Impact of surface functional group modification on cellular internalization and cytotoxicity of silica nanoparticles

Abstract

Background

Silica nanoparticles (SiO₂NPs) are widely used in industrial products. Surface modification of SiO₂NPs is one of the promising strategies to develop safer nanomaterials by design. The present study was designed to determine the effects of amino or carboxyl functionalization of rhodamine-labeled SiO₂NPs on cellular uptake and cytotoxicity.

Methods

In the in vivo arm of the study, male mice were randomly divided into seven groups (n = 6, each) and exposed to either amino (NH₂)- or carboxyl (COOH)-functionalized, or non-functionalized (OH)-rhodamine-labeled SiO₂NPs at 2 or 10 mg/kg bw, or endotoxin-free water as a control, by pharyngeal aspiration. At 24 h after administration, the mice were euthanized and bronchoalveolar lavage fluid (BALF) was collected for differential cell count and assessment of silica nanoparticle uptake using confocal microscopy. In the in vitro arm of the study, murine RAW264.7 macrophages were exposed to NH₂-or COOH-functionalized or OH- rhodamine-labeled SiO₂NPs. Nonspecific caspase inhibitor, necroptosis inhibitor, pyroptosis inhibitor and autophagy inhibitor were used to determine the roles of cell death signaling in cytotoxicity.

Results

The in vivo studies demonstrated significant increase in lung weight at 2 and 10 mg/kg bw by OH-SiO₂NPs but not the other two SiO₂NPs. At 10 mg/kg bw, COOH-SiO2NPs induced a significant increase in BALF macrophages, whereas OH- SiO₂NPs significantly decreased macrophages. OH-SiO₂NPs at 2 mg/kg bw and NH₂- and COOH-SiO₂NPs at 10 mg/kg bw significantly increased BALF neutrophiles. The in vitro studies showed greater NH₂-SiO₂NPs internalization into RAW264.7 macrophages than OH-SiO₂NPs, while OH-SiO₂NPs induced cytotoxicity and upregulation of IL-1β and TNF-α to greater extent than the other two types. Co-treatment with pan-caspase inhibitor and necroptosis inhibitor attenuated (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium) (MTS) cytotoxicity of OH-SiO₂NPs.

Conclusion

NH₂- or COOH-functionalization reduced the harmful changes observed with OH- SiO₂NPs, which included increase in lung weight and BALF neutrophils at low dose in mice as well as decrease in cell viability and upregulation of proinflammatory cytokines in RAW264.7 macrophages. The results suggested that OH-SiO₂NPs-induced cytotoxicity against macrophages was mediated at least in part through apoptotic/necroptotic signaling but was not related to internalization of particles. The results imply possible development of safer silica nanoparticles by amino- or carboxyl-functionalization of their silanols.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12989-025-00653-6.

Introduction

Silicon dioxide (silica) nanoparticles (SiO₂NPs) are synthetic amorphous silicon dioxide (SAS) nanoparticles used widely in industrial products as additives for rubber and plastics and as strengthening fillers for concrete. In addition, they are used in the biomedical field for drug delivery and theranostic purposes [1–6].

Silica materials exist in both crystalline and amorphous forms. The most common form of crystalline silica is quartz, whose toxicity has been studied for many years and is linked to chronic bronchitis, emphysema and silicosis [5–7]. Compared to the crystalline micron-sized SiO₂, the amorphous form is known to be less toxic [1, 2]. The toxicological profile of SiO₂NPs has been widely studied recently, with increasing body of literature on the potential adverse effects of SiO₂NP exposure. With regard to the mechanism of toxicity, in vitro studies have shown that the toxic effects of SiO₂NPs are mainly mediated through the induction of oxidative stress and activation of intrinsic or mitochondrial apoptotic pathways [8–13]. Reactive oxygen species (ROS)-mediated cell death is considered one of the main mechanisms of action of many different types of nanomaterials, including SiO₂NPs. However, the results of a recent study indicated that the toxicity of SiO₂NPs is not mediated through intracellular ROS but rather by total silanol content, cell membrane damage, and cell viability [14]. The majority of the in vivo toxicological data are based on acute exposure studies, which usually include intra-tracheal instillation, intravenous injection or oral exposure [1, 2]. Submicron amorphous silica particles were found to have greater inflammatory and cytotoxic potential compared to their bigger counterparts [15]. In the study of Morris et al. [20], C57BL/6 mice were intratracheally instilled with 4 or 20 mg SiO₂NPs /kg body weight. 24 h after instillation, approximately tenfold increase in the cell number was observed in the bronchoalveolar lavage fluid (BALF) of mice treated with the bare SiO2NPs at high dose of 20 mg/kg, compared to the control mice; neutrophils were also increased about 1000- and 500-fold, in BALF of mice treated with the bare SiO₂NPs and amine-functionalized SiO₂NPs at the high dose of 20mg/kg, respectively [16]. Other studies demonstrated acute and chronic exposure to SiO₂NPs aggravated airway inflammation [17–21].

One of the strategies to build “safe by design” NPs is to apply various types of surface modifications to coat the NPs and modulate their internalization into cells or their surface reactivity, thereby decreasing their toxic effects. For example, surface modification of SiO₂NPs was found to reduce their aggregation and nonspecific binding [22], while functionalization with amino or phosphate groups was reported to mitigate their pro-inflammatory and immunomodulatory effects in allergic airway inflammation [17]. Interestingly, coating of SiO₂NPs with polyethylene glycol polymer (PEG) did not efficiently reduce their pro-inflammatory potential in vivo [17, 23]. Furthermore, a few in vitro studies have shown that surface modification of SiO₂NPs reduced their potential for inflammasome activation and cytotoxicity [24, 25].

The present study was designed to determine the effect of surface modification of 30 nm amorphous SiO₂NPs both in vivo (C57BL/6JJcl mice) and in vitro (murine macrophage RAW 264.7 cell line), focusing on the internalization process of these particles into the macrophages, as well as their pro-inflammatory and cytotoxic potentials. Our original hypothesis was that surface modification of amorphous SiO₂NPs should affect their internalization into the macrophages, which is known to be linked to inflammation. However, the results of the in vivo study showed the unexpected results of OH-SiO₂NPs-induced decrease in BALF macrophages. Accordingly, we decided to extend the study to determine the cytotoxicity of OH-SiO₂NPs in vitro and its relation to cell internalization of NPs and gene expression of proinflammatory cytokines, as well as the involvement of different cell death pathways in the cytotoxicity. These studies were designed to help understand the mechanism of cytotoxicity induced by non-functionalized OH-SiO₂NPs.

Materials and methods

Silica nanoparticles

Rhodamine-labeled synthetic colloidal amorphous SiO₂NPs, “Sicastar “, of 30 nm in diameter, functionalized with amino group (NH₂-SiO₂NPs, catalog #40-01-301), carboxyl (COOH- SiO₂NPs, catalog #40-02-301) and non-functionalized (OH-SiO₂NPs, catalog #40-00-301) were purchased from Micromod Partikeltechnologie (Rostock, Germany) and used in both the in vitro and in vivo arms of this study.

The NH₂- and COOH-functionalization was introduced by diethylentriamin linker and (triethoxysilyl) propylsuccinic anhydride linker, respectively. The rhodamine-labeled particles were produced by hydrolysis of orthosilicates and compounds with fluorescence. The three types of rhodamine-labeled particles differed only in their surface features but not their crystal structures. All NPs were spherical, nonporous, 2.0 g/cm³ in density and dispersed in water at 25 mg/mL. The size of the NPs and polydispersity index in water or in complete cell culture medium was characterized by DLS. Zeta-potential was measured with Photal LEZA-600 (Otsuka Denshi Co., Osaka, Japan). The percentages of nitrogen (N), oxygen (O), carbon (C), and silicon (Si) by weight and by atom were measured using scanning electron microscopy (SEM) equipped with energy-dispersive X-ray spectroscopy (EDX) (VE-7800, KEYENCE, Osaka, Japan). The fluorescence intensities of the three types of silica NPs were measured at different concentrations using ARVO^TMMX 1420 Multilabel Counter (Perkin Elmer, Waltham, MA). The slopes of the regression lines for the independent variable of concentration and the dependent variable of fluorescence intensity were calculated to obtain the relative fluorescence intensity of the different types of silica NPs labeled with rhodamine.

In vivo studies

Animals

Forty-two male C57BL/6JJcl mice (7 week-old) were purchased from CLEA Japan, Inc. (Tokyo). All mice were housed and acclimatized to the laboratory environment for 1 week in a pathogen-free animal room controlled at 23–25 °C and 55–60% humidity. Light was set within a 12 h light–dark cycle (on at 09:00 and off at 21:00), and food and water were provided ad libitum.

