals

All animal procedures in this study were approved by the Animal Care and Use Committee of Huazhong University of Science and Technology (Approval No. S2884) and conducted according to the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals. All rats in this study were housed individually with a standardized diet in a temperature-controlled environment and monitored daily for potential complications or abnormal behavior. To enable a clearer and more rigorous in vivo assessment of each biological function, the dual functions of the GaBG hydrogel were independently evaluated using a calvarial defect model for bone regeneration and a subcutaneous tumor model for osteosarcoma suppression.

In vivo Bone Repair Assays

Thirty female SD rats (250–300 g) were randomly assigned to five groups (n = 6 per group): (1) control group (the defects with nothing filled); (2) CNS; (3) BG/CNS; (4) 10GaBG/CNS; (5) 15GaBG/CN. The composition of the hydrogel used in each experiment is detailed in Table 2. The SD rats were anesthetized using 1.0% (w/v) pentobarbital sodium (40 mg/kg, intraperitoneal injection). A 1.5–2 cm cranial incision was made slightly off the sagittal midline, and the pericranium was carefully removed. Calvarial defects with 4 mm in diameter were created using a trephine burr.25 Defects in the control group were left unfilled, while those in the experimental groups were injected with the corresponding hydrogels. The incision was then closed with 3–0 absorbable silk sutures. After the operation, all rats received antimicrobial treatment with penicillin (total dose: 16,000 IU, intramuscular injection) for three consecutive days. At the predefined experimental endpoint, animals were euthanized using an overdose of sodium pentobarbital (150 mg/kg, intraperitoneal), followed by cervical dislocation to ensure death, in compliance with AVMA guidelines.

At 8 weeks post-surgery, the calvarial samples were harvested for the evaluation of bone regeneration. New bone formation was quantitatively analyzed by micro-computed tomography (μ-CT), followed by histological assessment using hematoxylin and eosin (H&E) and Masson’s trichrome staining. Detailed analytical procedures are provided in the Supplementary Information.

In vivo Anti-Osteosarcoma Assays

To study the therapeutic effects of Ga ions, an osteosarcoma model was set up by injecting UMR-106 cells into nude mice. The experiment was performed as reported as previously,26 but a little change. UMR-106 cells were suspended in the DMEM medium as the concentration was 1×107/mL and 100 μL of cells suspension solution were injected into the armpit of the right anterior limb of nude mice (female, 5 weeks old). When the volume of osteosarcoma arrived at about 300mm3, the nude mice were randomly divided into 5 groups (n=4). Then, the tumors were surgically removed, and 10% residual tumor was left to mimic residual tumors after surgery. After closing the skin incision, 100 μL hydrogels, including CNS, BG/CNS, 10GaBG/CNS, and 15GaBG/CNS (The corresponding composition of these hydrogels are summarized in Table 2), were injected around residual tumors by subcutaneous injection, respectively. The group without any treatment was used as the control group. The tumor dimensions and body weight of the mice were recorded every two days. Tumor volume was determined using the formula: , where L and W denote the longest and shortest diameters of the tumor, respectively. Following 10 days of treatment, mice were euthanized humanely using sodium pentobarbital overdose (150 mg/kg, intraperitoneal injection), and major organs were also collected and stained with H/E.

Statistical Analysis

Data are presented as mean ± standard deviation. Statistical significance was evaluated using two-way analysis of variance, with p < 0.05 considered significant.

Results and Discussion Characterization of Bioactive Glasses

BG and gallium-doped BG (GaBG) nanoparticles were synthesized under optimized conditions. TEM images (Figure 1A) confirmed the mesoporous structure of the bioactive glasses. Notably, the particle sizes of the GaBG nanoparticles did not differ significantly from those of the BG nanoparticles. Upon doping with Ga ions, the pore structures of the resulting 5GaBG, 10GaBG, and 15GaBG nanoparticles underwent significant changes. The addition of Ga2O3 reduced mesoporous order, likely due to the dual network-forming and -modifying behavior of Ga ions in silicate systems.19 Scanning electron microscopy (SEM) images (Figure 1B) further revealed that the nanoparticles were spherical with a smooth surface, exhibiting good dispersity and an average particle size of approximately 350 nm. The N2 adsorption-desorption isotherms (Figure 1C) exhibited type IV isotherms with type H3 hysteresis loops, confirming the existence of mesopores in the nanoparticles. The BJH pore size distribution (Figure 1D) revealed that the mesopore sizes ranged from 3.53 nm to 3.61 nm. Given the hierarchical porosity of the structures, the bioactive glass nanoparticles were characterized by a large pore volume and a high specific surface area (Table 1). However, the incorporation of Ga2O3 led to a marked reduction in the pore volume and surface area of the bioactive glass, which is consistent with the observations reported by Aina et al.27 DLS measurements exhibited that these particles negatively charged surfaces with a zeta potential ranging from –14.5 mV to –11.9 mV, which aids in their uniform distribution within chitosan-based hydrogels (Table 1). The mean hydrated size of the nanoparticles ranged from 564 nm to 783 nm, with a small polydispersity index (PDI). Elements present in the BG and GaBG nanoparticles were detected by EDS (Figure 1E), and the quantitative compositions of the nanoparticles were determined to be 77SiO2-23CaO, 71SiO2-17CaO-12Ga2O3, 70SiO2-14CaO-16Ga2O3, and 64SiO2-17CaO-19Ga2O3 (mol%), respectively.

