Characterization and absorption of SiO2 NPs
In this study, fumed SiO2 NPs (Fumed NPs) were obtained from Sigma Co. Ltd. The structure was amorphous (as indicated by the absence of peaks in the X-ray diffraction pattern), and 3.5–4.5 hydroxyl groups were present per square millimicron of SiO2 surface. As controls, stober SiO2 NPs (Stober NPs), fumed SiO2 fine-particles (Fumed FPs) and stober SiO2 fine-particles (Stober FPs) were also used in this study. Scanning electron microscopy (SEM), transmission electron microscopy (TEM), and dynamic light scattering (DLS) were used to characterize SiO2 NPs and FPs. The SEM and TEM images showed near-spherical shape and good dispersion of the SiO2 NPs and FPs. The average diameters of the Fumed NPs, Stober NPs, Fumed FPs, and Stober FPs were 35.76 ± 4.71 nm, 31.65 ± 2.33 nm, 227.63 ± 9.85 nm, and 213.37 ± 3.04 nm, respectively (Fig. 1a). The hydrodynamic sizes of the Fumed NPs, Stober NPs, Fumed FPs, and Stober FPs in PBS were 56.89 ± 4.81 nm, 51.16 ± 4.47 nm, 409.11 ± 10.37 nm, and 386.50 ± 7.99 nm, respectively (Fig. 1b). All zeta potentials were negative (Fig. 1c). The SEM and TEM results suggested that SiO2 NPs were well-dispersed. The level of endotoxin in all SiO2 NPs and FPs was below 0.01 EU/ml for all oral dose, these levels of contaminating endotoxin were determined to be inconsequential for these studies. To analyze absorption of SiO2 NPs at different concentrations, mice were orally administered Fumed NPs (0, 25, 50, 100, or 200 mg/kg bw). Mice were also treated simultaneously with 4-PBA (4-phenylbutyric acid) and NAC (N-acetyl-cysteine) in addition to 100 mg/kg bw Fumed NPs. Blood was collected at 0, 0.5, 1, 2, 4, 6, 8, 10, 12, and 24 h. Blood silicon levels increased and peaked at 1 h after oral administration (Fig. 1d). Then Fumed NPs were orally administered daily to mice. At the end of week 18, silicon levels were significantly increased in the livers, pancreases, spleens, kidneys, and small intestines in the groups treated with 25, 50, 100, and 200 mg/kg bw Fumed NPs compared to those in control mice (Fig. 1e). In addition, 4-PBA and NAC did not affect the absorption of SiO2 NPs in mice (Fig. 1d, e). To compare the absorption of different types and sizes of silicon NPs, mice were orally administrated Fumed NPs, Stober NPs, Fumed FPs, and Stober FPs (100 mg/kg bw each). Blood was collected at 0, 0.5, 1, 2, 4, 6, 8, 10, 12, and 24 h. Silicon levels increased, then peaked at 1 h after oral administration of Stober NPs, but did not change following oral administration of Fumed FPs and Stober FPs (Fig. 1f). Then, Fumed NPs, Stober NPs, Fumed FPs, and Stober FPs were orally administered daily to mice. At the end of week 18, silicon levels were significantly increased in the livers, pancreases, spleens, kidneys, and small intestines in the groups treated with 100 mg/kg of Fumed NPs and Stober NPs, and silicon levels were not significantly different between these two groups (Fig. 1g). However, silicon levels in the livers, pancreases, spleens, and kidneys of the Fumed FPs and Stober FPs groups were similar to those in the control group (Fig. 1 g). Scanning electron microscopy (SEM) and energy dispersive X-ray analysis (EDXA) showed that SiO2 NPs (nanosized spherical white objects), and not larger aggregates, were found in the livers, pancreases, spleens, kidneys, and small intestines of mice treated with 25, 50, 100, and 200 mg/kg bw Fumed NPs (Fig. 1h and Additional file 1 Figure S1a). In addition, scanning electron microscopy and EDXA showed that SiO2 NPs (nanosized spherical white objects), and not larger aggregates were found in the livers, pancreases, spleens, kidneys, and small intestines in the Fumed NPs and Stober NPs groups, but not in the livers, pancreases, spleens, and kidneys in the Fumed FPs or Stober FPs groups (Fig. 1i and Additional file 1 Figure S1b). The silicon levels in the small intestines in the Fumed FPs and Stober FPs groups were significantly increased, and SEM and EDXA showed that SiO2 FPs (fine-sized spherical white objects) were found in the small intestines in both groups. This observation may have been due to residual SiO2 FPs in the small intestines (Fig. 1g and Additional file 1 Figure S1b).