The study was conducted according to the Japanese law on the protection and control of animals and the Animal Experimental Guidelines of Tokyo University of Science. The experimental protocol was approved by the Animal Ethics Committee of Tokyo University of Science (#Y16023).

Mice (mean body weight 22.3 ± 1.1 g, ± SD) were randomly divided into seven groups (n = 6, each) and exposed to either endotoxin-free water (as the control) or NH₂-SiO₂NPs, COOH-SiO₂NPs or OH-SiO₂NPs at 2 or 10 mg/kg bw, which were equivalent to 40 or 200 µg per mouse if body weight was 20 g. These exposure levels were half of those adopted in a previous study, which demonstrated that exposure to SiO₂NPs by intratracheal instillation increased the number of macrophages in BALF at 0.5 mg silica/mouse (20 mg/kg bw) but not at 0.1 mg silica/mouse (4 mg/kg bw) [16]. SiO₂NPs dispersed in water at 25 mg/mL were vortexed and then further diluted with endotoxin-free water to obtain the NP solution at 1 and 5 mg/mL.

Mice were anesthetized with pentobarbital and then exposed to 40 µL aliquot of samples of SiO₂NPs by pharyngeal aspiration, as described previously [26]. The technique of pharyngeal aspiration involved placement of the NPs suspension on the back of the tongue followed by pulling of the tongue to induce a reflex gasp with resultant aspiration of the droplets. At 24 h after administration, the mice were euthanized by intraperitoneal injection of pentobarbital. Bronchoalveolar lavage (BAL) was performed by cannulation of the trachea with 18-gauge needle, and infusion and collection of 5/6 mL of saline was repeated six times. The 24 h post-administration time point was selected to observe early inflammatory responses.

BALF total and differential cell counts

The recovered bronchoalveolar lavage fluid (BALF) was centrifuged (1,500 rpm, 5 min, 4°C), and the cell pellet was mixed with 1 mL of ACK lysing buffer (Gibco-Thermo Fischer Scientific, Waltham, MA) for hemolysis. Next, 10 mL of Dulbecco’s phosphate buffer saline (DPBS) was added before centrifugation at 1500 rpm and 4 °C. The resultant pellet was re-suspended in DPBS for total and differential cell counts. BALF total cell count was measured using hematocell counter. Aliquots of 5 × 10⁴ cells in 400 μl DPBS per slide were prepared for cytospins. The cell mixture was added to EZ Single Cytofunnel^®, Thermo, UK and centrifuged for 10 min at 1000 rpm with Cytospin, using cytoslides. The slides were dried overnight at room temperature and then stained with the Differential Quik Stain Kit (Modified Giemsa, Sysmex Co., Kobe, Japan) for differential cell count in 10 fields on each slide (20 ×magnification).

Fluorescence immunocytology

The slides obtained by Cytospin were washed three times in DPBS and then incubated with blocking agent (1% BSA) for one hour. The slides were further incubated with Biotin anti-Ly6G and Ly6C (Gr-1) (BD, Franklin Lakes, NJ), which was diluted 400 folds in 1% BSA, for one hour at room temperature, washed in DPBS three times and then incubated with 200-fold diluted Cy5 streptavidin (BioLegend, San Diego, CA) for 30 min at room temperature to stain the neutrophils. Ly6G and Ly6C (Gr-1) were used as markers of neutrophils, as their expression levels are known to correlate with differentiation and maturation of granulocytes [27, 28] and are only expressed transiently on bone marrow cells in the monocyte lineage [28]. The slides were counterstained with Hochest33342 for 10 min at room temperature and enclosed with Fluorescent Mounting Media (Dako, Agilent, Santa Clara, CA). Cells and SiO₂NPs were observed with a confocal microscope (model Fv10i, Olympus, Tokyo).

In vitro studies

Cell culture

Murine macrophages RAW 264.7 cell were kindly provided by Prof. Kenneth Dawson, University of College Dublin and grown in Dulbecco’s Modified Eagle Medium (DMEM, high glucose) with L-glutamine and phenol red (Wako, catalog #044-29765) supplemented with 1 mM sodium pyruvate (100 mM, Gibco, catalog #11,360–070), 100 U/mL penicillin, 100 mg/mL streptomycin and 250 ng/mL amphotericin B (anti-anti, Gibco, catalog #15,240-062), 2 mM glutamine and 10% FBS. All experiments were performed with cells from passages 4 to 15. Cells were grown in T25-flasks (Violamo, AS ONE Co., Osaka) in monolayers. Exponentially growing cells were maintained under a humidified atmosphere of 5% CO₂ and 95% air at 37 °C and were passaged once every two days using a cell scraper (NEST 100071).

NP treatment

Depending on the experiment and after reaching 70–80% confluence, the RAW 264.7 cells were seeded onto appropriate cell culture plates and treated with one of the three SiO₂NPs. Before exposure to SiO₂NPs, the RAW 264.7 cells were rinsed with PBS to eliminate trace amounts of FBS. Treatments were performed under FBS-free condition for three reasons: (1) serum is reported to modulate NPs uptake [29], and (2) to mimic in vivo condition whereby bronchial cells are not directly exposed to serum proteins. (3) bovine serum albumin (BSA) is known to induce agglomeration of SiO₂NPs [30]. Stock of 30 nm rhodamine-labeled silica nanoparticles (25 mg/mL in water) was vortexed shortly before the preparation of the final dilution for the treatment.

MTS assay

The (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium) (MTS) assay was conducted using CellTiter 96® Aqueous One Solution Cell Proliferation Assay (Promega, Madison, WI), as described previously [31] using the instructions provided by the manufacturer. Briefly, RAW264.7 cells were seeded at 1.5 × 10⁴ cells/well onto 96-well plates and incubated at 37 ºC under a humidified atmosphere of 5% CO₂ and 95% air for 24 h. After incubation, the cell culture medium was removed from each well with a multichannel pipette, and the cells were washed three times with DPBS to remove FBS. The cells were incubated for 4 or 24 h with one of the three types of SiO₂NPs dispersed in FBS-free cell culture medium at a final concentration ranging from 0.3 to 30 µg/cm². After incubation with the SiO₂NPs, the cells were washed twice with DPBS, and incubated with MTS reagent (1.4 mL CellTiter 96® AQueous One Solution Reagent and 7.1 mL of complete phenol red-free cell culture medium with FBS). Cell viability was determined by measuring absorbance at 490 nm, which reflected the reduction of {3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium} (MTS) to formazan by mitochondria in viable cells.

Evaluation of OH-SiO₂NPs-induced interference in MTS assay

Possible interference by hydroxylated silica nanoparticles (OH-SiO₂NPs) in the MTS assay was evaluated following the method of Hirsch [32]. OH-SiO₂NPs were dispersed in fetal bovine serum (FBS)-free cell culture medium at final concentrations ranging from 0.3 to 30 µg/cm² and incubated for 60 min at 37 °C in 5% CO₂ with one of the following: (1) phenol red-free cell culture medium; (2) phenol red-free medium mixed with CellTiter 96^® AQueous One Solution Reagent at a 5:1 ratio; or (3) 5 mL of phenol red-free medium containing 1 mL of CellTiter 96® AQueous One Solution Reagent and 50 µL of Na₂SO₃. After incubation, absorbance was measured at 490 nm.

Confocal microscopy

RAW 264.7 cells were seeded in 8 well Lab-TekII^™ chambered cover glasses (Nunc, Thermo Scientific, Dominique Dutscher, Brumath, France) at 16,590 cells/well in complete cell culture medium, and then incubated for 24 h. Upon 70–80% confluence, the cells were washed three times with DPBS to remove FBS and treated with 0.3 mL/well of NPs at pre-selected concentrations for 24 h at 37 °C and 5% CO₂. After removal of the medium, the cells were washed once and fixed in 4% paraformaldehyde for 20 min at 25 °C, rinsed twice with DPBS and then incubated with Cell Mask Green Plasma Membrane Stain (C37608 Thermo Fisher Science, Waltham, MA) and Hoechst for 10 min. The cells were washed with DPBS twice. After embedding the cells in Mounting Medium, they were observed by confocal microscopy (model FV.10, Olympus, Tokyo). The maximum concentration of silica NPs of 35.2 × 10⁵ µg/cell (8.37 µg/cm²) was determined to be equivalent to exposure level of 40 µg/mouse in vivo, given that the average number of macrophages collected in BALF was 1.14 × 10⁵ cells/mouse.

