2.1 Synthesis and Characterization of BOS@SOR
A bismuth-based nanomaterial (Na0.2Bi0.8O0.35F1.91, for short, NBOF) was first synthesized by a co-precipitation method. A mesoporous nanomaterial (Bi@BiO2 − x@Bi2S3) was then synthesized using a thermal reduction reaction by adding the sulfur source into NBOF aqueous solution. Subsequently, SOR was loaded to obtain Bi@BiO2 − x@Bi2S3–SOR, and its surface was modified with amphiphilic polyethylene glycol to obtain nanocomposite Bi@BiO2 − x@Bi2S3–SOR–PEG (BOS@SOR; Fig. 1A). The NBOF was spherical, approximately 220 nm (Figure S1). Using NBOF as a precursor, the Bi@BiO2 − x@Bi2S3 obtained by reduction also has a spherical shape and an average particle size of about 230 nm (Fig. 1B, C). From the elemental map, these nanoparticles contained three elements, Bi, O, and S, and the element distribution was relatively uniform (Fig. 1D). The electron diffraction peaks observed in the XRD results corresponded to the standard card of the three substances (PDF 44-1246 for Bi, 47-1057 for BiO2 − x, and 17–0320 for Bi2S3), indicating that the synthesized nanostructures are Bi@BiO2 − x@Bi2S3 composite structures (Fig. 1E). Furthermore, high-resolution TEM revealed that the lattice spacings were 0.328 nm, 0.274 nm, and 0.252 nm, respectively, which were well matched with Bi (012), BiO2 − x (200), and Bi2S3 (240) crystal plane (Fig. 1F). The XPS results showed that Bi and O characteristic peaks were observed in the full spectrum (Fig. 1G). The position of the S peak was further determined by the enlarged XPS spectrum (Fig. 1H), suggesting that Bi contains 0 and + 3 valences, and O and S are both − 2 valences (Figure S2). The XPS results showed that the elemental composition in the heterojunction was in line with expectations. As demonstrated in Fig. 1I, after SOR was loaded and surface modified, the characteristic absorption peak of SOR in the UV–vis spectrum confirmed that SOR was successfully loaded. Similarly, in the DLS results, the hydrated particle size increased continuously with SOR loading and surface modification using amphiphilic PEG (Fig. 1J).
2.2 Loading capacity and encapsulation efficiency of BOS@SOR
To test the different mass ratios of SOR to Bi@BiO2 − x@Bi2S3, the loading capacity (LC) and encapsulation efficiency (EE) values of SOR were evaluated. When the feed ratio was increased from 0.5 to 1.3, the drug loading and encapsulation rate of SOR increased (Fig. 2A). When the feeding ratio was 0.5, the drug loading content was too low to be ignored. The LC was 7.5% and 10.3% at feed ratios of 0.8 and 1.0, respectively. When the feed ratio was increased to 1.3, both LC and EE values were significantly improved. Based on the results of the subsequent cytotoxicity experiments, the loading ratio of 0.8 was chosen for SOR.
Glutathione (GSH), a reducing substance significantly upregulating in the tumor cells, protects cells from oxidative injury caused by reactive oxygen species (ROS). The strategy of consuming GSH to disrupt the redox balance in the tumor cells and to enhance the antitumor efficacy is increasingly recognized (24, 25). Bi-based nanomaterials, such as Pt/Bi, BiOCl/Cu2+, and Mn/Bi nanocomposites, have an excellent ability to regulate redox levels based on the characteristics of Bi in coordination with GSH (16, 19, 26–28). In the case of the coexistence of pH 5.5 (mimics the weakly acidic tumor microenvironment) and GSH, the drug release rate was much higher than that of pure pH 7.4, and the condition of the coexistence of GSH, and the release rate was close to 100% after 72 hours of release. Under the conditions of pH 7.4 alone and pH 7.4 plus GSH, the 72-hour release rates were 26.8% and 43.5%, respectively, indicating that both pH and GSH affect SOR release (Fig. 2B). Under the condition of pH 5.5 plus GSH, the morphology of BOS@SOR collapsed, and the particle size continued to decrease with time, effectively indicating that the composite could be degraded, which provides a basis for improving SOR bioavailability and realizing sustained drug release (Fig. 2C). The absorption spectrum showed that the absorbance at 348 nm was observed to rise continuously with time, indicating that the Bi-GSH complex formed after the Bi3+ was always coordinated with GSH and consumed the free GSH (Fig. 2D).