Figure 1 Characterization of bioactive glasses nanoparticles. TEM images of nanoparticles (A), and their respective SEM images (B), nitrogen adsorption-desorption isotherms (C), pore size distributions (D), EDX spectra (E), scale bar:200μm.

Gel Formation and Rheological Characteristics

To determine the maximum content of NPs that can be added to hydrogels, various compositions for chitosan-N-acetylcysteine/silk fibroin/bioactive glass (CS-NAC/SF/BG) solutions were formulated. The gelation times of the resulting solutions were then examined, as presented in Table 2 and Supplementary Table 1. The CS-NAC/SF/BG solutions were found to be capable of forming hydrogels and the gelation time was significantly influenced by the addition of BG. Specifically, when the BG concentration exceeded 1.5%, the CS-NAC/BG solution formed a hydrogel during stirring, indicating an accelerated gelation process. Given that the gelation time of 1.5BG/CNS was around 50 seconds, it is unsuitable for practical clinical applications. Based on the above observations, to ensure suitable gelation time and mechanical properties, the concentration of NPs in the hydrogel solution was set at 1% (w/v).

The parameters of the five hydrogel formulations are summarized in Table 2. Incorporating BG NPs significantly reduced the gelation time relative to the CNS group. This effect is attributed to the release of alkaline ions from BG, which increases the solution pH and promotes thiol ionization, thereby accelerating disulfide bond formation.28,29 Conversely, Ga incorporation lowers the pH by increasing hydrogen ion concentration,30 leading to prolonged gelation times. As the concentration of Ga ions increases, the pH of the bioactive glass in aqueous solution typically decreases, leading to an extension in gelation time.

The sol-gel transition of the hydrogel can be observed in Figure 2A. The pre-gel solution at 37°C can be completely gelled within 5 min during the gelation. In general, the frequency dependence of G′ within a relatively low frequency range is important to evaluating the strength of hydrogels.31,32Figure 2B and C show the variation of G′ and G” against frequency sweep for five kinds of hydrogels, and their respective average G′ at 1 Hz is detailed in Figure 2D. The G′ value is dependent on the composition of hydrogel, and the addition of bioactive glass NPs could notably enhance their G′. BG/CNS had the largest G′ value at close to 4 kPa, which was attributed to the higher pH of the CS-NAC/SF/BG solution causing a small portion of the CS-NAC molecular chain to precipitate, thus enhancing the strength of this hydrogel. Furthermore, the G′ of xGaBG/CNS (x = 5, 10, 15) increased with Ga content. Gallium ions coordinates with the amino or hydroxyl groups in chitosan molecules to form stable complexes, enhancing the mechanical properties of the gel.33 Compared with CS- or SF-based injectable hydrogels, which often suffer from insufficient mechanical strength or overly rapid degradation, the present CS-NAC/SF/GaBG system achieves a more balanced integration of injectability, mechanical robustness, and structural stability.12,34 The ratio of G′ to G″ is commonly employed to assess gel strength. Hydrogels exhibiting high mechanical strength are characterized by large G′ values, typically exceeding G″ by one to two orders of magnitude.35,36 As shown in Figure 2B–D, the hydrogels exhibited G′ values of approximately 2.2, 3.6, 2.8, 2.9, and 3.3 kPa, with corresponding G′/G″ ratios of about 27.2, 14.5, 22.3, 23.6, and 25.3, indicating that all hydrogels displayed the characteristics of strong gels.

Figure 2 Characterizations of hydrogels. Sol-gel transformation of hydrogels (A); frequency-dependent changes of G′ and G′′ (B and C), average values of G′ at 1 Hz (D), shear-rate dependent variations of viscosity ((E): 25°C; (F): 37°C for different gels), and strain sweep spectra (G) (*p < 0.05; **p < 0.01; ns, no significant difference; see Table 2 for their parameters).

To examine the injectability of the hydrogels, the shear-rate dependence of the solutions’ viscosity was tested at 25°C and 37°C, respectively. The results are presented in Figure 2E and F. At 25°C, the viscosity of the hydrogels was lower than 50 pas, which decreased as shear rate increased. At 37°C, the viscosity of all solutions increased significantly at lower shear rates compared to their corresponding measurements at 25°C. However, a decrease in viscosity was observed at shear rates above 10−1 s, indicating that the shear-thinning behavior of the hydrogels was preserved. Given that gel injections are typically performed at room temperature, these results suggest that all hydrogels possess satisfactory injectability.

