Safety of Nonporous Silica Nanoparticles in Human Corneal Endothelial Cells

Nonporous silica nanoparticles (SiNPs) are promising drug carrier platforms for intraocular drug delivery. In this study, we investigated the safety of three different sizes of SiNPs (50, 100, and 150 nm) in a human corneal endothelial cell (HCEC) line, B4G12. The HCECs were exposed to different concentrations (0, 25, 50, and 100 µg/ml) of three sizes of SiNPs for up to 48 h. Cellular viability, autophagy, lactate dehydrogenase (LDH) assay, and mammalian target of rapamycin (mTOR) pathway activation were evaluated. Intracellular distribution of the SiNPs was evaluated with transmission electron microscopy (TEM). TEM revealed that the SiNPs were up-taken by the HCECs inside cytoplasmic vacuoles. No mitochondrial structural damage was observed. Both cellular viability and LDH level remained unchanged with up to 100 µg/mL of SiNP treatment. Autophagy showed a significant dose-dependent activation with 50, 100, and 150 nm SiNPs. However, the mTOR activation remained unchanged. Human corneal tissue culture with 100 µg/ml concentrations of SiNPs for 72 h revealed no significant endothelial toxicity. In vivo corneal safety of the SiNPs (0.05 ml intracameral injection, 200 mg/ml concentration) was also verified in rabbit models. These findings suggested that 50, 100, and 150 nm SiNPs did not induce acute significant cytotoxicity in corneal endothelial cells at concentrations up to 100 µg/mL. However, long-term toxicity of SiNPs remains unknown.

mTOR Pathway Activation. We measured the expression level of phosphorylated mTOR (p-mTOR) and mTOR (Fig. 4). The increase in phosphorylated mTOR indicated the activation of the pathway. The expression of p-mTOR showed no significant change with the SiNP addition when compared with the normal control. Our finding suggests that 50, 100, and 150 nm SiNPs do not inhibit mTOR signal transduction, which is one of the most important cell survival pathways.
Ex Vivo HCEC Toxicity Assay. We used a low concentration of trypan blue (0.005%) mixed with minimum essential media (MEM) to minimize the endothelial cell toxicity during the staining procedure 13 . The trypan blue-stained area showed no significant difference compared with the baseline in all SiNP-treated corneas after 72 h of incubation (Fig. 6). After the SiNP treatment, the corneal endothelial cells effectively maintained a normal hexagonal actin skeleton, as revealed by the phalloidin staining (Fig. 6).
In Vivo Corneal Endothelial Cell Toxicity Assay. The rabbit corneas maintained their transparency for two weeks after the intracameral injection of SiNPs. No corneal haze, edema, or limbal vascular abnormality was observed both in the control and in the SiNP-treated eyes. The iris appeared normal, and the crystalline lens was clear both in the control and in the SiNP-treated eyes. The hexagonal structure of corneal endothelium was well maintained, and no significant difference was found in the endothelial cell density between the control and treated groups (Fig. 7). The histopathologic evaluation revealed no difference in the corneal endothelial appearance between SiNPs-treated and control corneas (Fig. 7).