Flow cytometry

RAW264.7 cells were seeded onto 6-well plates at 2.37 × 10⁴ cells/cm² in complete cell culture medium and incubated for 24 h before treatment. After treatment with 3.0 mL/well of SiO₂NPs at preselected concentrations for 1 or 4 h in dark, the medium was removed, cultures were thoroughly washed with PBS three times and treated with 0.1% trypan blue for 1 min to quench the fluorescence of rhodamine on the cell surface. The concentration of trypan blue for quenching the fluorescence of rhodamine was determined beforehand by plotting trypan blue and rhodamine intensity in RAW264.7 cells exposed to OH-rhodamine-labeled SiO₂NPs. The cells were washed with PBS, mixed with 500 µL of FACS buffer (PBS containing 0.5% FBS and 0.1% NaN₃) and harvested by cell scraper. Cell-associated fluorescence was detected using FACSCalibur^™ and results were analyzed with FlowJo software (BD, Franklin Lakes, NJ). The mean fluorescence intensity (MFI) of rhodamine (excitation 488 nm, Filter range 564–606 nm) from three different size fractions indicated by the forward scatter (FS) was computed by the flow cytometer. The results are reported as the median of the distribution of cell fluorescence intensity obtained by analyzing cells in the gate. To adjust for the differences in the fluorescence intensity relative to weight, the intensity of rhodamine fluorescence was measured at different concentrations of three types of S SiO₂NPs using ARVOMx-fla system (485 nm/535 nm 1.0 s).

LDH cytotoxicity assay

The LDH cytotoxicity assay was conducted to quantify cytotoxicity by measuring the release of the enzyme lactate dehydrogenase (LDH) from damaged cells into the culture medium. The assay was conducted using Pierce LDH cytotoxicity assay kit following the instructions provided by the manufacturer (Thermo Fisher Scientific, Waltham, MA). Briefly, RAW264.1 cells were plated at 10⁴ cells/well in 100 µl of medium in a 96-well tissue culture plate. After incubation at 37  C under 5% CO₂ for 24 h, the cells were exposed to NH₂-, COOH-functionalized, OH-rhodamine-labeled non-functionalized SiO₂NPs at 5.85 µg/cm² (19.5 µg/mL). After exposure for 1, 4, 12, 24, 36 and 48 h, the supernatant of the culture medium was collected by centrifugation, and dispensed at 50 µl/well into another 96-well plate. Each cell was incubated in the presence of 50 µl of LDH reaction mixture at room temperature for 30 min under darkness. The reaction was stopped by adding 50 µl of stop solution, and absorbance at 490 and 680 nm was read by Plate reader Gen5 (BioBik, Osaka).

Evaluation of cell death by cell proliferation assay

Cell death pathways involved in the cytotoxicity of OH-SiO₂NPs were examined using the following inhibitors: cell-permeable pan-caspase inhibitor Z-VAD-FMK (#G7231A, Promega, Madison, WI, USA), necroptosis inhibitors (RIP1-specific inhibitors) necrostatin-1, (#11,658, Cayman Chemical, Ann Arbor, MI, USA) and GSK”963 (#S864201, Sellekchem, Houston, TX, USA), pyroptosis inhibitor (caspase-1 and -4 inhibitor) Z-YVAD-FMK (#ab141388, Abcam, Cambridge, UK) and autophagy inhibitor (inhibitor of autophagosome-lysosome fusion) chloroquine diphosphate salt (#C-6628, Sigma-Aldrich, Saint Louis, MO, USA). Z-VAD-FMK, necrostatin-1, GSK’963 and Z-YVAD-FMK were each dissolved at 20 or 50 mM in dimethyl sulfoxide (DMSO), and chloroquine diphosphate salt was dissolved at 40 mM in endotoxin-free water. Each of the resultant inhibitor in vehicle was further diluted by the culture medium to the final concentration. RAW264.1 cells were preincubated with Z-VAD-FMK, necrostatin-1, Z-YVAD-FMK or chloroquine, or co-treated with Z-VAD-FMK and GSK’963, or corresponding vehicle for 1 h, and exposed to plain rhodamine-labeled silica NPs at 28 µg/mL for 18 h. The cytotoxicity assay was conducted for each inhibitor to determine the maximum concentration of the inhibitor associated with lack of decrease in cell viability. Cytotoxicity was evaluated using CellTiter 96^® Aqueous One Solution Cell Proliferation Assay (Promega) as mentioned above.

Nuclear staining by DAPI

RAW264.1 cells were plated at a density of 1.5 × 10⁴ cells per well in 200 µL of medium in a 96-well tissue culture plate. After incubation at 37 °C under 5% CO₂ for 24 h, the cells were exposed to OH–rhodamine-labeled non-functionalized SiO₂NPs at a concentration of 28 µg/cm². Following an 18 h exposure, the culture medium was removed, and the cells were washed once with D-PBS, then fixed with 4% paraformaldehyde (PFA) for 10 min. After removal of the fixative, the cells were washed three times with DPBS, stained with 4',6-diamidino-2-phenylindole (DAPI)for 1 min, and mounted using mounting medium (#ab104135, Abcam, Cambridge, UK). The cells were observed under a fluorescence microscope (DMI6000 B, Leica, Wetzlar, Germany) equipped with a digital camera (DFC365 FX, Leica, Wetzlar, Germany).

Quantitative real-time PCR

Quantitative RT-PCR was used to quantify the mRNA levels of various pro-inflammatory cytokines induced by SiO₂NPs. For this purpose, RAW264.1 cells were seeded onto 6-well plates at 23,700 cells/cm² and exposed to silica NPs at concentrations of 0.33, 1.67 and 8.37 µg/cm² (1.06, 5.30, and 26.50 µg/mL, respectively). The cells were collected by centrifugation at 1000 rpm for 5 min at 4 ºC. Total RNA from the cells was isolated by using ReliaPrep™ RNA Cell Miniprep System (Promega, Fitchburg, WI). The concentration of total RNA was quantified by spectrophotometry (ND-1000; NanoDrop Technologies, Wilmington, DE). RNA was reverse transcribed to single-strand cDNA using SuperScript III First-Strand Synthesis System for RT-PCXR (Life Technologies). The cDNA was subjected to quantitative real-time PCR (qRT-PCR) analysis with Thunderbird cyber green master mix (TOYOBO, Osaka) and primers designed by TAKARA (Kusatsu, Japan). The primers were.

5ʹ—GTCCCTCAACGGAAGAACCAA—3ʹ (forward) and 5ʹ—TCTCAGACAGCGAGGCACAT—3ʹ (reverse) for MIP-2, 5ʹ-

5’- CATCCACGTGTTGGCTCA (forward) and 5’- GATCATCTTGCTGGTGAATGAGT-3’ (reverse) for MCP-1, 5’-AGCTTCAGGCAGGCAGTATC-3’ (forward) and 5’-GTCACAGAGGATGGGCTCTT-3’ (reverse) for IL-1β,

5’-GATCGGTCCCCAAAGGGATG-3’ (forward) and 5’-GTGGTTTGTGAGTGTGAGGGT-3’ (reverse) for TNF-α,

and 5’-GATCATTGCTCCTCCTGAGC-3’ (forward) and 5’-ACTCCTGCTTGCTGATCCA-3’ (reverse) for β-actin. 5’-GCCTGGAGAAACCTGCCAA-3’ (forward) and 5’-TGAAGTCGCAGGAGACAACC-3’ (reverse) for GAPDH, 5’-GTTCCAGCACATTTTGCGAGT-3’ (forward) and 5’- GGTGAGGTCGATGTCTGCTT-3’ (reverse) for 18S ribosomal RNA,

A standard curve was constructed by serial dilutions of the cDNA sample. All PCR reactions were performed in duplicates and the values were expressed relative to the geometric mean of β-actin mRNA, GAPDH mRNA, and 18S r RNA level. The average of the two reactions was used as the representative mRNA value for each mouse.

Statistical analysis

Data are expressed as mean ± standard deviation (SD). Differences between the test samples and control or among the groups were analyzed respectively by one-way ANOVA followed by Dunnett or Tukey multiple comparison test. Multiple regression analysis with dummy variables for the different types of particles was conducted to test the interaction between SiO₂NP level and the type of NPs. Dummy variable [NH₂] = 1, when the type was NH₂-SiO₂NPs; [NH₂] = − 1, when the type was OH- SiO₂NPs; and [NH₂] = 0, otherwise. Dummy variable [COOH] = 1, when the type was COOH- SiO₂NPs; [COOH] = − 1, when the type was OH-SiO₂NPs; and [COOH] = 0, otherwise. Statistical significance was set at 5%. For the analysis of relative fluorescence intensity among different types of NPs, regression lines were obtained by forcing the intercept to zero using Excel 2016 (Microsoft, Redmond, WA). All statistical analyses were performed using JMP (version 14, SAS Institute, Cary, NC).