The magnitude of G′ and G′′ as a function of percent strain of hydrogels provides insight into the elasticity of the hydrogels.36–38 Generally, the crossover points of G′ and G” observed in a strain sweep correspond to the critical strain, or yielding strain, which is commonly used to evaluate hydrogel elasticity. At this point, the applied strain becomes out of phase with the stress, indicating a disruption of the hydrogel’s three-dimensional network.37,38Figure 2G indicates that the five types of hydrogels had various yielding strains, for CNS, BG/CNS, 5GaBG/CNS, 10GaBG/CNS, and 15GaBG/CNS, the yielding strains were approximately 91%, 78%, 89%, 81% and 59% respectively. The results reveal that although the addition of NPs reduced the critical strain, all the hydrogels maintained good elasticity.

Analysis of Dry Gels

Dry gel samples were systematically characterized using SEM to evaluate this critical structural parameter. Representative cross-sectional images (Figure 3A–E) revealed three-dimensional networks of highly interconnected pores across all formulations. These interconnected channels allow osteoblasts and bone marrow mesenchymal stem cells to readily infiltrate the hydrogel rather than grow only on the surface, which is crucial for the formation of complete three-dimensional tissue.39 Furthermore, endothelial cells can migrate along these pores to form new blood vessels, providing a fundamental basis for bone repair and tissue regeneration.40 While pore size distribution patterns indicated similarities between CNS and BG/CNS gels, a notable broadening of distribution intervals was observed in these two groups compared to the GaBG-containing formulations. This morphological variation suggests that gallium ions influence the gelation process. Generally, scaffolds used for bone regeneration should have pore sizes greater than 100µm,41 and the average pore size of all dry gels meets this requirement, ensuring suitability for osteoblast growth. A comparison of average pore sizes revealed no significant differences (p > 0.05) among CNS, BG/CNS, 5GaBG/CNS, and 10GaBG/CNS gels, whereas a significantly larger average pore size (p < 0.01) was observed in the 15GaBG/CNS gels. This can be attributed to the higher concentration of Ga ions in 15GaBG, which, upon complexation with chitosan, leads to an uneven cross-linking network within the gel, resulting in larger pores in certain areas. A closer examination of the pores present in these gels reveals that BG/CNS, 5GaBG/CNS, 10GaBG/CNS, and 15GaBG/CNS exhibit more small pores along the pore walls compared to CNS, suggesting that the incorporation of BG and GaBG nanoparticles into the gels can affect the pore morphology and pore-size distribution of CNS hydrogels.

Figure 3 Properties of dried hydrogels. Pore-size distributions and representative images of dried hydrogels (A–E); swelling indexes of dried hydrogels (F); EDS spectra (G–J), FTIR spectra (K) for CS-NAC, SF, BG, CNS, 5GaBG/CNS, 10GaBG/CNS and 15GaBG/CNS gel; ion release patterns for 15GaBG/CNS gel (L). (see Table 2 for their parameters; scale bar: 200μm).

The complementary swelling index (SI) measurements (Figure 3F) provided further insights into pore network functionality. All dried hydrogels exhibited high SI, with no significant difference observed among gels of different compositions (p > 0.05). Dry hydrogels with high porosity and “open-cell” pore structures promote rapid water convection, resulting in elevated SI values and short times need to reach swelling equilibrium.42 In our case, each gel reached its swelling equilibrium in under 30 minutes, confirming the “open-cell” pore characteristics observed in SEM. The comparable SI across all compositions suggests that, despite subtle differences in pore size distribution, the overall network openness and hydrophilicity remain largely unaffected by BG or GaBG loading.

Compositional analysis through EDS (Figure 3G–J) confirmed successful incorporation of bioactive glass components, with the peak area of gallium ions Ga3⁺ proportionally with gallium content in the bioactive glass. FTIR characterization (Figure 3K) provided critical evidence of physical blending among CS-NAC, SF, and bioactive glass within the hydrogel system. For BG NPs, a broad band observed around 1099 cm−1 corresponds to the Si-O-Si symmetric stretching, while another band at 810 cm−1 is attributed to the Si-O-Si rocking mode, indicating the presence of a silica network from BG. In the case of CNS, a shoulder at approximately 1656 cm−1 can be attributed to the C=O stretching vibration of CS, which is typically observed in highly deacetylated CS.43 Two bonds at 1535 cm−1 and 1231 cm−1 indicate the presence of amide II and random coil structures of SF, respectively. Characteristic peaks of chitosan and silk fibroin are observed in the FTIR spectra of CNS. The peaks of CS-NAC and SF remain unchanged, indicating that CS-NAC and SF coexisted in the gel through physical blending. Furthermore, after mixing with bioactive glass nanoparticles, characteristic absorption peaks of bioactive glass were observed in BG/CNS, 5GaBG/CNS, 10GaBG/CNS, and 15GaBG/CNS gel samples. Importantly, the absorption peaks of bioactive glass in these samples did not shift, suggesting that bioactive glass is physically blended within the gel matrix. This physical integration mechanism offers distinct benefits for therapeutic ion delivery. The absence of strong chemical bonding between GaBG nanoparticles and the polymer matrix suggests the potential for sustained ion release through diffusion-controlled mechanisms, while maintaining structural integrity through physical entanglement.

