Vitamin Supplementation Protects against Nanomaterial-Induced Oxidative Stress and Inflammation Damages: A Meta-Analysis of In Vitro and In Vivo Studies

The extensive applications of nanomaterials have increased their toxicities to human health. As a commonly recommended health care product, vitamins have been reported to exert protective roles against nanomaterial-induced oxidative stress and inflammatory responses. However, there have been some controversial conclusions in regards to this field of research. This meta-analysis aimed to comprehensively evaluate the roles and mechanisms of vitamins for cells and animals exposed to nanomaterials. Nineteen studies (seven in vitro, eleven in vivo and one in both) were enrolled by searching PubMed, EMBASE, and Cochrane Library databases. STATA 15.0 software analysis showed vitamin E treatment could significantly decrease the levels of oxidants [reactive oxygen species (ROS), total oxidant status (TOS), malondialdehyde (MDA)], increase anti-oxidant glutathione peroxidase (GPx), suppress inflammatory mediators (tumor necrosis factor-α, interleukin-6, C-reactive protein, IgE), improve cytotoxicity (manifested by an increase in cell viability and a decrease in pro-apoptotic caspase-3 activity), and genotoxicity (represented by a reduction in the tail length). These results were less changed after subgroup analyses. Pooled analysis of in vitro studies indicated vitamin C increased cell viability and decreased ROS levels, but its anti-oxidant potential was not observed in the meta-analysis of in vivo studies. Vitamin A could decrease MDA, TOS and increase GPx, but its effects on these indicators were weaker than vitamin E. Also, the combination of vitamin A with vitamin E did not provide greater anti-oxidant effects than vitamin E alone. In summary, we suggest vitamin E alone supplementation may be a cost-effective option to prevent nanomaterial-induced injuries.


Introduction
With the rapid development of nanotechnology in the last decade, nanomaterials have been omnipresent in industrial products, medicines, food, and cosmetics due to their unique chemical and physical properties [1][2][3]. These extensive applications make nanomaterials unavoidably enter human bodies through respiratory inhalation, dermal penetration, oral ingestion, or injection routes, which subsequently results in toxic damages on various organs and tissues (e.g., lung, liver, kidney, spleen, heart, testis, et al.) and the pooled SMDs if a significant publication bias was detected (p < 0.05). The robustness of the pooled conclusions was assessed using a sensitivity analysis based on stepwise removing one study at a time.

Quality Assessment
The Toxrtool score was 16 for all studies, indicating that all in vitro studies were of high quality (Table 2). A low risk of bias was considered for the items of performance (random housing), attrition, reporting and other bias in most of animal studies except of the study of Yin et al. [26] (Table 2). Although an unclear risk of bias was assigned for most of the rest items because they were not sufficiently reported, this did not affect the determination of the indicators and the results. Thus, the quality of included in vivo studies was also acceptable.    (14) are the study endpoint(s) and their method(s) of determination clearly described?; (15) is the description of the study results for all endpoints investigated transparent and complete?; (16) are the statistical methods for data analysis given and applied in a transparent manner?; (17) is the study design chosen appropriate for obtaining the substance-specific data aimed at?; (18) are the quantitative study results reliable? Each item of Toxrtool obtained a score of 1 if the article satisfied the criteria; 1; otherwise, the score of 0 was given for the articles. SG, sequence generation; BC, baseline characteristics; AC, allocation concealment; RH, random housing; BI, blinding of investigators; ROA, random outcome assessment; BOA, blinding of outcome assessor; IOD, incomplete outcome data; SOR, selective outcome reporting.

Effects of Vitamin E Treatment on Cell Viability
The studies of Wang et al. [12], Yan et al. [31] and Faedmaleki et al. [32] provided the results of cell viability after treatment with six dosages of vitamin E. Cells were exposed to vitamin E for 24 h and 48 h in the study of Wang et al., [12]. Only a 24-h time period was designed for other three studies [27,31,32]. Thus, 25 items of data were used for this meta-analysis. The pooled results showed that vitamin E treatment could significantly improve the cell viability compared with the nanomaterial exposure group (SMD = 4.89; 95%CI, 3.65-6.14; p < 0.001; I 2 = 85.2%; p < 0.001) (Table 3; Figure 2). This significant effect was not changed in the subgroup analyses (Table 4).

