Preparation of self-assembly silica redox nanoparticles to improve drug encapsulation and suppress the adverse effect of doxorubicin

Chemicals

Dimethylformamide (DMF, Sigma-Aldrich, USA), permeable membrane tube (MWCO 3.5 kDa, Spectrum Laboratories Inc., Japan), tetraethyl orthosilicate (TEOS, Sigma-Aldrich, USA), ammonia (NH₃, China), Doxorubicin hydrochloride (DOX, Wako, Japan), 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT, Roche Diagnostics, Japan), dimethyl sulfoxide (DMSO, China), Dulbecco’s Modified Eagle’s Medium (DMEM, Gibco, USA), fetal bovine serum (FBS, Sigma-Aldrich, USA), and antibiotics (a mixture of penicillin, streptomycin, and neomycin, Sigma-Aldrich, USA) were purchased.

Preparation and characterization of RNP^N and siRNP

RNP^N was prepared by self-assembly amphiphilic copolymers PEG-b-PMNT, as reported in previous research [18]. Briefly, 15 mg of PEG-b-PMNT were dissolved in 1 mL of DMF and then stirred on a magnetic stirrer. The mixture was then put into a semi-permeable membrane tube before being dialyzed against distilled water for 24 h. To prepare siRNP, a similar method was conducted, except that 50 μL of TEOS and 50 μL of NH₃ were added during the stirring process (Figure 1). The average size, polydispersity index (PdI), and size distribution of RNP^N and siRNP were then characterized by the dynamic light scattering (DLS, Malvern Zetasizer, UK) system.

The stability of RNP^N and siRNP

Phosphate-buffered saline (PBS, 10 mM) was prepared at different pHs, including 3.0, 6.5, and 7.4, to mimic the pH levels of the gastric, tumour extracellular, and physiological environments, respectively. RNP^N and siRNP were then diluted in the prepared buffers. The stability of RNP^N and siRNP was evaluated by measuring the change in light scattering intensity using the DLS system at 25 °C for 2 h.

Preparation and characterization of DOX@RNP^N and DOX@siRNP

DOX@RNP^N and DOX@siRNP were prepared by co-dissolving 1 mg of DOX and 15 mg of PEG-b-PMNT in 1 mL of DMF. Similar steps were followed in the preparation of RNP^N and siRNP, respectively (Figure 1). Additionally, various ratios of TEOS:NH₃ (200 μL TEOS: 200 μL NH₃, 100 μL TEOS: 100 μL NH₃, 50 μL TEOS: 50 μL NH₃, 100 μL TEOS: 50 μL NH₃, 200 μL TEOS: 50 μL NH₃) were examined to optimize the size of DOX@siRNP. The DLS system was used to characterize the average size, PdI, and size distribution of DOX@RNP^N and DOX@siRNP.

Drug encapsulated efficiency and loading capacity of DOX@RNP^N and DOX@siRNP

The encapsulation efficiency (EE / %) and loading capacity (LC / %) of DOX@RNP^N and DOX@siRNP were investigated based on the amount of loaded DOX evaluated by a microplate reader (Thermo Fisher Scientific, USA) with a calibration curve of DOX measured absorbance at 480 nm. The EE and LC were calculated by usingequation (1) and(2):

(1)

(2)

Release profile of DOX@RNP^N and DOX@siRNP

2 mL of DOX@RNP^N or DOX@siRNP were loaded into separate dialysis bags and placed in different beakers under stirring conditions, each with distinct pH including 3, 6.5, and 7.4. At predetermined time points (0, 1, 3, 6, 9, 20, and 24 h), samples were collected and the absorbance at 480 nm was measured to evaluate the amount of drug released from the NPs.