Ion Release Patterns

The ion release behavior of 15GaBG/CNS gel was evaluated due to the gel’s higher Ga ions content (see Table 2). As depicted in Figure 3L, Si ions were released gradually throughout the entire sampling period, without exhibiting a burst release (approximately 39 mg/mL at 28 days). Both calcium and gallium ions were released in a nearly linear manner for over four weeks, reaching about 11.24 mg/mL and 16.8 mg/mL, respectively, at 28 days. Notably, during the first seven days, the release rate of calcium ions was higher than that of gallium ions; however, the gallium ion release rate subsequently increased and surpassed that of calcium ions. This phenomenon is partly due to the higher Ga ions content compared to Ca ions in 15GaBG (64SiO2-17CaO-19Ga2O3 mol%). Additionally, gallium ions act as both network formers and modifiers in GaBG nanoparticles,27,44 and as the glass network degrades with silicon ion release, both forms of gallium are liberated, resulting in a higher gallium ion concentration than calcium in the solution. The absence of burst release for any ions underscores the CNS matrix’s role in modulating diffusion, likely through hydrogen bonding with leached ions—a property critical for minimizing cytotoxicity during the acute post-implantation phase. These results suggest that 15GaBG/CNS could release Si, Ca, and Ga ions for up to at least 28 days under physiological conditions, which is crucial for its physiological functions in cells and tissues.

In vitro Apatite Formation

Figure 4 presents the SEM micrographs and EDS spectra of the hydrogels after 28 days of immersion in simulated body fluid (SBF). After 28 days, all gel samples developed surface deposits with the characteristic cauliflower-like morphology of hydroxyapatite (HA) (Figure 4A, C, E, and G), a feature commonly seen on bioactive glasses treated with SBF.45 As shown in Figure 4B, D, F, and H, the EDS spectrum of the gels mineralized in SBF exhibited additional phosphorus (P) peaks compared with those of the unmineralized gels (Figure 3G–J). Notably, the Ca/P molar ratio of the deposited apatite was higher on the surfaces of BG/CNS and 5GaBG/CNS gels than on those of 10GaBG/CNS and 15GaBG/CNS. This inhibitory effect on apatite formation is consistent with Ga’s function as a network modifier in bioactive glasses. When gallium ions substitute for silicon atoms in the glass network, negative charges are generated.44 These charges are neutralized by protons, forming Brønsted acid sites [Si(OH)+Ga−]. The resulting protonic acidity increases the acidity of both the glass surface and the surrounding SBF, thereby affecting the glass’s reactivity and bioactivity.46 Additionally, the reduced micropore volume and surface area (as shown in Table 1) contribute to the delayed growth of the apatite-like layer in the SBF medium. Supplementary Figure 1 shows the FTIR spectra of the gels after 28 days of immersion in SBF. The absorption bands at approximately 596 and 604 cm−1 were assigned to P-O vibrations in PO43− groups of crystalline Ca-P. Additionally, one band near 874 cm−1 and two characteristic bands around 1400 and 1440 cm−1 were attributed to C-O stretching vibrations, indicating the presence of carbonated hydroxyapatite.47 These results indicate that BG- and GaBG-incorporated gels promote apatite formation under physiological conditions.

Figure 4 SEM images and EDS spectra of gel samples after being immersed in SBF for 28 days. BG/CNS (A and B); 5GaBG/CNS (C and D); 10GaBG/CNS (E and F) and 15GaBG/CNS (G and H) gels (scale bar: 10 μm).