Effects of Vitamin E Treatment on Cell Viability
The studies of Wang et al. [12], Yan et al. [31] and Faedmaleki et al. [32] provided the results of cell viability after treatment with six dosages of vitamin E. Cells were exposed to vitamin E for 24 h and 48 h in the study of Wang et al., [12]. Only a 24-h time period was designed for other three studies [27,31,32]. Thus, 25 items of data were used for this meta-analysis. The pooled results showed that vitamin E treatment could significantly improve the cell viability compared with the nanomaterial exposure group (SMD = 4.89; 95%CI, 3.65-6.14; p < 0.001; I 2 = 85.2%; p < 0.001) (Table 3; Figure 2). This significant effect was not changed in the subgroup analyses (Table 4).      [27]. Thus, nine data were used for this meta-analysis. The pooled results showed that vitamin E treatment could significantly decrease the caspase-3 activity compared with the nanomaterial exposure group (SMD = −2.07; 95%CI, (−3.25)-(−0.89); p = 0.001; I 2 = 80.7%; p < 0.001) ( Table 3; Figure 3). The anti-apoptotic ability of vitamin E was only significant after exposure for 48 h with a dosage of ≤ 1000 µM (Table 4).

Effects of Vitamin E Treatment on Oxidative Stress
The effects of vitamin E treatment on ROS (an indicator of oxidative stress) were evaluated in three articles [12,27,31] with ten data because four dosages (50, 500, 1000, 2000 µM) and two treatment durations were included in the study of Wang et al. [12]. The pooled results revealed that vitamin E treatment was associated with reduced ROS levels compared with the nanomaterial exposure group (SMD = −13.07; 95%CI, (−17.85)-(−8.30); p < 0.001; I 2 = 90.8%; p < 0.001) (Table 3; Figure 4). Subgroup analyses' results demonstrated that vitamin E treatment with a dosage of ≤1000 µM significantly inhibited the formation of ROS regardless of nanomaterials types and treatment durations (Table 4).

Effects of Vitamin E Treatment on Oxidative Stress
The effects of vitamin E treatment on ROS (an indicator of oxidative stress) were evaluated in three articles [12,27,31] with ten data because four dosages (50, 500, 1000, 2000 μM) and two treatment durations were included in the study of Wang et al. [12]. The pooled results revealed that vitamin E treatment was associated with reduced ROS levels compared with the nanomaterial exposure group (SMD = −13.07; 95%CI, (−17.85)-(−8.30); p < 0.001; I 2 = 90.8%; p < 0.001) (Table 3; Figure 4). Subgroup analyses' results demonstrated that vitamin E treatment with a dosage of ≤1000 μM significantly inhibited the formation of ROS regardless of nanomaterials types and treatment durations (Table 4).  Two studies [14,21] with four data (since three concentration gradients were designed for NiONPs in the study of Ahamed et al. [14]) investigated the effects of vitamin C treatment on the cell viability. The meta-analysis results demonstrated that vitamin C intervention could significantly increase the cell viability compared with the nanomaterial

Effects of Vitamin C Treatment on Cell Viability
Two studies [14,21] with four data (since three concentration gradients were designed for NiONPs in the study of Ahamed et al. [14]) investigated the effects of vitamin C treatment on the cell viability. The meta-analysis results demonstrated that vitamin C intervention could significantly increase the cell viability compared with the nanomaterial exposure group (SMD = 4.19; 95%CI, 2.37-6.01; p < 0.001; I 2 = 45.3%; p = 0.140) ( Table 3).

Effects of Vitamin C Treatment on Oxidative Stress
Four studies [13,14,21,30] with six data (due to the three concentrations included in the study of Ahamed et al. [14]) investigated the effects of vitamin C treatment on ROS levels.

Meta-Analysis for In Vivo studies 3.5.1. Effects of Vitamin E Treatment on Body Weight
Five studies [10,17,18,26,29] with six data (including two ZnONP concentrations included in the study of Baky et al. [17]) monitored the body weight of animals after administration of vitamin E. The meta-analysis results revealed no significant differences in the body weight between vitamin E and nanomaterial exposure groups (p = 0.328) ( Table 3). Although the subgroup analysis showed vitamin E could increase the body weight for animals exposed to AgNPs, only one study reported this effect (Table 4) and thus, the conclusion remained indefinite.

