Low molecular weight (5,000–10,000 dalton) hyaluronic acid (HA) was acquired from Shandong Freda Biotechnology (Shandong, China). Dimethyl sulfoxide, deuterium oxide, and DMSO were acquired from Sinopharm Chemical Reagent (China). Dialysis bags (7,000 kDa and 8,000–14,000 kDa) were acquired from Shanghai Yuanye Bio-Technology (Shanghai, China). Polyethylene glycol (PEG, 8,000 dalton), tetrabutylammonium hydroxide (TBA, 40% in H2O), and N,N’-dicyclohexylcarbodiimide (DCC) were acquired from Aladdin Reagent (Shanghai, China). Dulbecco’s modified Eagle’s medium/high modified (DMEM-H), phosphate buffer saline (PBS), antibiotic–antimycotic, and fetal bovine serum (10099-141) were acquired from Gibco (United States). Red Blood Cell Lysis Buffer (C3702) and Cell counting kit-8 were acquired from Beyotime Biotechnology (Shanghai, China). Sodium zoledronate was acquired from Solarbio (Beijing, China). Nano hydroxyapatite was acquired from Aladdin Reagent (Shanghai, China). CD68-PE, CD3-FITC, and CD8-APC antibodies were acquired from BioLegend (United States). LIVE/DEAD Cell Imaging Kit was acquired from Dojindo (Japan). CD68 and CD3 antibodies were acquired from Santa-Cruz (United States). Anti-goat and anti-rabbit secondary antibodies were acquired from Bosterbio (United States). Alanine aminotransferase ALT assay kit (C052), aspartate aminotransferase AST assay kit (C072), alkaline phosphatase assay kit (C003), blood urea nitrogen determination kit (C010), and creatinine assay kit (C074) were acquired from Changchun Huili Biotech (Changchun, China).

A total of 293T cells were acquired from ATCC and cultured in DMEM supplemented with 10% fetal bovine serum (FBS), 1% penicillin and streptomycin at 37°C, and 5% CO₂ (Heal Force, China). A culture dish, 24-well plates, and 96-well plates were acquired from Corning. C57BL/6 mice (6–8 weeks) were acquired from Hunan SJA Laboratory Animal Co., Ltd (Hunan, China). All mice were sacrificed by 1% pentobarbital sodium injection (80 mg/kg). The animal protocol was approved by the Institutional Animal Care and Use Committee (IACUC) of Hunan Cancer Hospital (KYJJ-2016-009).

Transmission electron microscopy (TEM) was used to perform morphological analysis of NPs. HA-PEG-nHA-ZOL NPs with 2% phosphorous acid stained were dropped on a copper mesh coated with a carbon support film. After drying, the morphology was observed under a transmission electron microscope.

Dynamic light scattering (DLS) was used to measure the size distribution and stability of NPs. The prepared HA-PEG-nHA NP and HA-PEG-nHA-ZOL NP nanoparticles were put into a cuvette, and we put the cuvette into a dynamic light scattering particle size analyzer for testing. Each sample tested for 3 times, 1 min each time. The test conditions were as follows: argon ion laser, wavelength 658 nm, temperature 25 ± 0.1°C, and dynamic light scattering angle 90 degrees. The zeta potential was simultaneously measured; the operating conditions are 11.4 v/cm, 13.0 mA, and 25°C; and the sample solvent was diluted with distilled water. DLS was measured for three times and the average calculated.

A small amount of solid nHA, HA-TBA, and HA-PEG-nHA NP samples was taken out and mixed with dried potassium bromide powder, after being grinded and pressed into a transparent sample potassium bromide tablet, and we measured the infrared spectrum of the sample.

HA-PEG-nHA-ZOL NPs and 10 mg each of ZOL were taken and dispersed in 1 ml of deuterated heavy water, and a nuclear magnetic resonance proton spectrometer was used to calculate the drug loading of HA-PEG-nHA-ZOL NPs according to the following formula:

Where i = ZOL or HA-PEG-nHA -ZOL NPs, A_{δ 7}.₂₆ and A_{δ 7}.₄₁ are both a characteristic hydrogen peak on the imidazole group. Measured NMR for three times and calculate the average.

A concentration-absorbance standard curve was established for ZOL phosphoric acid at 210 nm. A total of 20 mg of HA-PEG-nHA -ZOL NPs was accurately weighed and dispersed in 10 mL of deionized water under an ice bath. Then, this liquid was placed in a dialysis bag (10,000 kDa), and the dialysis bag was placed in a PBS/HCl–PBS buffer solution with pH of 7.4 and 5.8, respectively, and protected from light at 37°C and 100 rpm. Within a certain time, the entire release medium is replaced with fresh medium, and the old medium is retained. The ultraviolet-visible spectrophotometer was used to measure the ZOL concentration in old media, and the release rate was calculated according to the following formula:

Where t = 0, 0.5, 1, 2, 4, 8, 12, 16, 24, 36, 48, 60, 72 h and M_t is the product of the old medium volume and the concentration.