Cell Viability

To investigate the impact of Ga ions released from GaBG/CNS composite gels on the growth and proliferation of both osteoblasts and osteosarcoma cells, MC3T3-E1 pre-osteoblast cells and UMR-106 osteosarcoma cells were cocultured with hydrogels. Representative fluorescence images of stained MC3T3-E1 cells cultured with gels for varying durations are shown in Figure 5A. Notably, minimal cytotoxicity was observed across all groups after three and seven days of culture, with cell density markedly increasing by day 7 compared to day 3. By day 7, the cells exhibited elongated, spindle-like morphology with pronounced cellular extensions and robust adhesion to the gel matrix, indicative of active proliferation and extracellular matrix integration. These findings suggest that hydrogels not only support long-term cell viability but also promote osteoblast maturation and substrate interaction. The proliferation assays of MC3T3-E1 cells cultured on the gels for 24, 48 and 72 hours were quantified (Figure 5B). Two key findings can be drawn from this data: (1) After 24 hours of culture, cell viability in the CNS group was 99.9%, while the BG/CNS, 5GaBG/CNS, 10GaBG/CNS, and 15GaBG/CNS groups all exhibited viabilities approaching 90%; (2) after 72 hours of culture, the cell viability of the 15GaBG/CNS group remained above 80% without changing the culture medium, whereas the viability of the other groups remained consistent with their 24- and 48-hour values. Collectively, these results confirm the excellent biocompatibility of all tested hydrogels. Importantly, the Ga ion concentrations eluted from 5GaBG/CNS, 10GaBG/CNS, and 15GaBG/CNS gels were within thresholds deemed non-cytotoxic, as evidenced by sustained MC3T3-E1 proliferation.

Figure 5 Representative confocal images and cell viability of MC3T3-E1 cell (A and B) and UMR-106 cell (Live/Dead-staining) (C and D). Images represent z-projections of a 100 µm z-stack, (viable cells: green; dead cells: red; scale bar: 200 μm; *p < 0.05; **p < 0.01; ns, no significant difference; n=6).

In stark contrast to osteoblasts, UMR-106 osteosarcoma cells displayed pronounced dose- and time-dependent cytotoxicity in response to Ga ions (Figure 5C and D). After three days of culture, CNS and BG/CNS gels exhibited negligible cytotoxicity, with densely packed, viable osteosarcoma cells. However, in the 5GaBG/CNS, 10GaBG/CNS, and 15GaBG/CNS groups, cell density dropped significantly, and more dead cells were observed in these groups relative to the other groups. Furthermore, higher gallium ion levels in the gel led to an increase in dead cells. After seven days of culture, the CNS and BG/CNS gels still showed very few dead cells with well-spread, viable cells, whereas a significant increase in dead cells was observed in the 5GaBG/CNS, 10GaBG/CNS, and 15GaBG/CNS groups. In fact, nearly all cells in the 10GaBG/CNS and 15GaBG/CNS gels appeared to be dead. Figure 5D illustrates the proliferation of UMR-106 cells cultured with the gel for 24, 48, and 72 hours. At 24 hours, viability in the 5GaBG/CNS group dropped to 78%, declining further to 65% by 72 hours. Higher Ga concentrations (10GaBG/CNS and 15GaBG/CNS) caused more drastic effects: viabilities plummeted to 45% and 33% at 24 hours, respectively, with the 15GaBG/CNS group showing near-complete cytotoxicity (<5% viability) by 72 hours. No significant reduction in UMR-106 cell viability was observed between the CNS and BG/CNS, confirming Ga-specific toxicity. The selective cytotoxicity of Ga ions toward UMR-106 cells aligns with prior studies demonstrating Ga’s ability to disrupt iron-dependent pathways in cancer cells, inducing apoptosis via mitochondrial dysfunction and oxidative stress.48,49 Remarkably, this anticancer activity occurs without compromising osteoblast viability, even at the highest Ga concentration (15GaBG/CNS), suggesting a wide therapeutic window. Considering that these gels are used in bone regeneration and osteosarcoma inhibition and showed strong cytotoxicity toward cancer cells but no toxicity to normal pre-osteoblasts, 10GaBG/CNS and 15GaBG/CNS were selected for follow-up animal assessments.

The osteogenic potential of the gels was assessed by measuring the ALP activity of MC3T3-E1 cells (Supplementary Figure 2), as ALP is a widely used marker of early osteogenic differentiation. After 7 days, no significant differences were observed among the groups. By day 14, ALP activity in the CNS group remained largely unchanged, whereas BG/CNS, 5GaBG/CNS, 10GaBG/CNS, and 15GaBG/CNS gels all induced significantly higher ALP activity, with the 5GaBG/CNS group exhibiting the most pronounced effect. The enhanced ALP activity in the 5GaBG/CNS group may arise from Ga’s dual role in promoting osteoblast differentiation and inhibiting osteoclast resorption, as reported in studies on Ga-doped biomaterials.50,51 Conversely, the diminished ALP response at higher Ga concentrations likely reflects a prioritization of cytotoxic mechanisms over differentiation signal—a tradeoff critical for balancing bone repair and tumor suppression in clinical applications.

Collectively, the synthesized CNS and BG/CNS hydrogels demonstrate clear potential for bone tissue engineering due to their injectability, tissue-adaptive mechanics, and bioactive properties. In particular, gallium-doped variants provide additional local antitumor effects, enabling dual functionality. Furthermore, all components—including thiolated chitosan, silk fibroin, and bioactive glass—are readily available, inexpensive, and easy to scale up, supporting potential clinical translation.