Discussion
In this study, we found that monodisperse nonporous SiNPs of 50, 100, and 150 nm sizes induced no significant cytotoxicity in cultured HCECs up to 48 h and at a concentration of up to 100 µg/mL. Cellular autophagy showed a dose-dependent increase with high concentrations of SiNPs, but the mTOR pathway remained intact. The human corneal tissue culture model demonstrated no significant cytotoxicity of SiNPs (100 µg/mL) after 72 h of exposure. The safety of SiNPs was further verified in vivo with the rabbit intracameral injection model. For efficient intraocular drug delivery, nanoparticle-based ocular drug delivery systems, including gold, carbon, lipid, chitosan, liquid crystalline, and albumin nanoparticles, have been actively investigated [14][15][16][17][18][19] . As a nanocarrier system, SiNPs have a great potential for ocular drug delivery 20,21 . SiNPs have a stable chemical structure, a large surface area to volume ratio, ease of surface modification, and tolerable biodegradability, which all increase their attractiveness for biologic applications 3 . However, safety must be verified 22 , and as mentioned earlier, the safety of SiNPs has always been a challenge for clinical application. The crystalline form of silica is toxic to alveolar cells and can induce irreversible pulmonary silicosis 23,24 . Conversely, amorphous silica, such as nonporous silica particle, is considered relatively biologically safe because of its biodegradability [25][26][27] . Biodegradation and clearance are essential for biomedical use of nanoparticles because chronic accumulation of nanoparticles in human bodies can lead to serious long term complications. It was reported that 5 to 10 nm sized nonporous SiNPs degraded in two weeks 28 . Furthermore, smaller and porous SiNPs degraded faster than larger and nonporous SiNPs 28 . It was reported that mesoporous SiNPs biodegraded in human umbilical vein endothelial cells and human embro kidney (HEK) 293 T cells 29,30 . Nevertheless, negative reports have suggested the significant nanotoxicity caused by SiNPs [31][32][33][34][35] . The nanotoxicity of SiNPs seems to be dependent on size, dose, and cell types 9,36 . Therefore, the toxicity of silica nanoparticles should be evaluated independently in every target organ system by means of both in vivo and in vitro toxicity tests. For the ocular system, no direct toxicity of SiNPs in retinal and corneal cells has been previously reported 6,37 . In addition, our group recently reported the in vitro safety of 50, 100, and 150 nm-sized nonporous SiNPs by using the primary culture of human corneal epithelium and keratocytes 10,11 .  Although, the lack of acute toxicity of SiNPs on cultured human corneal epithelial cells and keratocytes was verified in our previous reports 10,11 , further verification of the safety of SiNPs in HCECs is madatory. Corneal epithelial cells and keratocytes possess regenerative capability and they can actively repopulate the damaged cells within several weeks after the injury. However, HCECs are lack of regenerative potential and any injury to HCECs can lead to the irreversible loss of corneal transparency and serious visual deterioration 2 . After penetrating the cornea, all nanoparticles disperse into aqueous humor. The aqueous outflow pathway is the most important clearance pathway for intraocular porous silicon micro-particles 12 . Until full clearance is obtained through the trabecular meshwork, SiNPs can remain in contact with HCECs for a prolonged time. Therefore, our current study regarding HCEC safety is an important step in the safety clearance for the future development of SiNPs as an ocular topical drug delivery system.
In this study, SiNPs treatment increased autophagy in HCECs. This finding is consistent with previous reports that SiNPs can induce cellular stress [38][39][40][41] . The similar elevation of autophagy by SiNPs was observed in our previous reports using human corneal epithelial cells and keratocytes 10,11 . Autophagy is known as a natural cellular process that cleans up unnecessary and dysfunctional cellular components for recycling 42 . It The mTOR signal activation did not significantly change with the treatment of SiNPs at a concentration of up to 100 μg/mL. Relative densitometry was calculated as a fold change from the control, and all values (mean ± SEM) were obtained from three independent experiments; each independent experiment was performed in triplicate (*p < 0.05, **p < 0.01, ***p < 0.001). also helps cells to overcome external stress and survive in harsh environments. Recently, it was reported that autophagy promotes the degradation of polyethyleneimine-alginate nanoparticles in endothelial progenitor cells 43 . Another study reported that the increased osteoblast proliferation was induced by autophagy after tantalum nanoparticle exposure 44 . Therefore we hypothesize autophagy system is one of the cellular mechanism to overcome SiNP-induced stress. It is noteworthy that mTOR pathway and cell viability remained intact even with increased autophagy in our study.
Unfortunately, the primary culture of HCECs has a limited proliferation capacity. Using primary cultured HCECs from different donors can cause donor-specific effects, such as age and topography-related (central vs. peripheral origin) effects on experimental results 2 . Therefore, because our studies required a great number of homogeneous cells, we used the HCECs cell line B4G12 in this study. This cell line is one of the two clonal lines obtained from the HCEC-12 cell line, which was made by the immortalization of parental cells from a 91-year-old woman (http://www.creative-bioarray.com/HCEC-B4G12-CSC-C3457-item-1468.htm). B4G12 can provide sufficient homogenous cell numbers for multiple experiments. As observed in many cell lines, the properties of HCECs and B4G12 are not always the same. Nevertheless, B4G12 adequately  expresses HCEC-specific 9.3. E-antigen, ZO-1, and occludin 45 . In addition, B4G12 transplanted to replace corneal endothelium in rabbit eyes was able to successfully restore the Na+/K+ pump function and maintain corneal transparency 46 . B4G12 has been actively used as an ideal model for differentiated HCECs because of its morphologic and functional similarity [47][48][49] .