Effects of SiO2 NPs on blood glucose
Blood was collected from the tail veins of mice during the oral administration phase to measure blood glucose. At doses of 100 mg/kg bw and higher of Fumed NPs, blood glucose increased significantly starting at week 10. However, administration of 25 and 50 mg/kg bw of SiO2 NPs did not affect blood glucose (Fig. 2a). Insulin secretion was similar in each group in this study (Fig. 2b). Administration of 100 mg/kg bw of Stober NPs also increased blood glucose in mice starting at week 10, but did not affect insulin secretion (Fig. 3a and b). Administration of 100 mg/kg bw Fumed FPs or Stober FPs did not affect blood glucose or insulin secretion (Fig. 3a and b). The oral glucose tolerance test (OGTT) conducted at weeks 10 and 18 showed that the areas under the curves (AUC) resulting from administration of 100 and 200 mg/kg bw of Fumed NPs were significantly higher than those in response to administration of 25 and 50 mg/kg bw of Fumed NPs, which indicated that administration of 100 and 200 mg/kg bw of Fumed NPs resulted in reduced glucose tolerance (Fig. 2c, f). Insulin levels were similar in all groups (Fig. 2d, g). In addition, the insulin tolerance test (ITT), conducted at weeks 10 and 18, also showed that administration of 100 and 200 mg/kg bw of Fumed NPs resulted in reduced insulin sensitivity (Fig. 2e, h). The OGTT and ITT conducted at weeks 10 and 18 also showed that administration of 100 mg/kg of Stober NPs induced IR in mice, but Fumed FPs and Stober FPs did not (Fig. 3c-h). Terminal deoxynucleotidyl transferase-mediated dUDP nick end-labeling (TUNEL) results showed no apoptotic cells in the pancreases of the mice in any group at either 10 or 18 weeks (Fig. S2a). These results showed that mice that received 100 and 200 mg/kg bw of Fumed NPs exhibited insulin resistance (IR), resulting in increased blood glucose starting at week 10.
RNA sequencing and transcriptome characterization of mice following oral administration of SiO2 NPs
The livers of three mice exposed to 100 mg/kg bw of Fumed NPs for 10 and 18 weeks were analyzed using RNA-seq. The reading frames were mapped to the mouse genome (version NCBIM37/mm9). The transcriptomic analyses showed significant changes in mRNA abundance for 702 genes following oral administration of Fumed NPs, with 517 genes upregulated and 183 genes downregulated. In addition, one gene was upregulated after mice were exposed to Fumed NPs for 10 weeks, but then downregulated after 18 weeks of exposure. In contrast, one gene was downregulated after 10 weeks of exposure to SiO2 NPs, but then upregulated after 18 weeks of exposure (Fig. 4a and Additional file 2 Table S1). Enrichment analysis was conducted on the 517 upregulated genes and the 183 downregulated genes. The resulting agglomerative hierarchical clustering diagram contained 311 Gene Ontology (GO) terms, including 258 Biological Process (BP) terms, 28 Cellular Component (CC) terms, and 25 Molecular Function (MF) terms. These GO terms showed that SiO2 NPs affected the generation of reactive oxygen species (ROS) and other oxidants (GO-ID: 0055114, 16,491, 16,705, 4497, etc.), endoplasmic reticulum (ER) stress (GO-ID: 34976, 6986, 34,620, 30,968, etc.), and the inflammatory response (GO-ID: 0006954, 2526, etc.) (Fig. 4b, c and Additional file 3 Table S2).
SiO2 NPs increased blood glucose in mice from week 10
To analyze the relationship between the SiO2 NP-induced increase in blood glucose and ROS and ER stress, mice were orally administered SiO2 NPs, 4-PBA, and NAC. The results showed that blood glucose increased from week 10 in the 100 mg/kg bw SiO2 NP group (Fig. 5a). Insulin secretion was similar in each group (Fig. 5b). Glucose tolerance was tested at week 10, and SiO2 NPs did not affect insulin secretion in mice. However, SiO2 NPs impaired glucose tolerance, which was rescued by 4-PBA and NAC (Fig. 5c, d). Similarly, SiO2 NPs reduced insulin sensitivity in mice at week 10, as evidenced by the ITT (Fig. 5e). TUNEL results showed that SiO2 NPs induced cellular apoptosis, but did not affect the protein expression of cleaved-caspase 3 in liver cells of mice (Additional file 1 Figure S2b), but increased phosphorylation levels of IRS1 and reduced phosphorylation levels of Akt in liver cells (Fig. 5f, g). 4-phenylbutyric acid inhibits ER stress and NAC is a ROS scavenger. In this study, both 4-PBA and NAC effectively inhibited the effects of SiO2 NPs on blood glucose, glucose tolerance, insulin sensitivity, and phosphorylation of IRS1 and Akt (Fig. 5a-g). In addition, Stober NPs did not affect the protein expression of cleaved caspase 3, but altered the phosphorylation of IRS1 and Akt in liver cells of mice after 18 weeks of exposure (Additional file 1 Figure S2c, d). However, neither Fumed FPs nor Stober FPs affected the protein expression of cleaved caspase 3 or phosphorylation levels of IRS1 and Akt in mice.