Cell viability was tested by the cell counting kit-8 (CCK-8). A total of 8,000 of 293T cells were seeded into 96-well plates, each absorbance group for 4 wells. The groups are listed in Table 1. Notably, NPs that had the concentrations lower than 2.5 μg/ml were considered as “low-dose” for cells; NPs that had the concentrations higher than 5 μg/ml were considered as “high-dose” for cells.

One day later, drugs were added into the wells and culture for 2 days, and 10 μl CCK-8 solution was added into each well and cultured for 3 h. A microplate reader (Tecan, Spark, Switzerland) was used to measure the absorbance at 450 nm, and the cell viability was calculated according to the following formula:

A total of 293T cells were cultured in a 24-well plate. After culturing for 1 day, the cells were treated with drugs, each group having 2 wells. Groups are listed in Table 2.

After a 2-day culture, a 200-μl culture medium which contains Calcein-AM (CAM) and PI (1:500 diluted) was added to each well of 293T cells. After 30 min of incubation, the cells were observed under a fluorescence microscope (Leica DMi8, Germany).

Nanoparticles were caudal intravenously injected into mice. The concentrations and doses of NPs used in the current study are listed in Table 3.

Healthy mice were treated with NPs in different concentrations (details in Table 3). Notably, NPs that had concentrations lower than 250 μg/ml were considered as “low-dose” for mice; NPs that had concentrations higher than 500 μg/ml were considered as “high-dose” for mice.

After 3 days, mice were sacrificed and organs including the kidney, lung, heart, liver, and brain were fixed with 4% paraformaldehyde overnight. The organs were then dehydrated and embedded in paraffin. The organs were then sectioned into 5-μm slices and stained with hematoxylin and eosin (H&E). The stained sections were imaged under an inverted phase-contrast microscope (Olympus CX23, Japan).

The lung, liver, spleen, and kidney were collected and embedded in paraffin 3 days after injection. After sectioning into 5-μm slices, CD3 and CD68 immunohistochemistry (IHC) staining was performed. For IHC staining, dewaxing and rehydration of paraffin sections were first treated with 3% hydrogen peroxide in methanol solution for 15 min. After PBS flush, the sections were blocked with 10% secondary antibody serum for 30 min. Diluted primary antibodies (1:100) were added and incubated overnight at 4°C. After PBS flush, the secondary antibody was added and incubated for 30 min at room temperature. Then, diaminobenzidine (DAB) coloration was performed and hematoxylin was used to restain nuclei. Dehydration and covering of the slide were performed.

Liver function was evaluated according to alanine aminotransferase (ALT), aspartate aminotransferase (AST), and alkaline phosphatase (ALP) level. Peripheral blood was collected from the caudal vein. After 10 min of 4000-rpm centrifuge, supernatant was collected and ALT, AST, and ALP assay was performed. Serum ALT, AST, and ALP levels were determined by an automatic biochemical analyzer Chemray 240 (Rayto Life and Analytical Sciences, United States).

Kidney function was evaluated according to blood urea nitrogen (BUN) and creatinine (CR) level. Peripheral blood was collected from the caudal vein. After 10 min of 4000-rpm centrifuge, supernatant was collected and BUN and creatinine assay was performed. Serum BUN and creatinine levels were determined by an Automatic biochemical analyzer Chemray 240.

Peripheral blood was collected 3 days after drug delivery. RBC lysis buffer (5 ml) was added into 1 ml blood, lysed for 5 min, and centrifuged at 450 g and 4°C for 5 min (Eppendorf Centrifuge 5424 R, Germany). The supernatant was discarded, resuspended with 1 ml ice-cold PBS buffer, filtered through a cell strainer (70 μm), and centrifuged at 450 g and 4°C for 15 min. The supernatant was discarded and resuspended with 100 μl ice-cold PBS buffer. CD68-PE, CD3-FITC, and CD8-APC antibodies (1:100) were added and incubated at 4°C for 30 min and centrifuged at 450 g and 4°C for 5 min. The supernatant was discarded and resuspended with 200 μl ice-cold PBS buffer. Novo Express was used to perform flow cytometry analysis.

After the normality test, data subjected to normal distribution were analyzed by one-way-ANOVA followed by multiple comparisons between groups using Tukey’s post hoc test. A Mann–Whitney U test was used for analyzing flow cytometry data. All statistical analysis was carried out with GraphPad Prism 8.0 software. A p-value of < 0.05 is considered significant.