To compensate for the limitation of using immortalized B4G12 cells, we further verified the safety of SiNPs by using ex vivo human corneas and an in vivo rabbit model. We found that 50, 100, and 150 nm-sized SiNPs induced no significant endothelial toxicity in both human and rabbit corneas. The well-maintained hexagonal architecture suggests that the SiNPs did not cause significant cellular damage at the tested concentration. Our study has several limitations. First, we tested limited sizes and concentrations of SiNPs (up to 100 µg/mL and three different sizes). A further increase in SiNP concentration to a higher range or with smaller particle sizes may possibly induce significant cellular damage in HCECs. Moreover, the uptake of SiNPs by cells may eventually lead to the chronic perturbation of intracellular mechanisms and cellular damage. Therefore, a close in vivo observation for an extended period of time (e.g., years) is necessary to further verify the safety of SiNPs. Second, the long-term exposure effects of SiNPs on HCECs are lacking. Note that ex vivo human corneal tissue culture is different from the in vivo environment because the dynamic circulation of aqueous humor is absent in a tissue culture setting.
In summary, we found SiNPs of sizes 50, 100, and 150 nm to be safe in HCECs at up to 72 h of exposure. The cellular uptake of SiNPs was localized inside the vacuoles of the cytoplasm with no nuclear membrane or mitochondrial damage. The cellular survival in the mTOR pathway remained intact, although some activation of autophagy was observed at higher concentrations of SiNPs. The cellular viability of HCECs was not affected at up to 48 h of exposure to SiNPs at the tested concentrations. The human corneal tissue culture revealed no significant toxicity of SiNPs at 100 μg/mL of concentration at up to 72 h. The in vivo rabbit model further verified the lack of toxicity of SiNPs in corneal endothelial cells.

Conclusions
Our current study verified the safety of 0, 100, 150 nm sized SiNPs in human HCECs. These findings can facilitate the future development of the SiNP-based ophthalmic drug delivery method for intractable eye diseases.

Materials and Methods
Synthesis and Characterization of Nanoparticles. Three sizes of silica (SiO 2 ) nanoparticles (SiNPs; sizes: 50, 100, and 150 nm) were manufactured using the Stöber synthesis method as previously reported 10,11 . The synthesis of 50 nm of SiNPs started with mixing 2 mL of ammonia (NH 4 OH, 28%, Junsei, Tokyo, Japan) and 50 mL of ethyl alcohol (EtOH, anhydrous, 99.5%, Daejung, Kyeonggi, Korea) were first mixed. Then, 1 mL of tetraethyl orthosilicate (TEOS, Samchun, Gyunggi, Korea) was added. Similarly, 100 nm and 150 nm of SiNPs were prepared using equal molar ratios of ingredients. Afterward, 1.5 mL of TEOS was added to the mixture of 3 mL solution of ammonia in 50 mL of ethyl alcohol. Then, the solutions were stirred continuously for 12 h at an ambient condition (25 °C, 1 atm). Smaller-sized SiNPs could be produced by quickly adding TEOS while stirring the solution. The prepared SiNPs were washed with EtOH three times using centrifugation (10,000 rpm, 15 min). The final SiNP precipitates were dispersed in distilled water.
The surface charge of the prepared SiNPs was measured by the zeta potential (SZ-100, Horiba, Kyoto, Japan) in both distilled water and DPBS. In addition, SEM (SIGMA, Carl Zeiss, Oberkochen, Germany) images and ImageJ software were used to analyze the size and distribution of SiNPs. The dispersity of SiNPs was defined as the coefficient of variation (Dispersity (%) = d 100 σ ×  , where σ is the standard deviation and d is the mean size) 50 .