Cytotoxic assay

The murine fibroblast cells (L929), human breast cancer cells (MCF-7), and human hepatocellular carcinoma cells (HepG2) were obtained from the American Type Culture Collection (ATCC, USA). The cells were cultured in DMEM containing 10 % FBS and 1 % antibiotics (a mixture of penicillin, streptomycin, and neomycin) at 37 °C and 5 % CO₂. The cells were then seeded on 96-well plates (10⁴ cells/well) and incubated for 24 h at 37 °C and 5 % CO₂. The tested samples, including RNP^N, siRNP, DOX, DOX@RNP^N, and DOX@siRNP, were added to each well to obtain final concentrations of DOX at 0.25, 0.5, 1, 2.5, and 5 μg/mL, and incubated for 24 h. After that, the medium was removed, and MTT solution was added to each well and continuously incubated for 4 h. Finally, DMSO was added to dissolve crystals of formazan inside the cells, and the plates were measured absorbance at 540 nm by a microplate reader (Thermo Fisher Scientific, USA). The cell viability (%) was calculated by usingequation (3):

(3)

where OD_sample is the absorbance of the sample and OD_control is the absorbance of the solvent.

Statistical analysis

Statistical analysis was performed using one-way analysis of variance (ANOVA) to determine significant differences among various groups. A p-value <0.05 was considered statistically significant. Data were presented as the mean ± standard deviation (SD).

Characterization of RNP^N and siRNP

RNP^N and siRNP were prepared by self-assembly amphiphilic copolymers PEG-b-PMNT in DMF, followed by dialysis against distilled water for 24 h to remove the organic solvent. The results showed that both RNP^N and siRNP were transparent (Figure 2A), with average sizes of 45.5 and 85.5 nm, respectively (Figure 2B). Furthermore, the PdI of both NPs was less than 0.3, indicating a narrow size distribution. The increase in the size of siRNP as compared to RNP^N might be explained by a sol-gel reaction of TEOS in the RNP^N core [22]. The size of siRNP was approximately 85 nm after the condensation reaction, suggesting no significant aggregation (Figure 2C), even though silica particles formed by sol-gel can reach up to 2000 nm. We next examined the influence of different pH buffers on the stability of NPs [23]. As shown inFigure 2D, RNP^N disintegrated under acidic pH (3.0 and 6.5) and stabilized at pH 7.4. In contrast, siRNP was stable regardless of pH change. This might be explained that the amino groups of the PMNT segment disintegrated due to their protonation at low pH. However, adding silica would maintain the structure of NPs at acidic pH, resulting in enhanced stability.

Characterization of DOX@RNP^N and DOX@siRNP

DOX@RNP^N and DOX@siRNP were similarly prepared as RNP^N and siRNP, respectively, except for the addition of DOX during the preparation process. Different concentrations of NH₃ might result in a change in the size of silica NPs despite its catalyzation role in sol-gel reactions [24]. Thus, different ratios of TEOS:NH₃ were examined to prepare DOX@siRNP with the nano-size distribution. As shown inFigure 3A, DOX@RNP^N, DOX@siRNP with ratios of 50 μL TEOS: 50 μL NH₃ and 100 μL TEOS: 50 μL NH₃ were transparent, resulting in the nano-size distribution as compared to other samples with turbidity in solution and micro-size distribution (data not shown). Hence, DOX@siRNP with a ratio of 50 μL TEOS: 50 μL NH₃ was chosen for further experiments because of its suitable size. As compared to NPs, the size of both DOX@RNP^N and DOX@siRNP significantly increased to 56 and 90 nm, respectively, implying the encapsulation of DOX in NPs (Figure 3B). The critical cut-off point for endocytosis has been estimated from 20 to 500 nm for M cells and from approximately 50 to around 100 nm for enterocyte cells [25]. Therefore, DOX@RNP^N and DOX@siRNP showed the potential to increase the oral bioavailability of DOX. Moreover, the PdI of both NPs was around 0.3, indicating the narrow size distribution even after drug encapsulation (Figure 3C). Conventionally, polymeric NPs exhibit a low EE by physical entrapment of the drug (less than 10 %) [26-28]. For example, the LC of DOX-loaded silica NPs was reported at 4.8 % [29]. As shown inFigure 3D, the EE was 20.3 and 74.9 %, and the LC was 1.5 and 5 % for DOX@RNP^N and DOX@siRNP, respectively. This result indicates the appropriate design of the DOX encapsulation because the silanol groups in the core of siRNP significantly improved the EE and LC of DOX. This result could be explained by the pKa of the amino group in DOX being around 9.9, thus most DOX molecules are in the positively charged form at physiological pH [30]. Consequently, silanol moieties possessing a negatively charged surface could interact with DOX by electrostatics, suggesting the enhancement in EE and LC of DOX@siRNP.