Results

Larger hydrodynamic diameter of OH-SiO₂NPs relative to NH₂ and COOH-SiO₂NPs

To understand the physicochemical features of the three types of SiO₂NPs, the hydrodynamic diameters and zeta potentials were determined by DLS and Zeta potential measurements, both in water (since the materials were dispersed in endotoxin-free, ultra-pure water for pharyngeal aspiration in mice) and in the cell culture medium (relevant to the exposure of RAW264.8 cells). The mean hydrodynamic diameter of the OH-SiO₂NPs was significantly larger than those of the NH₂- and COOH- SiO₂NPs in water and in FBS-free cell culture medium (Table 1). On the other hand, polydispersity index was significantly greater in COOH-SiO₂NPs than OH-SiO₂NPs but was less than 0.7 both in water and FBS-free cell culture medium, indicating appropriate condition for DLS measurement. The physicochemical features of the three types of SiO₂NPs (in water and in FBS-free cell culture medium) had no effect on the Zeta potential. The pH of OH- SiO₂NPs was significantly lower than that of NH₂- SiO₂NPs in water while there was no significant difference in the pH among the three types of SiO₂NPs in FBS-free cell culture medium.

Table 1.

Physicochemical properties of SiO₂NPs dispersed in water and in complete cell culture medium, studied by DLS and Zeta potential measurement

Type of SiO2NPs	Solvent	Hydrodynamic diameter (nm)	Polydispersity index	Zeta potential (mV)	pH
NH2-SiO2NPs	Water	30.91 ± 0.02§	0.12 ± 0.02§¶	− 21.0 ± 6.2§	6.51 ± 0.12§
COOH-SiO2NPs	Water	29.25 ± 0.48§	0.15 ± 0.04§	− 29.4 ± 7.3§	6.26 ± 0.27§¶
OH- SiO2NPs	Water	33.87 ± 1.11¶	0.12 ± 0.02¶	− 22.3 ± 2.7§	5.92 ± 0.10¶
NH2-SiO2NPs	FBS-free medium	31.68 ± 0.40§	0.19 ± 0.01§¶	− 20.0 ± 0.8§	7.97 ± 0.12§
COOH-SiO2NPs	FBS-free medium	29.36 ± 0.28¶	0.25 ± 0.08§	− 19.8 ± 1.4§	7.94 ± 0.12§
OH- SiO2NPs	FBS-free medium	34.87 ± 0.38†	0.17 ± 0.03¶	− 16.1 ± 2.9§	7.93 ± 0.17§

Data are mean ± SD of three measurements. Tukey–Kramer multiple comparison was conducted following ANOVA between NH₂-SiO₂NPs, COOH-SiO₂NPs and OH-SiO₂NPs in each of Water or FBS-free medium. There is a significant difference between the different types of SiO₂NPs, which are not marked by the same symbols. Significance level was set at 5%

OH-SiO₂NPs, NH₂-SiO₂NPs and COOH-SiO₂NPs showed relatively higher abundance of O, N and Si element, respectively, compared with the other types of SiO₂NPs

SEM-DEX analysis revealed that the surface of NH₂-SiO₂NPs contained significantly higher amounts of nitrogen (N) and carbon (C), both by weight and by atom, than that of OH-SiO₂NPs (Table 2). In contrast, OH-SiO₂NPs exhibited significantly higher oxygen (O) content, both by weight and by atom, than the other types of SiO₂NPs. The surface of COOH-SiO₂NPs contained significantly more silicon (Si), both by weight and by atom, whereas NH₂-SiO₂NPs had significantly less Si compared with OH-SiO₂NPs.

Table 2.

Element ratio of SiO₂NPs quantified by SEM–EDX

			Percentage
Element	Type of SiO2NPs	N	By weight	By atom
N	NH2-SiO2NPs	3	4.24 ± 0.25*	5.55 ± 0.31*
	COOH-SiO2NPs	3	3.61 ± 0.04	4.85 ± 0.05
	OH-SiO2NPs	3	3.18 ± 0.39	4.24 ± 0.50
O	NH2-SiO2NPs	3	40.7 ± 0.2*	46.6 ± 0.2*
	COOH-SiO2NPs	3	41.9 ± 0.3*	49.3 ± 0.4*
	OH-SiO2NPs	3	44.7 ± 0.4	52.2 ± 0.7
C	NH2-SiO2NPs	3	13.8 ± 0.2*	21.0 ± 0.2*
	COOH-SiO2NPs	3	10.3 ± 0.3	16.2 ± 0.5
	OH-SiO2NPs	3	10.0 ± 0.5	15.6 ± 0.7
Si	NH2-SiO2NPs	3	41.3 ± 0.3*	26.9 ± 0.3*
	COOH-SiO2NPs	3	44.2 ± 0.1*	29.6 ± 0.1*
	OH-SiO2NPs	3	42.1 ± 0.5	28.0 ± 0.5

*P < 0.05, compared to OH-SiO₂NPs by ANOVA followed by Dunnett’s multiple comparison. SEM–EDX: scanning electron microscopy equipped with energy-dispersive X-ray spectroscopy

In vivo experiments

OH-SiO₂NPs increased lung weight in vivo

To evaluate the effects of each functional group on the toxicity of the three types of SiO₂NPs, we measured body and lung weights in mice exposed to each type of SiO₂NPs. Pharyngeal aspiration of these SiO₂NPs at 2 or 10 m/kg bw had no effect on body weight (Table 3). On the other hand, the non-functional OH- SiO₂NPs, but not NH₂- or COOH- SiO₂NPs, induced a significant increase in lung weight at both 2 and 10 mg/kg bw. These results suggest that OH- SiO₂NP is injurious to the lung of mice, causing increased lung weight in vivo and that NH₂- and COOH-functionalization mitigated this harmful effect.

Table 3.

Weight of body and lungs of mice at 24 h after exposure to SiO₂NP by pharyngeal aspiration

	Concentration of SiO2NPs (mg/kg bw)	n	Body weight (g)	Lung weight (mg)
Vehicle	0	6	20.9 ± 1.4	293 ± 22
NH2-SiO2NPs
	2	6	21.6 ± 0.4	301 ± 5
	10	6	21.3 ± 1.4	321 ± 23
COOH-SiO2NPs
	2	6	21.6 ± 1.2	305 ± 16
	10	6	21.5 ± 1.0	322 ± 6
OH-SiO2NPs
	2	6	21 ± 1	324 ± 25*
	10	6	20 ± 1	331 ± 16*

^*P < 0.05, compared to the vehicle control, by ANOVA followed by Dunnett’s multiple comparison. Multiple regression analysis with dummy variables for types of particles did not show any significant interaction for body and lung weights

OH-SiO₂NPs induced a greater increase in BALF neutrophils COOH- or NH₂-SiO₂NPs at the low dose

To evaluate NPs-induced lung inflammatory response, BALF was collected for total and differential cell counts. At 2 mg/kg bw, any type of SiO₂NPs did not significantly change BALF total cell (Fig. 1A) or macrophage count (Fig. 1B), while only OH- SiO₂NPs, significantly increased BALF neutrophil count (Fig. 1C).

Fig. 1.

BALF total and differential cell count in mice exposed to SiO₂NPs for 24 h via pharyngeal aspiration. OH-, COOH- and NH₂-functonalized rhodamine-labeled SiO₂NPs by pharyngeal aspiration at 2 or 10 mg/kg bw. At 24 h after the treatment, BALF was collected and total and differential cell counts were determined using Differential Quick Stain Kit and hematocell counter. About 400 cells in three fields were counted on a slide from one animal each. Data are mean ± SD (n = 6 each) A BALF total cell count. B BALF macrophage cell count. C BALF neutrophil cell count. *P < 0.05, compared with the vehicle group, by ANOVA followed by Dunnett multiple comparison test. There is a significant difference between the different types of SiO₂NPs at the same concentration, which are not marked by the same symbols (§,¶,Ͱ and†,‡)

COOH-SiO₂NPs increased BALF neutrophils to a greater extent than NH₂-or OH-SiO₂NPs at the high dose

At 10 mg/kg bw, COOH-SiO₂NPs induced the largest increase in BALF macrophage count (Fig. 1B), and NH₂- and COOH-SiO₂NPs each significantly increased total BALF cell and neutrophil counts (Fig. 1A and C), while OH- SiO₂NPs significantly decreased BALF macrophage count (Fig. 1B). Multiple regression analysis showed significant interaction between type of SiO₂NPs and dose for BALF total cell or neutrophil count, but not for macrophage count (Suppl. Table 1), indicating significant effect of type of SiO₂NPs on the intensity or direction of SiO₂NPs -induced change in BALF total cell or neutrophil count. Additional multiple regression analysis without interaction showed significant effect of type of SiO₂NPs in BALF macrophage count. These results suggest that OH- SiO₂NPs induced lung inflammatory response marked by neutrophil accumulation at low dose and reduced BALF macrophages at high dose.