Cell Culture. The limited proliferative capability of cultured primary HCECs makes any in vitro experiment requiring large numbers of these cells unrealistic. Therefore, we used an established HCEC line (i.e., B4G12 cells) in this study. B4G12 cells (Cat no.CSC-C3457-CRA, Creative Bioarray, Shirley, NY, USA) were used. The cells were cultured in tissue culture-treated plastic at 4 × 10 4 cells/cm 2 . The culture medium recommended by the company, which contains human endothelial serum free medium (Cat no. CM-345L7, Creative Bioarray, Shirley, NY, USA) and 10 ng/ml of fibroblast growth factor-2 (Cat no. CSC-CTK0134, Creative Bioarray, Shirley, NY, USA), was used. After reaching confluency, the cells were harvested and resuspended in the culture medium. The cells were plated in 75 cm 2 tissue flasks and then maintained at 37 °C in a 5% CO 2 and 95% air-humidified atmosphere. The culture medium was changed every three days, and the cells were passaged using 0.25% Trypsin-ethylenediaminetetraacetic acid (EDTA) (Gibco BRL, Carlsbad, CA, USA). The passage number ≤5 was used for the study.

Treatment of SiNPs.
SiNPs with sizes of 50, 100, and 150 nm were confirmed using SEM. Before being mixed into the HCECs culture medium, the stock solution of the SiNPs, which was 10 mg/mL in DPBS (Gibco), were sonicated for 30 min. HCECs were cultured in the SiNP-containing culture medium in a 5% CO 2 and 95% air-humidified atmosphere at 37 °C for 24 h or 48 h.

Electron Microscopy Analysis. The intracellular distribution of SiNPs in HCECs was investigated by TEM as
previously described 11 . The HCECs were treated with three sizes of SiNPs (100 µg/ml) for 24 h, and then the cells were fixed in 3. LDH Assay. Necrotic cell death with plasma membrane damage was evaluated using an LDH cytotoxicity detection kit (Takara Bio Inc., Shiga, Japan) 10 . The experimental procedure was previously described and performed following the manufacturer's protocol 11 . Briefly, HCECs were cultured at 3 × 10 3 cells/well in a 96-well plate and incubated for 24 h and 48 h. Following the adherence of cells, 50, 100, and 150 nm SiNPs were applied to SCIEntIfIC REPoRtS | 7: 14566 | DOI:10.1038/s41598-017-15247-2 the cells for 48 h dose dependently at 0, 25, 50, and 100 µg/ml. The wells with no SiNP addition and the wells with 1% triton X-100 addition were used as the negative and positive controls, respectively. Following the incubation of cells, cell free supernatants were transferred into a new 96-well plate. The wells were incubated with the reaction mixture for 20 min at room temperature (RT). Absorbance at 490 nm was measured.
Cell Viability Assay. A commercial cell counting kit (CCK-8; Dojindo Molecular Technologies, Inc., Kumamoto, Japan) was used to measure HCECs's viability according to the manufacturer's protocol 11 . Briefly, the HCECs were cultured at 3 × 10 3 cells/well in a 96-well plate and incubated for 24 h. Following the adherence of cells, 50, 100, and 150 nm SiNPs were added to the culture media for 24 h and 48 h over a range of concentrations (0, 25, 50, and 100 µg/ml). After the appropriate incubation, 10 µL of CCK-8 solution was added to each cultured well, and absorbance at 450 nm was determined after 2 h of incubation of the HCECs with the reagent at 37 °C.
Western Blot Analysis. Western blot analysis was performed following the previously reported method 11 .