Effects of Vitamin E Treatment on Oxidative Stress
Six publications [8,10,11,18,25,28] with seven data (two GO concentrations included in the study of Shang et al. [25]) measured MDA levels; four articles [9,11,25,28] with six data (two concentrations included in the studies of Shang et al. [25] and AL-RASHEED et al. [9]) analyzed GSH levels; three studies [8,10,18] detected TOS, TAC, SOD and GPx; two studies examined OSI [10,18], CAT [8,10], SOD mRNA [10,18], GPx mRNA [10,18] and Nrf2 mRNA expression levels [18,28]. The summary analysis showed that the levels of pro-  (Table 3). No significant differences in the CAT activity, the mRNA expression levels of SOD and Nrf2 were present between two groups (p > 0.05) ( Table 3). Subgroup analyses demonstrated that only vitamin E treatment for more than two weeks (regardless of dosages) significantly decreased MDA and increased GSH. The improvement effects of vitamin E on GSH may be more sensitive for mice than rats, for ZnONPs, TiO 2 NPs and GNPs than GO (Table 4). posure group (Table 3). No significant differences in the CAT activity, the mRNA expression levels of SOD and Nrf2 were present between two groups (p > 0.05) ( Table 3). Subgroup analyses demonstrated that only vitamin E treatment for more than two weeks (regardless of dosages) significantly decreased MDA and increased GSH. The improvement effects of vitamin E on GSH may be more sensitive for mice than rats, for ZnONPs, TiO2NPs and GNPs than GO (Table 4).  [8,10,11,18,25,28]

Effects of Vitamin E Treatment on Inflammation
Five publications [8,9,17,28,29] with seven data measured TNF-α levels; four articles [9,17,28,29] with six data collected IL-6 levels; two studies [8,10,18] with four data assessed CRP levels. The number of real data for analysis of these three inflammatory indicators was larger than the number of articles because two concentrations included in the studies of Baky et al. [17] and AL-RASHEED et al. [9]. Two studies [25,27] with three data investigated the levels of IgE. The mRNA expression level of NF-κB was analyzed in the studies

Effects of Vitamin E Treatment on Inflammation
Five publications [8,9,17,28,29] with seven data measured TNF-α levels; four articles [9,17,28,29] with six data collected IL-6 levels; two studies [8,10,18] with four data assessed CRP levels. The number of real data for analysis of these three inflammatory indicators was larger than the number of articles because two concentrations included in the studies of Baky et al. [17] and AL-RASHEED et al. [9]. Two studies [25,27] with three data investigated the levels of IgE. The mRNA expression level of NF-κB was analyzed in the studies of Moradi et al. [8] and Azim et al. [28]. The summary analysis showed that except of NF-κB, the levels of all other pro-inflammatory indicators were lower in the vitamin E treatment group than those in the nanomaterial exposure group [TNF-α ( Subgroup analyses indicated that vitamin E treatment only inhibited the production of TNF-α at the early stage (treatment for less than two weeks), but IL-6 at both of the early (≤two weeks) and later (>two weeks) stages (Table 4).

Effects of Vitamin E Treatment on Apoptosis
The pooled analysis of four studies with six data [9,17,26,28] showed that compared with the nanomaterial exposure group, the caspase-3 activity was significantly decreased by vitamin E treatment (SMD = −7.10; 95%CI, (−10.49)-(−3.72); p < 0.001) (Table 3; Figure 8). Subgroup analyses showed that except of AgNPs, vitamin E treatment suppressed apoptosis induced by all other nanomaterials regardless of durations (Table 4). Subgroup analyses indicated that vitamin E treatment only inhibited the production of TNF-α at the early stage (treatment for less than two weeks), but IL-6 at both of the early (≤two weeks) and later (>two weeks) stages (Table 4).   Subgroup analyses indicated that vitamin E treatment only inhibited the production of TNF-α at the early stage (treatment for less than two weeks), but IL-6 at both of the early (≤two weeks) and later (>two weeks) stages (Table 4).