Fewer macrophages and neutrophiles with internalized OH-SiO₂NPs

To determine the impact of the functional groups on the fate of NPs after pharyngeal aspiration, the macrophages and neutrophils were differentially stained and examined by confocal microscopy. COOH- and NH₂- SiO₂NPs were internalized in macrophages regardless of the dose, while internalization of OH-SiO₂NPs in the macrophages was less pronounced (Fig. 2A). Quantitative analysis confirmed that around 60% of macrophages showed internalization of NH₂- and COOH- SiO₂NPs at 2 and 10 mg/kg bw, compared with less than 40% for OH-SiO₂NPs (Fig. 2B and C). With regard to neutrophils, the numbers or percentages of neutrophils that showed COOH- and NH2-SiO₂NPs internalization increased dose-dependently, but no such dose-dependent increase was noted for OH-SiO₂NPs (Fig. 2D and E). Multiple regression analysis showed significant interaction between type of SiO₂NPs and dose in multiple regression model for the number of macrophages with internalized SiO₂NPs and the number or percentage of neutrophils with internalized SiO₂NPs, indicating significant effect of type of SiO₂NPs on the intensity or direction of SiO₂NPs -induced change in the above responses, but not for the percentage of macrophages with internalized SiO₂NPs (Suppl. Table 2). Additional multiple regression analysis without interaction showed significant effect of type of SiO₂NPs. These results suggest a weaker in vivo OH- SiO₂NPs internalization in macrophages and neutrophiles.

Fig. 2.

OH-, NH₂- or COOH-SiO₂NPs internalized in macrophages and neutrophils in BALF collected from mice exposed to each ofthese NPs by pharyngeal aspiration. At 24 h after the treatment, BALF was collected and cytospinned on slides. A Confocal images of the cells recovered from BALF and stained using Differential Quick Stain Kit. Rhodamine-labelled SiO₂NPs are stained red, while the nuclei and cytoplasm are stained dark blue and light blue, respectively. Scale bars = 50 µm. B Count of macrophages with internalized rhodamine-labelled SiO₂NPs. All 71–576 cells in three fields on the slide from each animal were counted. C Counts of neutrophils with internalized rhodamine-labelled SiO₂NPs. All 0–490 cells in three fields on the slide from each animal were counted. D Percentage of the macrophages with internalized NPs relative to the total number of macrophages. E Percentage of neutrophils with internalized NPs relative to the total neutrophils. ANOVA followed by Tukey multiple comparison test was used to determine the significance of the differences between the different types of SiO₂NPs at the same concentration. There is a significant difference between the different types of SiO₂NPs at the same concentration, which are not marked by the same symbols (§,¶,Ͱ and†,‡). Significance level was set at 5%

In vitro experiments

OH-SiO₂NPs is the most toxic to cultured murine RAW264.7 macrophages

Since OH-SiO₂NPs dose-dependently decreased BALF macrophages (Fig. 1B), we updated our hypothesis: that OH-SiO₂NPs are the most cytotoxic NPs. To test this hypothesis, we conducted MTS cell viability assay and LDH cytotoxicity assay for the three SiO₂NPs in RAW264.7 macrophage cell line. Treatment of these cells with OH- SiO₂NPs, but not NH₂- and COOH- SiO₂NPs, significantly decreased cell viability measured by MTS, in time- and dose-dependent manners (Fig. 3 and Suppl. Figure 1). Multiple regression analysis showed significant interaction between type of SiO₂NPs and dose, both after 4 h and 24 h exposure, indicating significant effect of type of SiO₂NPs on the intensity of SiO₂NPs -induced decrease in MTS cell viability (Suppl. Table 3). The LDH cytotoxicity test also showed higher LDH release from cells exposed to OH-NPs compared to the other types of NPs (Suppl. Figure 2). Since OH-SiO₂NPs caused the greatest reduction in MTS cell viability among the tested particles, we evaluated the potential interference of OH-SiO₂NPs in the MTS assay, including possible absorbance generation by the nanoparticles themselves, inhibition of formazan formation, or alteration of formazan absorbance. The results showed no evidence of interference by OH-SiO₂NPs at concentrations ranging from 0.3 to 30 µg/cm² (Suppl. Figure 3). These results are in agreement with those of the in vivo experiments and suggest that OH-SiO₂NPs are the most toxic particles to murine RAW264.7 macrophage cell line, and that NH₂- and COOH-functionalization mitigated this harmful effect.

Fig. 3.

Viability of RAW264.8 cells, as determined by MTS, at 4 or 24 h after treatment with SiO₂NPs at the indicated concentrations. RAW264.8 cells were treated with OH-, NH₂ or COOH-modified rhodamine-labelled SiO₂NPs dispersed in the cell culture medium at 0.3 to 30 µg/cm² (1–100 µg/mL). After the treatment, cell viability was assessed by measuring absorbance at 490 nm, which reflected the reduction of {3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium} (MTS) to formazan by mitochondria in viable cells. Data are mean ± SD (n = 6). *P < 0.05, compared with untreated control (0 µg/cm²), by Dunnett’s test following ANOVA

Weaker OH-SiO₂NPs internalization into RAW264.8 than NH₂- SiO₂NPs

Next, we assessed the internalization of the three types of NPs into RAW264.8 macrophages in order to understand how internalization is associated with OH- SiO₂NPs toxicity. Confocal microscopy showed that OH-SiO₂NPs were mainly localized in the plasma membrane of RAW264.8 macrophages, irrespective of the dose used (Fig. 4A, B, C and D). This was remarkable after 1 h of treatment at the largest concentration of NPs and even more prominent after 4 and 24 h of treatment. These results suggest weaker OH- SiO₂NPs internalization into RAW264.8 macrophage cell line, which are in agreement with the results of the in vivo experiments and implied that internalization of these NPs does not play a role in their toxicity.

Fig. 4.

Colocalization of SiO₂NPs in RAW264.8 macrophages exposed to OH-, NH₂- or COOH-SiO₂NPs for 1, 4 and 24 h. RAW 264.8 cells were exposed to different concentrations of the three types of SiO₂NPs for A 1 h, (B, D) 4 h, and (C) 24 h. Internalization of NPs in the cells was assessed by confocal microscopy. Green – plasma membrane, red – SiO₂NPs. Scale bars = 100 μm. The magnification is the same in all photomicrographs of (A), (B), and (C), and in all photomicrographs of (D). Magnification of objective lens is × 63 for (A), (B), (C) and × 100 for (D)

Confirmation of weaker OH-SiO₂NPs internalization by flow cytometry

Internalization of NPs into the cells was further quantified by flow cytometry (Fig. 5 and Suppl. Figures 4). To eliminate any signal originating from NPs adsorbed on the surface of the cells, 0.1% trypan blue was added shortly before the analysis. The working concentration of trypan blue was optimized in a series of preliminary experiments (Suppl. Figure 4). Flow cytometry identified three populations of RAW264.4 cells, designated as gates A, B, and C, showing distinct forward scatter (FS) signals that reflected differences in cell size (Suppl. Figure 5). Because the slopes of the regression lines, with rhodamine intensity as the dependent variable and SiO₂NP concentration as the independent variable, differed among SiO₂NP types (Suppl. Figure 6), the rhodamine intensities of COOH-SiO₂NPs and OH-SiO₂NPs were normalized by dividing each value by the ratio of their respective slope to that of NH₂-SiO₂NPs, yielding the relative rhodamine intensity for comparison. Exposure to COOH-, NH₂- and OH-SiO₂NPs increased the mean value of the normalized fluorescence intensity in the population gated to A, B and C (Fig. 5 and Suppl. Figure 4), and such increase was dose-dependent.

Fig. 5.

Assessment of internalization of SiO₂NPs into RAW264.8 cells by flow cytometry. Cells were exposed to different concentrations of NPs (0.33, 1.67, and 8.37 mg/cm²) for 1h or 4 h. After the treatment, the cells were washed and incubated with 0.1% trypan blue for 1 min, in order to quench fluorescence emanating from NPs adsorbed on the surface of the cells. Mean fluorescence intensity of the cells was analyzed using excitation/emission wavelengths set corresponding to rhodamine dye. Data are MFI ± SD (n = 3, 10,000 cells were analyzed per sample). ANOVA followed by Tukey multiple comparison test was applied to test for differences among the three types of SiO₂NPs (*P < 0.05). n.s not significant

Following exposure of the cells to the three types of SiO₂NPs at 0.33 µg/cm² (representing the non-cytotoxic level of OH-SiO₂NPs, measured by the MTS assay), for 1 or 4 h, the mean relative rhodamine intensity for the population gated to A, B and C was higher in the order of NH2- > OH- > COOH-SiO₂NPs. However, at 1.66 or 8.37 µg/cm² for 1 or 4 h, the mean relative rhodamine intensity was higher in the order of NH2- > COOH- > OH-SiO₂NPs. Specifically, for 1.66 and 8.37 µg/cm², the mean relative rhodamine intensity increased time-dependently for NH₂ and COOH-SiO₂NPs, but the time-dependent increase was suppressed for OH-SiO2NPs (Fig. 5). The results of flow cytometry confirmed weaker OH-SiO₂NPs internalization into RAW264.7 macrophages, compared with COOH- and NH₂- SiO₂NPs.