In vivo Bone Repair

The 3D reconstructed micro-CT images of excised calvarial bones revealed minimal new bone formation within the defect area of the blank control group at 8 weeks post-operation (Figure 6A–E). This finding aligns with those of previous studies,52 which indicate that critical-sized defects are difficult to heal naturally without human intervention. In contrast, defects treated with CNS hydrogel displayed sporadic osteoid tissue formation along the defect periphery, suggesting partial osteogenic potential. Notably, the BG/CNS, 10GaBG/CNS, and 15GaBG/CNS groups exhibited substantial enhancements in bone regeneration compared to both the blank and CNS groups, highlighting the synergistic role of bioactive glass (BG) components. Quantitative analysis of residual defect areas (Figure 6F) revealed that the 15GaBG/CNS group exhibited a residual defect area of 55% ± 5%, which was comparable to other BG-containing hydrogel groups, suggesting that Ga substitution did not adversely affect the osteoconductive properties of the hydrogel. This aligns with previous reports that controlled release of silicon and calcium ions from BG-based systems stimulates osteoblast proliferation and extracellular matrix mineralization,53 while Ga ions, which have historically been associated with anti-osteoclastic activity,19 may synergistically enhance bone remodeling without cytotoxic effects. Morphometric parameters, including bone volume fraction (BV/TV), trabecular number (Tb. N), and trabecular thickness (Tb. Th) were quantitatively assessed (Figure 6G–I). All BG-incorporated hydrogels significantly outperformed CNS hydrogel in all metrics (p < 0.05), although inter-group variations among BG formulations lacked statistical significance. These findings collectively validate the notion that BG functionalization—regardless of Ga doping levels—effectively augments trabecular microarchitecture, a critical determinant of biomechanical competence in regenerated bone.

Figure 6 Micro-CT images and quantitative analyses of BV/TV, trabecular number, and thickness in hydrogel-treated calvarial defects. (A) blank control; (B) CNS; (C) BG/CN; (D) 10GaBG/CNS; (E) 15GaBG/CNS; (F) remaining defect area; (G) BV/TV ratio; (H) trabecular number; (I) trabecular thickness (repair period: 8 weeks; scale bar:5mm; **p < 0.01; ns, no significant difference, n=6).

After eight weeks, the regenerated tissue at the defect site, along with the surrounding bone, was collected and analyzed histologically using H/E and Masson’s trichrome staining (Figure 7). The histological analysis confirmed that all hydrogel groups exhibited excellent histocompatibility, with no signs of tissue rejection. Minimal amounts of newly formed collagen and bone were observed in the CNS gel-treated defects, whereas the BG/CNS, 10GaBG/CNS, and 15GaBG/CNS groups showed more pronounced collagen and new bone formation. H/E staining further revealed that bioactive glass-loaded hydrogels facilitated rapid new bone growth, progressing from the surface inward through macropores that served as migration channels. These histological findings are consistent with micro-CT observations.

Figure 7 H/E staining ((a–d)and (a(i)–d(i))) and Masson’s trichrome staining ((e–h) and e(i)-h(i))) of regenerated bone tissues.

Abbreviations: NB, new bone; FT, fibrous tissue; IM, implant materials.

In vivo Antitumor Efficiency

In vivo antitumor efficacy was evaluated in UMR-106 osteosarcoma-bearing mice treated with various hydrogels, with results summarized in Figure 8. Notably, the 15GaBG/CNS group exhibited the most pronounced tumor suppression compared to those treated with the control, CNS, BG/CNS, and 10GaBG/CNS gels (Figure 8A and B). While the blank control group displayed exponential tumor progression, CNS hydrogel treatment transiently retarded tumor expansion through physical confinement—a mechanical barrier effect that diminished as tumor cells breached the hydrogel periphery over time. Intriguingly, BG/CNS and 10GaBG/CNS hydrogels paradoxically enhanced tumor growth relative to the CNS group. This is likely attributable to the effects of sustained Ca2+ and Si4+ release, which can enhance cellular metabolism and angiogenesis and may support osteosarcoma growth. Similarly, the released Ga ions from the 10GaBG/CNS group did not reach a dose sufficient for inhibiting tumor growth. The tumor volume in the 10GaBG/CNS group may have been insufficient to overcome the pro-proliferative effects, leading to a slightly larger tumor volume than that in the CNS group, although the difference was not statistically significant. In contrast, a significant reduction in tumor volume was observed in the 15GaBG/CNS group compared with the other four groups (Figure 8B). These findings suggest that Ga contents below this threshold may be insufficient for tumor inhibition and should be applied with caution in tumor-related indications. The high anticancer efficacy observed in the 15GaBG/CNS group can be attributed to the ability of Ga ions to combat cancer cells through various physiological interference mechanisms, such as inhibiting DNA replication and disrupting its helical structure. Transferrin receptors, which are overexpressed on tumor cell surfaces but exhibit low expression in normal tissue cells, are involved in cellular iron uptake through interactions with transferrin.48,49,54 Ga ions can competitively bind to transferrin, which primarily transports Fe3+, thereby inhibiting the activity of ribonucleotide reductase and, consequently, suppressing DNA replication. This ultimately leads to the inhibition of tumor cell growth.55,56 Images of the excised tumors (Figure 8C) and tumor weight measurements (Figure 8D) further demonstrated that tumor growth was significantly more inhibited in the 15GaBG/CNS group than in the other groups. No significant differences in body weight were observed after 10 days of treatment (Figure 8E), indicating that the GaBG-incorporated gels developed in this study exhibit safety.