Effects of Vitamin E Treatment on Apoptosis
The pooled analysis of four studies with six data [9,17,26,28] showed that compared with the nanomaterial exposure group, the caspase-3 activity was significantly decreased by vitamin E treatment (SMD = −7.10; 95%CI, (−10.49)-(−3.72); p < 0.001) (Table 3; Figure  8). Subgroup analyses showed that except of AgNPs, vitamin E treatment suppressed apoptosis induced by all other nanomaterials regardless of durations (Table 4).

Effects of Vitamin E Treatment on DNA Damage
Comet assay was performed to evaluate DNA damage for three studies [9,17,28], after which the data about the tail DNA content and the tail length were obtained. The pooled analysis of these three studies with five data revealed a significant decrease in the tail length between two groups (SMD = −7.88; 95%CI, (−11.95)-(−3.81); p < 0.001) (Table 3; Figure 9). There was no significant difference in the tail DNA % (p = 0.283). The improvement effects of vitamin E treatment on the tail length remained significant after subgroup analyses stratified by nanomaterial types and animal model types (Table 4).

Effects of Vitamin E Treatment on DNA Damage
Comet assay was performed to evaluate DNA damage for three studies [9,17,28], after which the data about the tail DNA content and the tail length were obtained. The pooled analysis of these three studies with five data revealed a significant decrease in the tail length between two groups (SMD = −7.88; 95%CI, (−11.95)-(−3.81); p < 0.001) (Table 3; Figure 9). There was no significant difference in the tail DNA % (p = 0.283). The improvement effects of vitamin E treatment on the tail length remained significant after subgroup analyses stratified by nanomaterial types and animal model types (Table 4).

Effects of Vitamin E Treatment on Liver Function
Only liver function data (ALT, AST) could be combined for included studies [8,9,28] and thus, a meta-analysis was performed for them. The pooled analysis results showed that the level of ALT (SMD = −7.35; 95%CI, (−11.41)-(−3.29); p < 0.001) was significantly decreased by vitamin E treatment, but not the level of AST (Table 3).

Effects of Vitamin C Treatment on Oxidative Stress
MDA, SOD and CAT were analyzed in two studies to explore the anti-oxidative roles of vitamin C [10,19]. Unexpectedly, the pooled analysis did not detect significant differences in these three indicators between vitamin E and nanomaterial exposure groups (p > 0.05) ( Table 3)

Effects of Vitamin E Treatment on Liver Function
Only liver function data (ALT, AST) could be combined for included studies [8,9,28] and thus, a meta-analysis was performed for them. The pooled analysis results showed that the level of ALT (SMD = −7.35; 95%CI, (−11.41)-(−3.29); p < 0.001) was significantly decreased by vitamin E treatment, but not the level of AST (Table 3).

Effects of Vitamin C Treatment on Oxidative stress
MDA, SOD and CAT were analyzed in two studies to explore the anti-oxidative roles of vitamin C [10,19]. Unexpectedly, the pooled analysis did not detect significant differences in these three indicators between vitamin E and nanomaterial exposure groups (p > 0.05) ( Table 3).

Effects of Vitamin A Treatment on Body Weight
Meta-analysis of two studies [10,18] indicated vitamin A treatment could increase the body weight of animals relative to the nanomaterial exposure group (SMD = 2.1; 95%CI, 0.06-4.14; p = 0.043) ( Table 3).

Effects of Vitamin A Treatment on Oxidative Stress
Meta-analysis of three studies [8,10,18] Table 3). Meta-analysis of two studies [8,10] showed the activity of CAT was higher in the vitamin A treatment group relative to the nanomaterial exposure group (SMD = 3.22; 95%CI, 1.04-5.40; p = 0.004) ( Table 3).

Effects of Vitamin A + E Treatment on Oxidative Stress
Meta-analysis of two studies [8,18] showed that the level of MDA was reduced in the vitamin A + E treatment group compared with the nanomaterial exposure group (SMD =

Effects of Vitamin A Treatment on Body Weight
Meta-analysis of two studies [10,18] indicated vitamin A treatment could increase the body weight of animals relative to the nanomaterial exposure group (SMD = 2.1; 95%CI, 0.06-4.14; p = 0.043) ( Table 3).