Co-treatment with apoptosis and necroptosis inhibitors partially suppresses OH-SiO₂NPs-induced cell viability

Next, we determined the cell death signaling involved in the cytotoxic effect of OH-SiO₂NPs on RAW264.7 cells, using MTS assay with pan caspase, necroptosis, pyroptosis and autophagy inhibitors. No change in cell viability was noted in the presence OH-SiO₂NPs and any of these inhibitors alone for 18 h (Fig. 6A, B, C and D), suggesting that OH-SiO2NPs cytotoxicity is not explained by apoptosis, necroptosis, pyroptosis or autophagy. However, the OH-SiO₂NPs’ cytotoxic effect on cultured RAW264.7 cells was partially reduced in the combined presence of pan-caspase inhibitor and necroptosis inhibitor (Fig. 6E). Finally, DAPI nuclear staining of RAW264.7 cells exposed to 8.37 µg /cm^{2 (}28 µg/mL) OH- SiO₂NPs did not show nuclear fragmentation but showed chromatin condensation (Fig. 7). These results suggest that OH- SiO₂NPs-induced reduction in cell viability is not due to typical apoptosis of RAW264.7 cells.

Fig. 6.

Effect of treatment with inhibitors of pan-caspase, necroptosis, pyroptosis, and autophagy and co-treatment with pan-caspase inhibitor and necroptosis inhibitor on MTS cytotoxicity of OH-SiO₂NPs in vitro. There is a significant difference between the different treatments, which are not marked by the same symbols (§,¶,Ͱ) (by Tukey- Kramer method following ANOVA.). Significance level was set at 5%. A Effect of pan-caspase inhibitor Z-VAD-FMK on MTS viability of cells exposed to OH-SiO₂NPs. RAW264.1 cells were preincubated with 20 or 40 µM of pan-caspase inhibitor Z-VAD-FMK or vehicle for one hour, and exposed to OH-SiO₂NPs at 8.37 µg/cm² (28 µg/mL) for 18 h. Cytotoxicity was evaluated using CellTiter 96® Aqueous One Solution Cell Proliferation Assay (Promega). Concentrations of inhibitors larger than 20 µM were not used due to their cell toxicity. PCI: pan caspase inhibitor (Z-VAD-FMK), 20: 20 µM, 40: 40 µM. B Effect of necroptosis inhibitor necrostatin-1 (RIP1-specific inhibitor) on MTS viability of cells exposed to OH-SiO₂NPs. RAW264.1 cells were preincubated with necrostatin-1 at 10 or 20 µM for one hour, and exposed to OH-SiO₂NPs at 8.37 µg/cm² (28 µg/mL) for 18 h. Cytotoxicity was evaluated using CellTiter 96® Aqueous One Solution Cell Proliferation Assay (Promega). Concentrations of inhibitors larger than 20 µM were not used due to their cell toxicity. NEC necrostatin-1(RIP1 kinase inhibitor), 10: 10 µM, 20: 20 µM C Effect of pyroptosis inhibitor Z-YVAD-FMK (caspase-1 and -4 inhibitor) on MTS viability of cells exposed to OH-SiO₂NPs. RAW264.1 cells were preincubated with Z-YVAD-FMK inactive control at 20 µM for one hour, and exposed to OH-SiO₂NPs at 8.37 µg/cm² (28 µg/mL) for 18 h. Cytotoxicity was evaluated using CellTiter 96® Aqueous One Solution Cell Proliferation Assay (Promega). Concentrations of inhibitors larger than 20 µM were not used due to their cell toxicity. Z-YVAD: Z-YVAD-FMK, 20: 20 µM D Effect of autophagy inhibitor chloroquine disphosphate (inhibitor of autophagosome-lysosome fusion) on MTS viability of cells exposed to OH-SiO₂NPs. RAW264.1 cells were preincubated with chloroquine diphosphate at 5 µM for one hour, and exposed to OH-SiO₂NPs at 8.37 µg/cm² (28 µg/mL) for 18 h. Cytotoxicity was evaluated using CellTiter 96® Aqueous One Solution Cell Proliferation Assay (Promega). Concentrations of inhibitors larger than 5 µM were not used due to their cell toxicity. CQ: Autophagy inhibitor, chloroquine diphosphate, 5: 5 µM E Effect of co-treatment with pan-caspase inhibitor Z-VAD-FMK and necroptosis inhibitor GSK’963 (RIP1-specific inhibitors) on MTS viability of cells exposed to OH-SiO₂NPs. RAW264.1 cells were preincubated with Z-VAD-FMK at 10 µM and GSK’963 at 100 µM for one hour, and exposed to OH-SiO₂NPs at 8.37 µg/cm² (28 µg/mL) for 18 h. Cytotoxicity was evaluated using CellTiter 96® Aqueous One Solution Cell Proliferation Assay (Promega). Concentrations of pan-caspase inhibitor or necroptosis inhibitor larger than10 or 100 µM were not used due to their cell toxicity. PCI: pan caspase inhibitor (Z-VAD-FMK), GSK: GSK’963 (necroptosis inhibitor/RIP1 kinase inhibitor)

Fig. 7.

Fluorescence image of RAW264.7 nuclei stained by DAPI. We examined nuclear morphology to determine whether 28 µg/cm₂ OH-SiNP can induce apoptosis of RAW264.7. Note the lack of signs of apoptosis (e.g., nuclear fragmentation or chromatin limbing) in the presence of OH-SiO₂NPs a, b RAW264.7 cells exposed to the culture medium only for 18 h, c, d RAW264.7 cells exposed to OH-SiO₂NPs at 8.37 µg /cm² (28 µg/mL) for 18 h. Scale bars = 100 µm

Upregulation of IL-1β and TNFα by OH-SiO2NPs in RAW264.8 macrophages

Our in vivo experiments showed that pharyngeal aspiration of NPs in mice induced lung weight changes associated with BALF inflammatory cell accumulation. Accordingly, we assessed the effects of these NPs on the expression of pro-inflammatory cytokines or chemokines using our in vitro model of RAW264.8 macrophages and qRT-PCR. Since exposure to OH-SiO₂NPs increased BALF neutrophils at the low dose, the expressions of macrophage inflammatory protein-2 (MIP-2), which is known to recruit and activate neutrophils [33], and monocyte chemoattractant protein-1 (MCP-1), which recruits both neutrophils and monocytes [34], were measured. We evaluated the mRNA expression levels of cytokines and chemokines rather than their protein secretion, as the protein concentrations in the culture medium were near the detection limit of the commercially available ELISA kits. Exposure of the RAW264.8 cell line to SiO₂NPs at 8.37 µg/cm² for one hour significantly increased the mRNA level of MIP-2 in the order of OH- = NH2- > COOH-SiO₂NPs (Table 4). Furthermore, exposure to OH- or NH2-SiO₂NPs for one hour significantly increased the mRNA level of MCP-1 in the order of OH- > NH₂- SiO₂NPs. Multiple regression analysis showed significant interaction between type of SiO₂NPs and dose for all of the examined genes, indicating significant effect of type of SiO₂NPs on the intensity of SiO₂NPs-induced upregulation of the examined gene expression (Suppl. Table 4). Exposure to these SiO₂NPs at 8.37 µg/cm² for 4 h resulted in further increase in the mRNA levels of MIP-2 and MCP-1 but the effect of SiO₂NPs became insignificant. The mRNA level of IL-1β increased significantly during exposure to OH- SiO₂NPs at 8.37 µg/cm² for one hour but decreased during exposure to NH₂- and COOH- SiO₂NPs (Table 5). On the other hand, the mRNA level of TNF-α significantly increased after exposure to all three types of SiO₂NPs at 8.37 µg/cm² for one hour and the increase was in the order of OH- > NH2- > COOH- SiO₂NPs. However, the extent of IL-1β and TNF-α upregulation was smaller by 4-h exposure irrespective of the type of NPs and the differences in the upregulated levels among the three NPs became insignificant at 1.66 and 8.37 µg/cm². Multiple regression analysis showed no significant interaction between type of SiO₂NPs and dose, indicating no effect of type of SiO₂NPs on the intensity or direction of change in the expression of the examined genes (Suppl. Table 4). Considered together, the above results suggest that the cytotoxic effects of SiO₂NPs pharyngeal aspiration include lung inflammation and cell death through apoptotic and necroptotic pathways, and that the former is linked with the upregulation of pro-inflammatory cytokines.