2.3 Photothermal effect of BOS@SOR
Based on the above analysis of finding the optimal mass ratio of SOR and Bi@BiO2 − x@Bi2S3, the photothermal properties of the as-prepared BOS@SOR were then evaluated. Under the irradiation of 808 nm laser (1 W cm− 2), the temperature of BOS@SOR solution (100 µg mL− 1) increased with the irradiation time. When the temperature rose by 26.5°C after eight minutes of irradiation, the irradiation was stopped, and the solution temperature was gradually cooled to room temperature. At the same time, taking water as a reference, under the same laser irradiation conditions, the temperature rise of water was less than 1°C (Fig. 3A). The calculation showed that the photothermal conversion efficiency (η) of BOS@SOR under 808 nm irradiation was 42.3% (Fig. 3B). To study the stability of the material, we repeatedly irradiated the solution of BOS@SOR with the laser on and off four times, and its heating ability was repeatable. No apparent heating attenuation after repeated irradiation was observed (Fig. 3C). Under the same irradiation conditions, the heating capacity of BOS@SOR was positively correlated with the concentration; that is, the higher the concentration, the higher the temperature (Fig. 3D, E). In addition, the heating capacity of BOS@SOR was also positively correlated with the power density of the 808 nm laser, and the temperature increased with the increase of the power density (Fig. 3F). These results indicate that BOS@SOR can provide sufficient heat to realize thermocatalytic effect.
2.4 In Vitro Cytotoxicity, Chemotherapy, and photothermal therapy
Excellent biocompatibility is a prerequisite for nanoparticles as nano-drug-delivery vehicles. Herein, two HCC cell lines (HCC-LM3 and MHCC-97H) established in our institute were applied to evaluate the cytotoxicity of BOS by CCK-8 assay. After incubating with BOS at various concentrations from 0 to 200 µg mL− 1 for 24 h or 48 h, both HCC-LM3 and MHCC-97H cells exhibited negligible cytotoxicity. The cell viability was over 95%, even at 200 µg mL− 1 of BOS (Fig. 4A). The favorable evidence demonstrated the excellent biocompatibility of BOS, providing a solid foundation for further biological application in vitro and in vivo.
To select the appropriate concentration for verification of synergetic targeted and photothermal therapy, the single effect of targeted or photothermal therapy was detected at various BOS concentrations by the CCK-8 assay. The cell viability rate decreased gradually with an increase in BOS@SOR (from 0 to 200 µg mL − 1). However, the cells subjected to the same doses of free SOR (from 0 to 15 µg mL− 1) exhibited significantly lower cell viability in HCC-LM3 and MHCC-97H cells (Fig. 4B), which may be partially attributed to insufficient uptake of BOS@SOR and sustained release property. Next, the cells were incubated with BOS at various concentrations and irradiated with 808 nm light at 800 mW cm− 2 for 5 min. As shown in Fig. 4C, BOS or laser alone has negligible cytotoxicity. While combined with increasing concentration of BOS (from 0 to 200 µg mL− 1), the same dose of laser exhibited an increasingly inhibited effect on cell proliferation. Remarkably, 50 µg mL− 1 BOS could significantly inhibit 30–40% cell proliferation by monotherapy, thus leaving enough space for subsequent verification of the synergistic therapeutic effect. Therefore, the 50 µg mL− 1 BOS@SOR (BOS: 46.25 µg mL− 1, SOR: 3.75 µg mL− 1) was selected for further experiments.