Effects of Vitamin A Treatment on Oxidative Stress
Meta-analysis of three studies [8,10,18] Table 3). Meta-analysis of two studies [8,10] showed the activity of CAT was higher in the vitamin A treatment group relative to the nanomaterial exposure group (SMD = 3.22; 95%CI, 1.04-5.40; p = 0.004) ( Table 3).

Effects of Vitamin A + E Treatment on Oxidative Stress
Meta-analysis of two studies [8,18] showed that the level of MDA was reduced in the vitamin A + E treatment group compared with the nanomaterial exposure group (SMD = −8.42; 95%CI, (−11.17)-(−5.67); p = 0.013) ( Table 3). TOS, TAC, SOD and GPx were not significantly changed (Table 3).

Publication Bias and Sensitivity Analysis
Egger's test showed a publication bias existed in the analysis of several indicators with at least three data analyzed (except of the body weight, p = 0.081; TNF-α, p = 0.251 for in vivo studies with vitamin E treatment) (Table 3). However, significant results were still present for most of indicators [except of SOD (p = 0.435) and GSH (p = 0.198) in in vivo studies with vitamin E treatment, which were no longer significant after being adjusted by the trim and fill method]. Sensitivity analysis results also showed that no individual study affected the synthesized results ( Figure 10).
Egger's test showed a publication bias existed in the analysis of several indicators with at least three data analyzed (except of the body weight, p = 0.081; TNF-α, p = 0.251 for in vivo studies with vitamin E treatment) (Table 3). However, significant results were still present for most of indicators [except of SOD (p = 0.435) and GSH (p = 0.198) in in vivo studies with vitamin E treatment, which were no longer significant after being adjusted by the trim and fill method]. Sensitivity analysis results also showed that no individual study affected the synthesized results ( Figure 10). Figure 10. Sensitivity analysis for MDA levels of murine models treated with vitamin E. MDA, malonaldehyde; CI, confidence interval. [8,10,11,18,25,28]

Discussion
Although there had meta-analyses to demonstrate that vitamins can exert anti-oxidant and anti-inflammatory activities [15,16,34], no studies investigated their protective roles for nanomaterial-induced injuries until now. In addition, some meta-analysis results found the anti-oxidant and anti-inflammatory functions of vitamins were limited and even indicated vitamins exhibited potential toxic activities [35]. Thus, to prevent the hazard events induced by nanomaterials, but not cause the abuse of health care products, this study included 19 articles and performed a meta-analysis to comprehensively evaluate the roles and mechanisms of vitamins for cells and animals exposed to nanomaterials. Our meta-analysis results showed that vitamin E could antagonize nanomaterial-induced oxidative stress (mainly by reducing ROS, TOS, TAC, OSI, and MDA and increasing GPx), inflammation (significantly reducing the effects on TNF-α, IL-6, CRP, and IgE), improving cytotoxicity (manifested by an increase in the cell viability and a decrease in pro-apoptotic factor caspase-3) and genotoxicity (represented by a reduction in the tail length), which were less changed by subgroup stratifications. Pooled analysis of in vitro studies indicated that vitamin C treatment increased the cell viability and decreased ROS levels, but its antioxidant potential was not observed in the meta-analysis of in vivo studies. Vitamin A