Figure 8 In vivo antitumor evaluation. Photographs of UMR-106 tumor-bearing mice after 10 days of treatment with different hydrogel samples (A); tumor volume changes over time during treatment (B); photos of excised tumors following treatment (C); average weights of excised tumors (D); body weight changes (E); representative H/E-stained micrographs of major organs excised from treated mice (F). (organs harvested after 10 days; scale bar=100μm; **p < 0.01; ns, no significant difference).

Histological analyses of the heart, liver, spleen, lungs, and kidneys were performed using H/E staining, and the results are shown in Figure 8F. Micrographs of tissues from CNS, BG/CNS, 10GaBG/CNS, and 15GaBG/CNS-treated mice revealed normal histological structures comparable to those in the control group. These findings indicate that the GaBG-embedded gels did not induce detectable damage to the major organs of the treated mice.

Conclusion

A novel nanocomposite hydrogel was successfully developed using CS-NAC, SF and GaBG for applications in bone regeneration and osteosarcoma suppression. Hydrogel solutions containing BG or GaBG were shown to be injectable at both ambient and physiological temperatures, forming stable gels under physiological pH conditions. The incorporation of bioactive glass components significantly enhanced the mechanical strength and stiffness of the hydrogels, while their highly porous and interconnected structures endowed them with pronounced apatite-forming ability in vitro. The optimized gels enabled sustained and controlled release of Si, Ca, and Ga ions, while serving as injectable scaffolds that supported MC3T3-E1 cell proliferation and osteogenic differentiation, and simultaneously inhibited UMR-106 cell growth. In the vivo, the optimized cell-free gel successfully repaired critical-size calvarial bone defects in a rat model without the use of exogenous growth factors. Furthermore, results obtained from UMR-106-tumor-bearing mice proved that 15GaBG/CNS significantly inhibited advanced tumor growth, showing superior antitumor efficacy. Despite these promising results, long-term in vivo biosafety, bone remodeling, and systemic metabolism of Ga ions still need to be evaluated, and its potential antitumor mechanisms also require further investigation. Future studies will focus on mechanistic elucidation and optimization of hydrogel formulations for clinically relevant large bone defects and post-tumor-resection models. Overall, GaBG-based injectable nanocomposite hydrogels represent a multifunctional platform for simultaneous bone regeneration and osteosarcoma suppression.

Funding

This work was supported by the National Natural Science Foundation of China (81972065), the National Key R&D Program of China (Grant No. 2017YFC1104402) and the Higher Education Scientific Research Project of Anhui Province (2023AH050619).

Disclosure

The author(s) report no conflicts of interest in this work.

References

1. Kansara M, Teng MW, Smyth MJ, et al. Translational biology of osteosarcoma. Nat Rev Cancer. 2014;14(11):722–735. doi:10.1038/nrc3838

2. Wang C, Zhang Y, Kong W, et al. Delivery of miRNAs using nanoparticles for the treatment of osteosarcoma. Int J Nanomed. 2024;19:8641–8660. doi:10.2147/IJN.S471900

3. Liu X, Zhang Y, Wu H, et al. A conductive gelatin methacrylamide hydrogel for synergistic therapy of osteosarcoma and potential bone regeneration. Int J Biol Macromol. 2023;228:111–122. doi:10.1016/j.ijbiomac.2022.12.185

4. Li C, Zhang W, Wang R, et al. Nanocomposite multifunctional hydrogel for suppressing osteosarcoma recurrence and enhancing bone regeneration. Chem Eng J. 2022;435:134896. doi:10.1016/j.cej.2022.134896

5. Rahim SA, Bakhsheshi-Rad HR, Licavoli J, et al. Overview of biodegradable materials for bone repair and osteosarcoma treatment: from bulk to scaffolds. Biomater Adv. 2025;174:214317. doi:10.1016/j.bioadv.2025.214317