2.5 In Vitro Synergistic Therapy
To evaluate the synergistic therapy of BOS@SOR against HCC cells, a series of functional assays were carried out. Cells were assigned to seven treatment groups and subjected to complementary treatment, including (Ⅰ) PBS, (Ⅱ) BOS (46.25 µg mL− 1), (Ⅲ) SOR (3.75 µg mL− 1), (Ⅳ) BOS@SOR (50 µg mL− 1), (Ⅴ) laser (800 mW cm− 2 for 5 min), (Ⅵ) BOS plus laser (46.25 µg mL− 1, 800 mW cm− 2 for 5 min), and (Ⅶ) BOS@SOR plus laser (50 µg mL− 1, 800 mW cm− 2 for 5 min). Strikingly, cells treated with BOS@SOR plus laser (23.89 ± 5.54% for HCC-LM3 and 24.53 ± 5.40% for MHCC-97H) exhibited lower cell viability as compared with those of SOR (55.47 ± 4.69% for HCC-LM3 and 56.46 ± 2.91% for MHCC-97H), BOS@SOR (65.42 ± 2.27% for HCC-LM3 and 64.97 ± 5.99% for MHCC-97H) or BOS plus laser (67.15 ± 2.98% for HCC-LM3 and 61.63 ± 5.08% for MHCC-97H; Fig. 5A), respectively. Compared with the cells treated with PBS, no significant change in viability was observed in cells subjected to BOS or laser alone. To assess whether the composite would indiscriminately kill malignant and nonmalignant cells, a CCK-8 assay using normal hepatocytes (L02) was performed. Compared to the PBS group, cells treated with BOS@SOR plus laser (78.43 ± 1.40%), SOR (72.07 ± 1.66%), or BOS plus laser (84.80 ± 1.96%) presented decreased viability but significantly higher than those of HCC cells (Fig. 5B). Meanwhile, no significant viability changes were observed between the BOS@SOR (95.53 ± 2.03%) and PBS groups (P = 0.100). Consistent with our data in Fig. 2B and C, the encouraging results may be partially attributed to the inadequate GSH content in normal cells, which prevents the nanoparticles from degrading and the encapsulated SOR from being released.
Subsequently, to assess long-term growth, a colony formation assay demonstrated that synergistic photothermal and molecularly-targeted therapy of BOS@SOR (21.69 ± 3.19% for HCC-LM3 and 21.86 ± 4.21% for MHCC-97H) had a lower colony forming ability of hepatoma cells than that of SOR (45.78 ± 3.19% for HCC-LM3 and 46.96 ± 3.06% for MHCC-97H), BOS@SOR (62.25 ± 3.87% for HCC-LM3 and 62.35 ± 5.06% for MHCC-97H) or BOS plus laser (57.83 ± 2.41% for HCC-LM3 and 59.11 ± 3.71% for MHCC-97H; Fig. 5C), respectively. Still, the synergistic therapeutic efficiency of BOS@SOR was identified in the Live/Dead cells staining assay with Calcein-AM and PI reagent, where Calcein-AM fluoresces intense green in viable cells and PI fluoresces bright red in dead cells (23, 29). Consistently, the quantification of green fluorescence revealed that cells subjected to synergistic therapy had the lowest proportion of live cells compared with those in the other six groups (Fig. 5D).
Finally, cell apoptosis among each group was detected by flow cytometry. BOS@SOR plus laser group (53.63 ± 5.82%) exhibited enhanced apoptosis as compared with that in the SOR (17.48 ± 0.90%), BOS@SOR (10.74 ± 0.99%) or BOS@SOR plus laser group (15.47 ± 0.90%), respectively. However, these apoptosis rates were not wholly consistent with the results from previous CCK-8, colony formation, and live/dead cell staining assays, and Group Ⅶ, Ⅲ, Ⅳ, or Ⅵ should have shown higher apoptosis rates (Fig. 5E). This paradox could be partially attributed to the fact that BOS- or SOR-mediated cell death may not be limited to apoptosis; other forms of cell death, such as ferroptosis, may also be involved (30). The apoptosis markers were investigated to further investigate the association between apoptosis and BOS@SOR. Herein, the Bcl-2 family (including Bcl-2, Bcl-xl, and Mcl-1), playing a suppressive role in tumor cell apoptosis (31, 32), decreased significantly in BOS@SOR plus laser group as compared with those in other groups using Western blotting (Fig. 5F). Taken together, the above results demonstrate that synergistic photothermal and molecularly-targeted therapy of BOS@SOR markedly improved therapeutic efficiency than monotherapy and at least in part due to the increased cell apoptosis.