Release profile of DOX@RNP^N and DOX@siRNP

The release kinetics of DOX@RNP^N and DOX@siRNP were evaluated under different pH buffers, including pH 3.0, 6.5, and 7.4, which mimic the gastric, tumour extracellular, and physiological pH, respectively. Gastric pH is one of the challenging barriers for oral drug delivery systems. As shown inFigure 4A, DOX@siRNP was released significantly slower than DOX@RNP^N (17.8 % compared to 35.9 %) after 24 h, even though both experienced protonations of amino groups at the hydrophobic core. However, more than 80 % of DOX@RNP^N and 90 % of DOX@siRNP were retained after 6 h, suggesting their potential for in vivo application as they can withstand the maximal transit time of the stomach in the human body. The extracellular pH of tumors, which commonly drops around 6.5-7.0, is mainly caused by anaerobic glycolysis in hypoxia [31]. As shown inFigure 4B, the release profiles of DOX@siRNP and DOX@RNP^N at pH 6.5 were 16.4 and 33.0 %, respectively. These results are consistent with the previous stability study. It is evident that DOX@siRNP was less impacted by acidic pH than DOX@RNP^N due to the ability of nano-silica to stabilize the structure. After 24 h at physiological pH, DOX@RNP^N and DOX@siRNP leaked at 14 and 15 %, respectively (Figure 4C). Similarly, several studies have reported that the release profile of micelle NPs with pH-sensitive groups is higher at acidic pH than at neutral pH [32,33].

In vitro cytotoxicity

The cytotoxicity of NPs was assessed through MTT assay on normal murine fibroblast cells (L929) and cancer cells, including human breast cancer cells (MCF-7) and human hepatocellular carcinoma cells (HepG2). In addition to the well-known anticancer mechanism, which involves interaction with the topoisomerases II to interfere with DNA replication, it also induces the over-expression of free radicals to kill cells. Therefore, RNP^N with TEMPOL, a ROS scavenger in the hydrophobic core, was used to encapsulate DOX to reduce its side effects in an in vitro model. As shown inFigure 5A, DOX indicated dose-dependence, with approximately 50 % of L929 cell death at a concentration of 5 μg/mL DOX, while the NPs significantly enhanced cell viability, almost equal to the control sample. Despite dose-dependent toxicity observed in DOX@RNP^N and DOX@siRNP, there was more than 90 % of cell viability after 24 h of incubation due to the scavenging ability of TEMPO groups. In addition, silica NPs have been reported to cause cytotoxicity due to the overproduction of ROS [34], leading to lower cell viability of siRNP compared to RNP^N. However, siRNP treatment exhibited more than 90 % of cell viability, suggesting the biosafe profile of NPs. Similarly, anticancer effects of DOX were observed on human breast cancer cells (MCF-7) and human hepatocellular carcinoma cells (HepG2), with dose dependence (Figure 5B andFigure 5C). Remarkably, DOX@siRNP caused fewer cell deaths (around 20 and 25 %) compared to (27 and 32 %) at 5 μg/mL of free DOX treatment against MCF-7 and HepG2, respectively. This result may be due to DOX being kept inside the core of siRNP, resulting in less exposure to cells, although encapsulation of DOX in NPs was expected to be more effective at entering cells. In contrast, the anticancer effect caused by DOX@RNP^N (5 μg/mL of DOX) was 36 and 37 % for MCF-7 and HepG2, respectively. It is obvious that the cumulative release of DOX@RNP^N was faster as compared to DOX@siRNP, resulting in higher cytotoxicity when entering cells.