Discussion
Although there had meta-analyses to demonstrate that vitamins can exert anti-oxidant and anti-inflammatory activities [15,16,34], no studies investigated their protective roles for nanomaterial-induced injuries until now. In addition, some meta-analysis results found the anti-oxidant and anti-inflammatory functions of vitamins were limited and even indicated vitamins exhibited potential toxic activities [35]. Thus, to prevent the hazard events induced by nanomaterials, but not cause the abuse of health care products, this study included 19 articles and performed a meta-analysis to comprehensively evaluate the roles and mechanisms of vitamins for cells and animals exposed to nanomaterials. Our meta-analysis results showed that vitamin E could antagonize nanomaterial-induced oxidative stress (mainly by reducing ROS, TOS, TAC, OSI, and MDA and increasing GPx), inflammation (significantly reducing the effects on TNF-α, IL-6, CRP, and IgE), improving cytotoxicity (manifested by an increase in the cell viability and a decrease in pro-apoptotic factor caspase-3) and genotoxicity (represented by a reduction in the tail length), which were less changed by subgroup stratifications. Pooled analysis of in vitro studies indicated that vitamin C treatment increased the cell viability and decreased ROS levels, but its anti-oxidant potential was not observed in the meta-analysis of in vivo studies. Vitamin A treatment was shown to decrease MDA (SMD: −3.17 vs. −6.37), TOS (SMD: −1.34 vs. −5.89) and increase GPx (SMD: 2.73 vs. 3.99), but its effects on these indicators seemed to be weaker than vitamin E. Also, the combination of vitamin A with vitamin E seemed not to provide greater anti-oxidant effects than vitamin E (except of MDA that was further reduced by two-fold). Accordingly, we may consider that vitamin E alone supplementation may be more cost-effective to prevent nanomaterial-induced injuries and diseases, especially for populations with occupational exposure.
Based on our results, the preventive roles of vitamin E against nanomaterial-induced injuries (apoptosis and DNA damage) were speculated to be exerted mainly through the following mechanisms: (1) as a fat-soluble vitamin, vitamin E can penetrate the lipid bilayer of the cell membrane and interact with phospholipids to stabilize bilayer structures and decrease the permeability of bilayer membranes [36], which ultimately inhibits the entrance of toxic nanomaterials into human cells [37,38]; (2) vitamin E not only quenches nanomaterial (if they are accidentally permeated to cells)-induced ROS in cell membranes [6,12,31], but also reacts with a lipid hydroperoxyl radical (LOO • ) by donating hydrogen from its phenolic hydroxyl group at the C-6 position, resulting in the formation of lipid hydroperoxide which was subsequently transformed to non-toxic hydroxide after catalysis by GPx to terminate lipid peroxidation and decrease the levels of the end products of lipid peroxidation (MDA) [39]; (3) previous studies demonstrated that ROS induced an inflammatory response via activation of the mitogen-activated protein kinase-NF-κB signaling pathway [40,41]. Thus, the anti-inflammatory functions of vitamin E may indirectly result from its suppressive effects on oxidative stress. Furthermore, vitamin E was found to directly stimulate the production of cyclic adenosine monophosphate (cAMP) in human peripheral mononuclear cells via the EP2/EP4 receptors and adenylyl cyclase [42], which in turn activated its downstream proteins (protein kinase A and cAMP response element binding) and then suppressed the release of pro-inflammatory cytokines (such as TNF-α and IL-6) from monocytes [43]. Importantly, there was evidence to demonstrate that up-regulation of TNF-α triggered the production of IL-6 [44], while IL-6 stimulated the transcription and synthesis of CRP [45] and IgE [46]. This may be an underlying reason to explain that vitamin E suppressed the levels of TNF-α at the early stage and then IL-6 for a long time as reported in our subgroup analyses.
There are some limitations in this meta-analysis. First, the number of included in vivo and in vitro studies was still limited and the detected indicators were varying in studies, which led to less and no data pooled (such as the anti-inflammatory roles of vitamin C and A; damages on the renal, spleen, heart and brain tissues; the other vitamin types). Second, considerable heterogeneity was present among studies for the analysis of several indicators and the source of heterogeneity could not be removed by the subgroup analysis. Therefore, it is necessary to conduct more experiments on cells, animals, and humans to confirm the conclusions of our study.

Conclusions
This meta-analysis of 19 in vitro and in vivo studies provides evidence that supplementation with vitamins (especially vitamin E) may be beneficial to prevent nanomaterialinduced cytotoxicity and genotoxicity by exerting anti-oxidative and anti-inflammatory activities. Our findings support the clinical recommendations of vitamin E intake for workers with occupational exposure to nanomaterials. However, our conclusions are still needed to be confirmed by analysis of more studies of high-quality and lack of heterogeneity.

Conflicts of Interest:
All authors declare that they have no competing interest. Tong Wu is an employee of Shanghai Jing Rui Yang Industrial Co., Ltd. Wei Xu is an employee of Shanghai Nutriwoods Bio-Technology Co., Ltd. Qingyang Meng is an employee of Shanghai Pechoin Daily Chemical Co., Ltd. Also, no any hidden conflict of interest exists by these companies.