2.6 In Vivo Biosafety
Biocompatibility is a prerequisite for in vivo biomedical applications of nanocapsules (33, 34). Given that hemocompatibility is a vital indicator to assess biocompatibility in vivo, the hemolytic test was initially performed. No significant hemolysis was observed when red blood cells were incubated with 0.9% saline, or BOS@SOR at concentrations from 100 to 500 µg mL− 1, while significant hemolysis occurred in distilled water (Fig. 6A). Correspondingly, quantitative analysis of hemoglobin release by spectrophotometer showed that the hemolysis rates of the BOS@SOR at varying concentrations, even up to 500 µg mL− 1, were all less than 2% (Fig. 6B), which is below the safe hemolytic ratio for biomaterials according to ISO/TR7406 (35). Subsequently, BALB/c nude mice were divided into two groups and intravenously injected with saline (200 µL) or BOS@SOR (200 µL, 1 mg mL− 1). On day 14 after injection, the mice were sacrificed, and blood samples were collected for biochemical examination (n = 6 per group). As expected, no marked change was monitored in serological findings (including AST, ALT, and ALB for the liver, BUN, and CRE for the kidney, and LDH for the heart) between the two groups (Fig. 6C). Since pathological assessment is the gold standard for evaluating biomaterial toxicity, HE staining on mice's vital organs (including heart, liver, spleen, lung, and kidney) was further performed after injection. No pronounced morphological and inflammatory changes were observed in all five organs (Fig. 6D). These results identify that BOS@SOR presents excellent biocompatibility, making it possible for in vivo applications.
2.7 In Vivo CT Imaging, Biodistribution, and Photothermal Effect
Au and I were considered classical elements with high X-ray attenuation coefficients (Au, 5.16 cm2 kg− 1; I, 1.94 cm2 kg− 1), whereas Bi was found to possess a more significant coefficient (5.74 cm2 kg− 1 at 100 KeV) (36). Recently, some Bi-based nanomaterials, such as Bi2S3, Na0.2Bi0.8O0.35F1.91, and Bi-SR-PEG, have been reported to have contrast-enhanced CT imaging property (20, 22, 23). We wondered whether the Bi-based BOS@SOR could also be a CT imaging reagent. Here, BOS@SOR (200 µL, 1.0 mg mL− 1) was intravenously injected into the orthotopic xenograft tumor model in nude mice. The CT images of the orthotopic HCC mice were captured at different time points: pre-injection, 10 min, 20 min, 30 min, and 60 min post-injection. Compared with the CT image attained pre-injection, CT images acquired post-injection revealed that BOS@SOR presented a significant whitening effect in the tumor area, which remained stable during the observation period of one hour (Fig. 7A). The contrast-enhanced CT imaging indicated that BOS@SOR might qualify for a potential application in tumor imaging and supply a distinct structure of tumors for CT-guided synergistic therapy for tumors.
Identifying the distribution and enrichment levels of the biomaterial is of great significance for selecting the appropriate timing for treatment in vivo. Herein, BOS@SOR (200 µL, 1 mg mL− 1) was injected into HCC mice via the tail vein. After injection for 1, 3, 6, and 24 h, the major organs (including liver, spleen, lung, kidney, and heart) and tumor were collected for detecting the Bi concentration by ICP-OES. Bismuth was accumulated in all detected organs, with the highest concentration in the liver. Enrichment peaks were observed at 3 or 6 hours after injection in the organs, except for the kidney, which peaked at 12 hours post-injection, possibly because the kidney is the metabolic organ of the last resort (Fig. 7B). Notably, tumor tissue also reached its enrichment peak after 12 hours, which may be attributed to the permeability and retention effect. Moreover, the results suggested that the optimal antitumor efficacy of synergistic therapy could be realized at 12 h post-injection.
Moreover, the orthotopic xenograft HCC mice using MHCC-97H were injected with 0.9% saline (200 µL) or BOS@SOR (200 µL, 1 mg mL− 1) via the tail vein to validate the photothermal conversion performance. After 12 h, the mice were irradiated with an 808 nm laser at a power density of 800 mW cm− 2. The mouse injected with BOS@SOR caused an increase of 13°C within 5 min, while the mouse injected with saline showed a rise of 9°C (Fig. 7C). These results indicated that BOS@SOR enriched in the tumor site could yield a marked photothermal effect under light irradiation.