Briefly, the synthesis of HA-PEG-nHA-ZOL nanoparticles consisted of three steps. First, HA and TBA solutions were mixed, frozen, and vacuum freeze-dried to obtain the HA-TBA polymer. PEG and nHA were subsequently added into HA-TBA and ultrasonic dispersed to acquire HA-PEG-nHA nanoparticles. Finally, with ZOL drug loading, HA-PEG-nHA-ZOL nanoparticles were synthesized. The encapsulation process was checked by Fourier transform infrared spectrometry (FTIR), revealing the synthetic routes from HA to HA-TBA and finally HA-PEG-nHA (Figure 1B). The drug loading process and efficacy were measured by nuclear magnetic resonance (NMR). In Figures 1C.a,C.b, the peak at 4.65 ppm refers to the methylene peak on zoledronic acid sodium. In Figure 1C.a, peaks at 3 to 4 ppm refer to the methine peaks on HA and PEG, peaks at 1.9 to 2.6 ppm refer to the hydroxyl peaks on HA, and peaks at 7 to 8 ppm are characteristic peaks of hydrogen on the imidazole ring. According to the peak areas of the above peaks, free zoledronic sodium was used as the external standard, and the ZOL loading efficiency could be calculated, which was 41.3 ± 2.31%. Compared with traditional polymeric nanoparticles, the hybrid nanoparticle in this study had a much higher drug loading rate, which might be related to the high affinity between ZOL and nHA.

As shown in the schematic of the HA-PEG-nHA-ZOL nanoparticle, the hybrid nanoparticle was constructed by a nHA core recruiting hydrophilic ZOL molecules, wrapping with a PEG-HA polymer shell (Figure 1A). Transmission electron microscopy (TEM) images of blank and ZOL-loaded nanoparticles exhibited a rod-like geometrical profile, with about 150 nm length and 40 nm cross-sectional diameter (Figure 1D). Diameters of blank and ZOL-loaded nanoparticles were precisely measured by dynamic light scattering (DLS). As shown in Table 1 and Figures 1D,E, the average diameter of HA-PEG-nHA NPs was 159.0 ± 2.3 nm. When loaded with ZOL, the average diameter was 182.5 ± 1.7 nm. The structure and stability of NPs were analyzed according to the zeta potential. As shown in Table 4 and Figure 1E, the zeta potential of HA-PEG-nHA was −6.01 ± 0.23 mV while that of HA-PEG-nHA-ZOL was −7.03 ± 0.19 mV. The electronegativity of ZOL reduced the surface potential of nanoparticles after drug loading, which efficiently stabilized the structure.

Ultraviolet spectrophotometry was used to measure the drug release rate in vitro. As shown in Figure 1F, dissociative ZOL displayed a rapid release rate with 91.43% liberated drug within 12 h. Relatively, HA-PEG-nHA-ZOL NPs exhibited a slower and constant drug release rate. At pH of 5.4 and 7.4, HA-PEG-nHA-ZOL NPs were released 30.08 and 20.54% ZOL, respectively, within 12 h, 64.11%, and 49.62% ZOL, respectively, within 72 h. The results above indicated that controlled and sustained drug release could be realized using HA-PEG-nHA-ZOL NPs, maintaining the long-lasting therapeutic effect with a single injection. Moreover, the higher drug release rate in an acidic environment indicates that HA-PEG-nHA-ZOL NPs could be an acid-sensitive carrier to promote the drug release in acidic solid tumor sites or ischemic regions.

Cell live/dead dyeing and cell counting kit-8 (CCK-8) assay were performed to evaluate the cytotoxicity of NPs to 293T cells. A total of 293T cells were treated with gradient concentration of NPs and ZOL (details in Table 2), and cell viability was measured after 48 h. To achieve equivalent drug dose control, the NPs and ZOL concentrations were converted based on the 40% drug loading rate. With low-dose NP (<2.5 μg/ml) treatment, 293T cells showed high viability. When treated with a low dose of free ZOL (the same dose was loaded in NPs, < 1 μg/ml), 293T cells showed lower viability and poor cell condition (Figure 2A). However, under high-dose NPs (>5 μg/ml) and ZOL (>2 μg/ml) treatment, 293T cells exhibited low cell viability (Figures 2B, 3A–F). We propose that even with sustained drug release, high-dose NPs had toxic effect due to high ZOL concentration in culture medium after 48 h of drug release.

To examine the cytotoxicity of the HA-PEG-nHA NP itself, we performed both cell live/dead dyeing and CCK-8 assay 48 h after treating 293T cells with gradient concentrations of blank NPs (NPB) and nHA. As shown in Figure 2C and Figures 3A–F, at all concentrations of NPB and nHA, 293T cells showed high cell viability. CCK-8 assay showed similar results as live/dead dyeing. Notably, under high nHA treatment, the cell viability showed a slight reduction, while the cytotoxic effect was weaken by PEG-nHA coating, indicating that the hybridization effectively reduced the cytotoxicity non-tumorigenic cells like 293T cells. Taken together, the results above suggested the good biocompatibility and low cytotoxicity of our hybrid nanoparticles in vitro.