SiO2 NPs increased ROS levels in mice starting at week eight
RNA-seq results showed that SiO2 NPs affected the generation of ROS in mice (Fig. 4b, c). RT-qPCR results showed that SiO2 NPs did not affect the mRNA expression of SOD1, SOD2, GSS, GCLC, or GCLM, which are genes that encode superoxide dismutase (SOD) and glutathione (GSH) (Fig. 6a). However, the levels of SOD and GSH were significantly reduced in sera and livers of mice starting at week eight (Fig. 6b, c). Furthermore levels of malonyl dialdehyde (MDA), a product of lipid peroxidation, were significantly increased in sera and livers of mice starting at week eight (Fig. 6d). These results suggested that SiO2 NPs increased ROS levels in mice starting at week eight. In addition, both 4-PBA and NAC effectively inhibited the effects of SiO2 NPs on ROS levels in sera and livers of mice (Fig. 6a-d). Nqo1, Nrf2, and HO-1 are genes in the Nrf2 pathways associated with ROS generation. In this study, RNA-seq results showed that SiO2 NPs affected the mRNA expression of Nqo1, Nrf2, and HO-1 (Fig. 6e). The effects of SiO2 NPs on the Nrf2 pathway were confirmed by RT-qPCR, and the results showed that SiO2 NPs increased the mRNA expressions of Nqo1, Nrf2, and HO-1 starting at week seven (Additional file 1 Figure S3a-d). The effects of Fumed FPs, Stober NPs, and Stober FPs on ROS production were also tested in this study. After 18 weeks of exposure, Stober NPs did not affect the mRNA expression of SOD1, SOD2, GSS, GCLC, or GCLM, but altered the levels of SOD, GSH, and MDA in sera and livers of mice (Fig. S3a-e). However, neither Fumed FPs nor Stober FPs affected the mRNA levels of T-SOD, GSH, or MDA in mice (Additional file 1 Figure S3e-i).
SiO2 NPs induced ER stress in mice from week six
RNA-seq results showed that SiO2 NPs affected genes related to ER stress in mice (Fig. 7a). RT-qPCR results showed that SiO2 NPs increased the mRNA expression of GRP78 and CHOP in the livers of mice starting at week six (Fig. 7b, c), and SiO2 NPs increased the ratio of sheared-XBP1/total-XBP1 (XBP1-s/t) in the livers of mice starting at week six (Fig. 7d). Cyp2b9 is a cytochrome P450 (CYP) enzyme involved in ER stress. In this study, SiO2 NPs increased the mRNA expression of Cyp2b9 in the livers of mice starting at week three (Fig. 7e). Furthermore, agarose gel electrophoresis results showed that SiO2 NPs increased the ratio of XBP1-s/t in the livers of mice at week six (Fig. 7f, g). Western blot results showed that SiO2 NPs increased the protein expression of phosphylated-eif2α, GRP78, CHOP, and ATF6, and increased the ratio of XBP1-s/t in the livers of mice at week six (Fig. 7h, i). These results showed that SiO2 NPs increased the expression of CYPs, then induced ER stress in mice starting at week six. In this study, 4-PBA significantly inhibited SiO2 NP-induced ER stress in mice. However, NAC did not affect SiO2 NP-induced ER stress in mice (Fig. 7b-i). After 18 weeks of exposure, Stober NPs increased the mRNA expression of GRP78, CHOP, XBP1-s/t, and Cyp2b9, and the protein expression of GRP78, CHOP, ATF6, and XBP1-s/t (Additional file 1Fig. S4a-c). However, Fumed FPs and Stober FPs did not induce the same changes.
SiO2 NPs activated the NF-κB and MAPK pathways in mice
RNA-seq results showed that SiO2 NPs affected genes related to inflammation (Fig. 8a). Real time qPCR data confirmed this result (Fig. 8b). Western blot results showed that SiO2 NPs activated the NF-κB pathway, through phosphorylation of p65-NF-κB and IκB (Fig. 8c, d). In addition, SiO2 NPs activated the MAPK pathway, which can result in IR, through phosphorylation of JNK and p38-MAPK (Fig. 8c, d). Furthermore, ELISA results showed that SiO2 NPs increased the levels TNF-α and IL-6 in sera of mice (Fig. 8e). In addition to 4-PBA- and NAC-mediated inhibition of ER stress and ROS generation, 4-PBA and NAC also inhibited SiO2 NP-induced inflammation and SiO2 NP-induced activation of the NF-κB and MAPK pathways (Fig. 8a-e). Furthermore, after 18 weeks of exposure, Stober NPs also activated the NF-κB and MAPK pathways, and also increased the levels of TNF-α and IL-6. In contrast, neither Fumed FPs nor Stober FPs induced these effects (Fig. S5a-d).