2.8 In Vivo Synergistic Therapy
Given the satisfactory therapeutic effect in vitro and excellent biocompatibility in vivo, the impact of synergistic photothermal and molecularly-targeted therapy was further assessed in vivo. Our preliminary data demonstrated that BOS@SOR was mainly accumulated in the liver, thus being an ideal therapeutic agent against liver diseases. As the orthotopic xenograft model mimicked the tumor microenvironment more efficiently than the subcutaneous model, an orthotopic HCC xenograft model was thus established using MHCC-97H cells. Followingly, the mice were further irradiated with 808 nm light at a power density of 800 mW cm− 2 for 5 min at 12 h post-injection of BOS@SOR (200 µL, 1.0 mg mL− 1; Fig. 8A).
On day 7, after the second treatment, the mice were euthanized and detected by bioluminescent imaging (Fig. 8B). The quantification analysis exhibited that luciferase activity in BOS@SOR plus laser group was significantly lower than in the other groups (n = 6 per group), suggesting a synergistic antitumor effect on HCC (Fig. 8C). Furthermore, compared with the mice receiving 0.9% saline, the mice subjected to free SOR alone suffered from a marked decrease in body weight, whereas the mice receiving the combination therapy of SOR-loaded nanoparticles and laser irradiation showed a less fluctuant pattern (Fig. 8D). The results reflected the lower systematic toxicity of the synergistic therapy. Subsequently, the tumors were resected, weighed, and photographed (Fig. 8E). Consistent with the results from bioluminescent imaging, the mice in the synergistic therapy group had lower tumor weight (0.07 ± 0.03 g) and volume (83.58 ± 36.15 mm3; Fig. 8F, G) as compared with those in BOS@SOR group (tumor weight 0.16 ± 0.03 g, and tumor volume 223.13 ± 63.16 mm3) and BOS plus laser group (tumor weight: 0.16 ± 0.03 g, and volume 229.07 ± 50.77 mm3). These results further confirmed the superior efficacy of synergistic photothermal and molecularly-targeted therapy over monotherapy. Of note, the in vivo results exhibited that the mice treated with BOS@SOR had a better therapeutic effect than those injected with free SOR (tumor weight and volume: 0.16 ± 0.03 g vs. 0.28 ± 0.04 g, 223.13 ± 63.16 mm3 vs. 315.57 ± 21.89 mm3, respectively), whereas better in vitro therapeutic efficacy was obtained in SOR group. The inconsistency could be attributed to the poor bioavailability of SOR and the passive targeting of BOS@SOR in vivo.
Next, histopathological staining of tumor tissues was carried out to assess the anti-tumor effect. H&E staining exhibited more severe tissue damages, such as necrosis and apoptotic body in BOS@SOR plus laser group, compared with the groups receiving SOR, BOS@SOR, or BOS plus laser alone. Also, no apparent morphology damage was observed in the other groups (Fig. 8H). Moreover, TUNEL staining exhibited that tissue sections from the BOS@SOR plus laser group showed a higher proportion of apoptotic cells (green fluorescence) than those from SOR, BOS@SOR, or BOS plus laser group. In contrast, the apoptotic cells from the saline, BOS, or laser group were negligible. These results indicated that the synergistic therapeutic efficacy could be partially ascribed to apoptosis activation (Fig. 8H). Apart from activating apoptosis, other properties involved in combination therapy, such as the anti-proliferation and anti-angiogenesis activities, were also determined. The results suggested that the staining intensity of Ki-67 (a well-known biomarker indicating cell proliferation (37, 38)), as well as VEGFA and CD31 (both are the classical angiogenesis markers (39, 40)) in each group, exhibited a similar tendency as compared with the corresponding tumor weight or volume (Fig. 8I). These results suggest that pro-apoptosis, anti-proliferation, and anti-angiogenesis effects are involved in the synergistic therapeutic efficacy of BOS@SOR, although the detailed mechanisms need further investigation.