All SiNP-treated HCECs were lysed in an ice-cold radioimmunoprecipitation assay buffer [ Immunocytochemistry. As previously reported 11 , the HCECs were seeded at a density of 2 × 10 4 cells per milliliter and grown on 4-well Lab-Tek chamber slides (Nalgene Nunc International, Penfield, NY, USA), and 100 µg/mL of 50, 100, 150 nm-sized SiNPs was treated for 24 h. The HCECs were fixed with 3.7% paraformaldehyde for 10 min at RT, and permeabilization was performed using 0.1% triton x-100 for 5 min at RT 11 . After washing with DPBS, the blocking of non-specific antigen site was done using 1% BSA in DPBS for 30 min at RT. were provided by Eversight International (Seoul, South Korea). Two corneas were from a 63-year-old male donor and another two corneas were from a 59-year-old male donor. Informed consent regarding the possible research use of the corneas was obtained from the donors when they agreed on tissue donation form. The experiment followed the tenets of the Declaration of Helsinki and was approved by the institutional review board of Dongguk University, Ilsan Hospital, Goyang, South Korea. The corneas were harvested at the date of death of the donors and used for the experiment at 9 and 11 days after death. Fresh human corneal donor buttons (n = 4) were divided by quadrant segments. Each segment was stained with 0.005% trypan blue mixed with minimum essential medium (MEM) for 5 min. Corneal endothelial cell viability was assessed by examining the blue-stained area under an inverted-phase contrast microscope. After the baseline viability assessment, the corneal segments were divided into four groups (n = 4 in each group), and each segment was incubated at 37 °C in a 5% CO 2 and 95% air-humidified atmosphere for 72 h. The tissue culture medium was serum-free MEM containing L-glutamine (2 mM), NaHCO3 (20 g/L), penicillin (100 IE/mL), and streptomycin (0.1 mg/mL). Three different sizes (50, 100, and 150 nm) of SiNPs were mixed in the tissue culture medium at a 100 µg /mL concentration (n = 4 each group). The tissue culture medium with no SiNPs was used for the control group (n = 4). After 72 h of incubation with SiNPs, the corneal segments were stained again with 0.005% trypan blue mixed with MEM for 5 min. Any increase in the blue-stained area from the baseline was used as the indicator of corneal endothelial cell toxicity, and these were compared among the groups. After trypan blue evaluation, the corneal segments (n = 3 each group) were fixed in 10% formalin and incubated with TRITC-conjugated phalloidin (1 µg/mL; Sigma-Aldrich).
The tissue was rinsed three times with PBS (5 min per rinse), and the whole-mount corneas were mounted endothelial-side down on a slide and stained with DAPI (catalog number: P36931; Molecular Probes). The slides were examined with a fluorescent microscope.
In Vivo Rabbit Corneal Endothelial Cell Toxicity Assay. To investigate the in vivo corneal effect of SiNPs, we used 12 New Zealand white rabbits (males, weighing 2.5-3.0 kg). The animals were treated in compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Experimental protocol was approved by Institutional Animal Care and Use Committee of Dongguk University, Ilsan Hospital (reference number: 2016-03146). The rabbits were divided into four groups (n = 3, each group), including the negative control group, and treated with intracameral injection of the SiNP solution (0.05 ml) of each size of SiNP (50, 100, 150 nm-sized SiNPs, 200 μg/mL concentration mixed with balanced salt solution) to the right eyes. To consider the average aqueous humor volume of rabbits (290 μL), the final concentration of SiNP in the anterior chamber was estimated as 30 μg/mL. The left eyes were used for the negative control. Corneal photographs of both eyes were taken at baseline, days 1, 2, and 3, and weeks 1 and 2. Corneal transparency was evaluated according to these photographs. At two weeks after the intracameral injection, the rabbits were euthanized and both eyes were enucleated. Tissues were fixed in 10% formalin, and a histopathologic examination using hematoxylin and eosin staining was performed. One cornea from each group was incubated with TRITC-conjugated phalloidin (1 µg/ mL; Sigma-Aldrich). The tissue was rinsed three times with PBS (5 min per rinse), and the whole-mount corneas were mounted endothelial-side down on a slide and stained with DAPI (catalog number: P36931; Molecular Probes). The slides were examined with a fluorescent microscope. Using the fluorescent pictures, endothelial cell count (DAPI stained nucleus) in three 500 μm × 500 μm square areas was calculated in each cornea using ImageJ software (http://imagej.nih.gov/ij/).

Statistical
Analysis. Data were presented as mean ± standard error, and statistical significance was determined by ANOVA and Dunnett's multiple comparison test. P values of less than 0.05 were regarded as significant by GraphPad Prism Ver. 5.01 (GraphPad Software Inc., La Jolla, CA, USA).