6. Wang Y, Wen J, Lu T, et al. Mesenchymal stem cell-derived extracellular vesicles in bone-related diseases: intercellular communication messengers and therapeutic engineering protagonists. Int J Nanomed. 2024;19:3233–3257. doi:10.2147/IJN.S441467

7. Chu C, Qiu J, Zhao Q, et al. Injectable dual drug-loaded thermosensitive liposome-hydrogel composite scaffold for vascularised and innervated bone regeneration. Colloids Surf B. 2025;245:114203. doi:10.1016/j.colsurfb.2024.114203

8. Lin H, Zhang L, Ye X, et al. The surface modification of 3D-printed polyether ether ketone with bioactive hydrogel for bone repair. Colloids Surf B. 2025;254:114846. doi:10.1016/j.colsurfb.2025.114846

9. Hamza HM, Malik MM, Asad M, et al. Advances in orthopedic implants: the role of nanotechnology in enhancing performance and longevity. Regenerative Med Reports. 2025;2(1):15–21. doi:10.4103/REGENMED.REGENMED-D-24-00024

10. Scolari IR, Granero GE. Narrative review of the opportunities for bone tissue regeneration and osteomyelitis treatment: transition metal complexes with antibiotics as proof-of-concept. Regenerative Med Reports. 2025;10:4103.

11. Sun F, Cui Y, Zhi W, et al. Mechanisms and clinical advances in the in vivo regeneration of renal tissue: a narrative review. Regenerative Med Reports. 2025;2(4):149–160. doi:10.4103/REGENMED.REGENMED-D-25-00003

12. Zhang K, Liu Y, Zhao Z, et al. Magnesium-doped nano-hydroxyapatite/polyvinyl alcohol/chitosan composite hydrogel: preparation and characterization. Int J Nanomed. 2024;19:651–671. doi:10.2147/IJN.S434060

13. Chen S, Wang Y, Zhang X, et al. Double-crosslinked bifunctional hydrogels with encapsulated anti-cancer drug for bone tumor cell ablation and bone tissue regeneration. Colloids Surf B. 2022;213:112364. doi:10.1016/j.colsurfb.2022.112364

14. Liu J, Yang B, Li M, et al. Enhanced dual network hydrogels consisting of thiolated chitosan and silk fibroin for cartilage tissue engineering. Carbohydr Polym. 2020;227:115335. doi:10.1016/j.carbpol.2019.115335

15. Zhu H, Monavari M, Zheng K, et al. 3D bioprinting of multifunctional dynamic nanocomposite bioinks incorporating Cu-Doped mesoporous bioactive glass nanoparticles for bone tissue engineering. Small. 2022;18(12):2104996. doi:10.1002/smll.202104996

16. Nikolova V, Angelova S, Markova N, et al. Gallium as a therapeutic agent: a thermodynamic evaluation of the competition between Ga3+ and Fe3+ ions in metalloproteins. J Phys Chem B. 2016;120(9):2241–2248. doi:10.1021/acs.jpcb.6b01135

17. Kircheva N, Dudev T. Novel insights into gallium’s mechanism of therapeutic action: a DFT/PCM study of the interaction between Ga3+ and ribonucleotide reductase substrates. J Phys Chem B. 2019;123(26):5444–5451. doi:10.1021/acs.jpcb.9b03145

18. Rana K, Souza L, Isaacs MA, et al. Development and characterization of gallium-doped bioactive glasses for potential bone cancer applications. ACS Biomater Sci Eng. 2017;3(12):3425–3432. doi:10.1021/acsbiomaterials.7b00283

19. Gómez-Cerezo N, Verron E, Montouillout V, et al. The response of pre-osteoblasts and osteoclasts to gallium containing mesoporous bioactive glasses. Acta Biomater. 2018;76:333–343. doi:10.1016/j.actbio.2018.06.036

20. Li F, Liu F, Huang K, et al. Advancement of gallium and gallium-based compounds as antimicrobial agents. Front Bioeng Biotechnol. 2022;10:827960.

21. Pajor K, Michalicha A, Belcarz A, et al. Antibacterial and cytotoxicity evaluation of new hydroxyapatite-based granules containing silver or gallium ions with potential use as bone substitutes. Int J Mol Sci. 2022;23(13):7102. doi:10.3390/ijms23137102

22. Zheng H, Huang Z, Chen T, et al. Gallium ions incorporated silk fibroin hydrogel with antibacterial efficacy for promoting healing of Pseudomonas aeruginosa-infected wound. Front Chem. 2022;10. doi:10.3389/fchem.2022.1017548

23. Lu B, Zhu D, Yin J, et al. Incorporation of cerium oxide in hollow mesoporous bioglass scaffolds for enhanced bone regeneration by activating the ERK signaling pathway. Biofabrication. 2019;11(2):025012. doi:10.1088/1758-5090/ab0676