Table 4.

Relative mRNA levels of MIP-2 and MCP-1in RAW264.1 cells exposed to the three types of SiO₂NPs

		Concentration (µg/cm2)	0	0.33		1.67		8.37
	Duration of exposure (h)
MIP-2	1	Vehicle	0.66 ± 0.02
(× 10–1)		NH2-SiO2NPs		0.49 ± 0.03	§	0.77 ± 0.09	§	9.46 ± 0.99*	§
		COOH-SiO2NPs		0.60 ± 0.17	§	0.67 ± 0.08	§	5.39 ± 1.02*
		OH-SiO2NPs		0.45 ± 0.13	§	0.70 ± 0.07	§	9.01 ± 2.02*	§
	4	Vehicle	0.31 ± 0.05
		NH2-SiO2NPs		0.28 ± 0.17	§¶	22.0 ± 36.7	§	57.1 ± 49.2	§
		COOH-SiO2NPs		0.46 ± 0.00	§	1.02 ± 0.29	§	35.0 ± 7.1*	§
		OH-SiO2NPs		0.08 ± 0.05	¶	0.85 ± 0.26	§	62.4 ± 7.7*	§
MCP-1	1	Vehicle	1.01 ± 0.16
		NH2-SiO2NPs		0.55 ± 0.05*	§	0.62 ± 0.05*	§	1.74 ± 0.11*	§
		COOH-SiO2NPs		0.51 ± 0.02*	§	0.65 ± 0.08*	§	1.29 ± 0.21	¶
		OH-SiO2NPs		0.93 ± 0.11	¶	0.84 ± 0.05	¶	2.18 ± 0.27*	§
	4	Vehicle	0.33 ± 0.07
		NH2-SiO2NPs		0.26 ± 0.01	§	1.89 ± 2.64	§	4.40 ± 3.59	§
		COOH-SiO2NPs		0.27 ± 0.01	§¶	0.35 ± 0.03	§	2.86 ± 0.91	§
		OH-SiO2NPs		0.33 ± 0.05	¶	0.35 ± 0.03	§	4.10 ± 0.45*	§

MIP-2 and MCP-1 mRNA levels relative to geometric mean of β-actin and GAPDH mRNA, and 18S rRNA level

Data are mean ± SD (n = 3, each). Asterisk (*) denotes significant difference with the vehicle control (by Dunnett’s multiple comparison following ANOVA). There is a significant difference between the different types of SiO₂NPs at the same concentration, which are not marked by the same symbols (§ or ¶) (by Tukey–Kramer multiple comparison following ANOVA). Significance level was set at 5%

Table 5.

Relative mRNA levels of IL-1β and TNF-α in RAW264.1 cells exposed to the three types of SiO₂NPs

		Concentration (µg/cm2)	0	0.33		1.67		8.37
	Duration of exposure (h)
IL-1β	1	Vehicle	2.35 ± 0.16
		NH2-SiO2NPs		1.08 ± 0.11	§	1.31 ± 0.25*	§	1.88 ± 0.32*	§
		COOH-SiO2NPs		0.90 ± 0.25*	§	1.27 ± 0.21*	§	1.57 ± 0.27*	§
		OH-SiO2NPs		2.07 ± 0.16	¶	1.12 ± 0.09	§	3.61 ± 0.76*	¶
	4	Vehicle	0.78 ± 0.21
		NH2-SiO2NPs		0.74 ± 0.05	§	2.06 ± 1.34	§	1.25 ± 0.69	§
		COOH-SiO2NPs		0.83 ± 0.08	§	0.82 ± 0.27	§	0.30 ± 0.11*	§
		OH-SiO2NPs		1.31 ± 0.11	¶	1.56 ± 0.44	§	0.92 ± 0.15	§
							§
TNF-α	1	Vehicle	1.14 ± 0.04
		NH2-SiO2NPs		1.00 ± 0.06	§	1.07 ± 0.09	§	1.57 ± 0.07*	§
		COOH-SiO2NPs		0.93 ± 0.10*	§	1.00 ± 0.04	§	1.35 ± 0.10*	¶
		OH-SiO2NPs		1.05 ± 0.08	§	0.98 ± 0.09*	§	2.02 ± 0.04*	†
	4	Vehicle	0.63 ± 0.05
		NH2-SiO2NPs		0.60 ± 0.01	§	1.09 ± 0.43	§	1.29 ± 0.38	§
		COOH-SiO2NPs		0.63 ± 0.06	§	0.75 ± 0.01	§	1.18 ± 0.15*	§
		OH-SiO2NPs		0.82 ± 0.04	¶	0.83 ± 0.10	§	1.28 ± 0.17*	§

IL-1β and TNF-α mRNA levels relative to geometric mean of β-actin and GAPDH mRNA, and 18S r RNA level

Data are mean ± SD (n = 3, each). Asterisk (*) denotes significant difference with the vehicle control (by Dunnett’s multiple comparison following ANOVA). There is a significant difference between the different types of SiO₂NPs at the same concentration, which are not marked by the same symbols (§, ¶ or †) (by Tukey–Kramer multiple comparison following ANOVA). Significance level was set at 5%

Discussion

NH₂- and COOH-functionalization of OH-SiO₂NPs reduced OH-SiO₂NPs-induced increase in lung weight

Exposure to OH-SiO₂NPs at 2 and 10 mg/kg bw induced the largest increase in lung weight compared to the other two SiO₂NPs. Increased lung weight can reflect pulmonary inflammation, which is associated with intense protein/fluid leakage from the vasculature into the alveoli mediated by inflammatory signals from pulmonary cells [35].

At the high dose, COOH-SiO₂NPs induced a greater infiltration of BALF neutrophils than NH₂- or OH-SiO₂NPs, whereas COOH or NH₂ functionalization attenuated OH-SiO₂NP-induced neutrophil infiltration at the low dose

With regard to BALF cells, our results showed that the type of the response to exposure was dose-and NP-functionalization-dependent. Thus, at 10 mg/kb bw, COOH-SiO₂NPs significantly increased BALF macrophages and neutrophils, NH₂-SiO₂NPs significantly increased BALF neutrophils, and in contrast, OH-SiO₂NPs significantly decreased BALF macrophages. On the other hand, at 2 mg/kg bw, OH-SiO₂NPs increased BALF neutrophiles, while COOH- and NH₂-SiO₂NPs did not. The response of BALF cells at 2 mg/kg bw might be linked with cytotoxicity of OH-SiO₂NPs. OH-SiO₂NPs-induced decrease in macrophages or the lack of increase in neutrophils at 10 mg/kg bw could possibly be due to cell death.

Surface modification of OH-SiO₂NPs reduces its cytotoxicity against RAW264.7 macrophage

Our in vitro studies demonstrated that OH-SiO₂NPs, but not the other two types of NPs, dose-dependently decreased the viability of RAW264.7 macrophages. Specifically, exposure to OH-SiO₂NPs at ≥ 3 mg/cm² for 4 h significantly decreased cell viability, while exposure to the same NPs at larger dose (≥ 7.5 mg/cm²) for 24 h significantly decreased the cell viability. This discrepancy in the lowest level for cytotoxicity between 4- and 24-h exposure may be explained by proliferation of the cells after the 4-h exposure. This argument is supported by the higher absorbance at 24 h than at 4 h (Suppl. Figure 6). The results suggest that surface modification of OH-SiO₂NPs reduced its cytotoxicity against murine macrophages, which might also explain the in vivo results that showed the uniqueness of OH-SiO₂NPs in dose-dependently reducing BALF macrophages and increasing lung weight. The present results are consistent with previous studies. Non-functionalized StÖber colloidal silica exhibited higher cytotoxicity toward RAW264.7 cells than amine-modified StÖber colloidal silica [36]. Similarly, Lankoff et al. reported that unmodified SiO₂NPs induced greater cytotoxicity and genotoxicity in human peripheral lymphocytes than aminopropyl- or vinyl-modified SiO₂NPs [24]. The suppression of particle-induced cytotoxicity through surface modification has also been observed for 1000 nm amorphous SiO₂ particles [37] and mesoporous silica SBA-15[38]. However, the opposite trend was reported for amino- or carboxyl- modified SBA-15-type silica synthesized via a co-condensation route using ethanol [39].