H&E staining was performed to evaluate the in vivo toxicity of NPs at different concentrations. High-dose blank NPs (NPB, 1000 μg/ml) and PBS were used as control groups. Figure 4 shows the sections of the brain, heart, liver, lung, and kidney 3 days after drug treatment. When treated with NPB or PBS, all corresponding figures showed no tissue toxicity (Figure 4). For drug-loaded NP groups, when treated with low-dose NPs (<250 μg/ml), no obvious tissue toxicity was observed in all the sections. The brain, heart, liver, and lung did not show evident tissue toxicity under high-dose NP (>500 μg/ml) treatment, either (Figure 4). However, distinct cell death and tissue damage were observed in the kidney and liver when mice were treated with high-dose NPs (>500 μg/ml) (Figure 4). Especially when treated with 1000 μg/ml drug-loaded NPs, swollen renal tubules and liver tissue were observed. These results suggested the low in vivo tissue toxicity of HA-PEG-nHA-ZOL NPs in a proper dose range while tissue toxicity still existed under high-dose treatment.

Liver function was evaluated 2, 3, and 5 days after NP intravascular delivery according to serum ALT, AST, and ALP levels. A total of 2 days after NP injection, groups treated with 250 μg/ml NPs showed insignificant serum ALT and AST level variation (Figure 5A). However, the AST level had a significant increase under the 500-μg/ml NP treatment, while a more distinct increase in ALT, AST, and ALP levels was observed under the 1000-μg/ml NP treatment (Figure 5A). A total of 3 and 5 days after NP injection, the ALP value returned to normal level, while ALT and AST values were maintained at high levels, suggesting the possibility of long-term permanent liver injury or acute inflammation after high-dose NP treatment (Figures 5B,C).

On the other hand, BUN and CR levels were measured 2, 3, and 5 days after NP intravascular delivery to assess kidney functions. A total of 2 days after NP delivery, all groups had no significant variation in BUN and CR levels (Figure 5A). However, 3 and 5 days after NP injection, groups treated with low-dose NPs (<250 μg/ml) had no evident variation in BUN and CR levels, while high-dose (>500 μg/ml) NP injection led to a significant increase in BUN and CR levels, suggesting severe kidney dysfunction (Figures 5B,C).

In general, an appropriate amount of NP administration had no impact on liver and kidney function. However, high-dose NPs resulted in liver and kidney malfunction, which might be due to NP or NPB accumulation in glomerulus or renal tubules, as well as the high hepatotoxicity of circulating ZOL.

Gradient concentrations of NPs were injected via intravenous route (details in Table 3), and immunohistochemical (IHC) staining was performed 3 days after injection in order to investigate the tissue immune cell infiltration in the kidney, lung, liver, and spleen. High-dose blank NPs (NPB, 1000 μg/ml) and PBS were used as control groups. As shown in Figure 6, even with high-dose NP (>500 μg/ml) treatment, no apparent inflammation was observed in the lung, liver, and spleen. For kidney, with the 500-μg/ml NP treatment, CD3 cell (T lymphocyte) and CD68 cell (monocyte/macrophage) infiltration rates had no obvious increase, while with the 1000-μg/ml NP treatment, a higher immune cell filtration rate was observed (Figures 6A,B), indicating the severe inflammation which might be due to accumulation of NPs or NPB in glomerulus and renal tubules as predicted before. However, no obvious inflammation was observed in the liver, suggesting the possibility of liver injury due to high toxicity of circulating zoledronic acid.

ZOL therapy showed a moderate immune response with low-dose treatment. To compare the immune response between classical ZOL therapy and hybrid NP therapy, 100 μl of 250 μg/ml NPs and 100 μg/ml ZOL were injected into mice in 2 groups (n = 4). Peripheral blood was collected, and flow cytometry analysis was performed 3 days post injection to measure the relative T cell and macrophage numbers. According to Figure 7A, CD68 + cell (monocytes/macrophages) number had no significant difference between the NP group and ZOL group. Meanwhile, CD3 + cell (T lymphocyte) and CD8 + cell (cytotoxic T lymphocyte) numbers showed no significant variation (Figure 7B). However, one mouse showed a strong immune response (two folds of T lymphocytes, eight folds of cytotoxic T lymphocytes) after NP injection (Figures 7B,C). The anomaly might be due to individual tolerance difference, which required further experiments with a larger sample size. In general, ZOL and NPs showed a similar and acceptable immune response after systemic delivery, which suggested that NPs could be exploited in vivo as a replacement of traditional ZOL therapy in the future.