Internalization of SiO₂NPs in RAW264.7 macrophages does not explain OH-SiO₂NP toxicity

Our in vivo experiments showed a lower percentage of macrophages and neutrophiles internalized OH-SiO₂NPs compared with NH₂- and COOH-SiO₂NPs. This finding was consistent with our in vitro results, in which flow cytometric quantification revealed less internalization of OH-SiO₂NPs. Confocal microscopy of RAW264.8 macrophages also confirmed less cellular uptake of OH-SiO₂NPs ithan of NH₂- and COOH-SiO₂NPs. Taken together with the observation that OH-SiO₂NPs exhibited the highest cytotoxicity, these results suggest that internalization of OH-SiO₂NPs is unlikely to be the primary mechanism underlying OH-SiO₂NP-induced cytotoxicity. Similarly, in the study by Lankoff et al. [24], side scatter (SSC) signals in flowcytometry-an indicator of nanoparticle binding and uptake-were highest for aminopropyl-modified SiO₂NPs, followed by unmodified and then vinyl-modified SiO₂NPs [24]. This order differs from that of cytotoxicity, further supporting the lack of correlation between cellular uptake and toxicity of SiO₂NPs. However, internalization may be a prerequisite for cytotoxicity induced by micron-sized particles, as a previous study demonstrated that inhibitors of lysosomal function, such as CA-074-Me (a cathepsin B inhibitor) and bafilomycin A (a vacuolar-type ATPase inhibitor), suppressed the cytotoxicity of 1000 nm amorphous silica particles, indicating the involvement of lysosomes in their toxicity [25].

Does localization of OH-SiO₂NPs near cell membrane play a role in cytotoxicity?

Confocal microscopy of RAW264.7 macrophages revealed more distinct colocalization of OH-SiO₂NPs with the cell membrane compared with NH₂– or COOH-SiO₂NPs. This observation is consistent with a previous study showing that non-functionalized Stöber colloidal silica exhibits higher cytotoxicity and greater cellular association-measured as the mass of silicon per protein extracted from cells incubated with silica nanoparticles-than amine-modified Stöber silica [36]. These findings suggest an important contribution of physical contact between particles and the cell membrane to cytotoxicity. Recent studies have highlighted the role of surface silanol groups in the cytotoxicity of pyrogenic amorphous SiO₂NPs, including reports of cytolytic toxicity associated with silanol-rich particles[40], membranolysis driven by nearly free surface silanols [41], and the critical role of surface silanol content in determining cellular toxicity [14]. Our results extend this concept to colloidal amorphous silica, which is generally considered less hemolytic than pyrogenic silica [42], thereby underscoring the significance of surface silanol groups in mediating particle–cell membrane interactions and cytotoxicity.

Characterization of SiO₂NPs: greater hydrodynamic diameter of OH-SIO₂NPs indicates stronger interaction with water or hydrophilic molecules in medium

SEM–EDX analysis revealed a greater abundance of nitrogen (N) and carbon (C) elements in NH₂-SiO₂NPs, likely derived from the NH₂ functional group and the diethylenetriamine linker on the particle surface. The higher silicon (Si) content observed in COOH-SiO₂NPs may originate from the (triethoxysilyl)propylsuccinic anhydride linker. These findings confirm the successful NH₂- and COOH-functionalization of the particles. Interestingly, the zeta potential of NH₂-SiO₂NPs was negative, although NH₂ groups are expected to be protonated under neutral or acidic conditions. According to the manufacturer’s information, numerous acidic Si–OH groups remain on the surface of NH₂-SiO₂NPs, contributing to their overall negative zeta potential [43]. It is noteworthy that OH-SiO₂NPs exhibited a higher oxygen (O) content and a larger hydrodynamic diameter in both water and FBS-free medium compared with NH₂- or COOH-SiO₂NPs. These properties indicate stronger interactions of OH-SiO₂NPs with water or other hydrophilic molecules in the medium. In contrast, NH₂- and COOH-functionalization, through organic linkers, likely increases particle surface hydrophobicity, limiting their interaction with the medium and preventing an increase in hydrodynamic diameter.

OH-SiO₂NPs toxicity involves IL-1β and TNF-α upregulation

The higher expression levels of IL-1β and TNF-α after 1 h exposure to OH-SiO₂NPs or after 4 h exposure to OH-SiO₂NPs at non-cytotoxic level of 0.33 μg/cm² suggests the involvement of proinflammatory cytokines in the cytotoxicity of OH-SiO₂NPs.

Possible role of apoptotic/necroptotic pathways in OH-SiO₂NPs cytotoxicity

The use of each of inhibitors of apoptosis, necroptosis, pyroptosis and autophagy did not alter the cytotoxicity of OH-SiO2NPs. However, the combined use of apoptosis and necroptosis inhibitors reduced the cytotoxicity of OH-SiO₂NPs, suggesting the roles of these two pathways in OH- SiO₂NPs’ cytotoxicity. In this regard, DAPI staining did not show nuclear fragmentation of RAW264.7 murine macrophages exposed to OH-SiO₂NPs s at 8.37 µg /cm² (28 µg/mL) for 18 h, suggesting non-apoptotic cell death (Suppl. Figure 5). As numerous studies have shown, necroptosis can occur when apoptotic signaling is inhibited, including in the absence of caspase-8 function [44–47]. Our results may therefore indicate that the necroptosis inhibitor was effective when necroptotic signaling was activated by caspase inhibition. Moreover, recent studies have revealed cross-talk between apoptosis, necroptosis, and pyroptosis, leading to the concept of a combined cell death pathway termed PANoptosis[48, 49]. Investigation of the possible involvement of PANoptosis, however, is beyond the scope of the present study.

Rationale for NP exposure level used in the study

With regard to the relation between the exposure levels by oropharyngeal aspiration used in our study to those achieved by inhalation, Osier et al. [50] found deposition of 690 µg of titanium dioxide in the lungs of male rats weighing 175 to 225 g (probably 9-week-old) after 2-h exposure to ultrafine titanium dioxide at concentration of 125 mg/m³. Since the alveolar surface areas of 9-week-old rats and 8-week-old mice are 4000 and 530 cm², respectively [51, 52], the estimated lung deposition of titanium dioxide in mice would be 690 × 530/4000 = 91 µg following exposure at 125 mg/m³ for 2 h. Based on these estimations, the high dose of 200 µg used in our study is comparable to lung deposition after inhalation of ultrafine particles at concentration of 125 mg/m³ for 4.4 h.

Limitations of the study

This study has several limitations. First, the presence of a fluorescent label may influence surface charge, accessibility of functional groups, or induce masking effects, potentially altering the interactions of the particles with biological systems. Second, the characterization of SiO₂NPs was limited. SEM–EDX confirmed increased nitrogen content in NH₂-SiO₂NPs, consistent with NH₂ functionalization, and increased silicon content in COOH-SiO₂NPs, consistent with the COOH linker. However, the absolute degree of functionalization was not quantified. Attempts to evaluate NH₂ and COOH functionalization using Fourier-transform infrared spectroscopy (FTIR) with both attenuated total reflection (ATR) and KBr pellet methods failed to detect sufficient signals corresponding to the surface functional groups. Third, a greater hydrodynamic diameter of OH-SiO₂NPs indicates stronger interactions with water or other hydrophilic molecules, investigation of the protein corona formed around the particles would be meaningful; however, this was not examined in the present study.

Conclusions

The results of the present study demonstrate that surface modification of OH–SiO₂NPs with amino or carboxyl functional groups attenuated the OH-SiO₂NP-induced increase in lung weight in mice and reduced cytotoxicity in murine macrophages. Our findings further suggest that OH-SiO₂NP-induced cytotoxicity is not mediated by nanoparticle internalization but rather by apoptotic and necroptotic signaling triggered upon interaction of the nanoparticles with the cell membrane. These results highlight the potential for developing safer silica nanomaterials through amino- or carboxyl functionalization of surface silanols.

Author contributions

SV, SI, SB, LT and GI designed the study and interpreted the data. SV, EW, WW and SI conducted animal studies. EW, KY, KaMi, ST and YO conducted in vitro studies. CZ, TSa, AS, YH, AI, TW, KuMi, YT, KMa, TSu and RA contributed to acquisition of the data. SV, EW, SI and GI drafted and revised the manuscript. All authors read and approved the final manuscript.

Funding

This work was supported by Japan Society for the Promotion of Science (JSPS) International Fellowship for Research in Japan (Short-term Program) #PE13040, JSPS NEXT Program #LS056 and Japan Science and Technology Agency (JST) Strategic Japanese-EU Cooperative Program: Study on managing the potential health and environmental risks of engineered nanomaterials.

Availability of data and materials

All generated and analyzed data are included in this published article and its supplementary information material.

Declarations

Ethics approval

This study was conducted according to the Japanese law on the protection and control of animals and the Animal Experimental Guidelines of Tokyo University of Science. The experimental protocol was approved by the Animal Ethics Committee of Tokyo University of Science (approval number: Y14059).

Competing interests

The authors declare no competing interests.