Titanium Dioxide Nanoparticles Evoke Proinflammatory Response during Murine Norovirus Infection Despite Having Minimal Effects on Virus Replication

United States Food and Drug Administration, National Center for Toxicological Research, 1Division of Microbiology, 2Division of Genetic and Molecular Toxicology, Jefferson, Arkansas, 72079, USA, 3Center for Integrative Nanotechnology Sciences, University of Arkansas at Little Rock, Little Rock, Arkansas 72204 4Department of Medical Bioscience, College of Veterinary Medicine, University of Kirkuk, Kirkuk, Iraq.


INTRODUCTION
Nanomedicine has evolved into a discipline with high impact and potential in the fields of infectious disease therapeutics, vaccine delivery systems, and cancer therapeutics, among others (1)(2)(3)(4)(5). A variety of nanomaterials, including polymeric NPs, polymeric micelles, graphite, silver, gold, silica, carbon nanotubes and polylactide-coglycolides (PLGA) are being used as drug delivery vehicles or as adjuvants for vaccines (6). Several features of NPs, including size, shape, and their absorptive, catalytic, conductive and magnetic properties, make them attractive candidates for targeted delivery of drugs and biologics to humans (7). Although the foreign nature of NPs can elicit immune system stimulation, the convenient scaffold for engineering proteins and other biomaterials on the surface of the nanomaterial makes them attractive for use as vaccine platforms and as adjuvants to stimulate the immune response (8). In addition, there has been increased use of a variety of NPs in consumer products (9,10), with over 1500 products currently used in several sectors, including foods, beverages, cosmetics and home appliances (11,12).
TiO2 NPs are widely used in pharmaceuticals, cosmetics and food products (13,14). They are employed as a pigment to provide whiteness and opacity to paints, coatings, plastics, papers, inks, food products, pills, tablets and toothpastes (15). The use of TiO2 as food additive results in an estimated dietary consumption of 5.4 mg/person/day, and approximately 36% of TiO2 used in food products is nanosized (10,16). The use of TiO2 NPs in cosmetics, including lip balm and spray, allows for increased oral uptake, leading to increased exposure and accumulation in the GI tract (17,18). The levels of human exposure of TiO2 NPs warrant safety assessment on the risk potential in the GI tract and systemically. There have been mixed reports regarding the ability of TiO2 NPs to cross the intestinal wall and redistribute to various organs (19,20). However, Cho et al. (21) demonstrated that administration of high doses (1041.5 mg/kg) of TiO2 NPs for 13 weeks resulted in no free Ti in the bloodstream, suggesting a lack of intestinal permeability and localized accumulation of these NPs in the GI tract. Furthermore, several studies have reported toxicity of TiO2 NPs to in vitro human intestinal cells, including Caco-2, suggesting that accumulation of TiO2 NPs in the GI tract might result in toxicity (22,23).
Accumulation of TiO2 NPs in the GI tract can also result in interaction with foodborne viruses within the GI tract (24). While these interactions could affect the virus life cycle, they may also result in the exacerbation of viral replication, pathogenesis and host response. Among the enteric viruses, noroviruses (NoVs) are highly infectious and pathogenic, causing acute gastroenteritis in humans. The highly infectious nature of NoVs stems from their low infectious doses (as few as 10 viral particles can infect a host), their resistance to environmental stressors, and the absence of mechanisms in humans for long-term immunity to NoVs (25)(26)(27). NoVs cause roughly 80% of all human nonbacterial gastroenteritis infections worldwide (28,29). Within the US alone, NoVs cause ~19-21 million acute cases each year, resulting in 800 deaths and 71,000 hospitalizations, accounting for the majority of US foodborne illness episodes. In healthy individuals, NoV infection is generally mild and of short duration, but NoV infection can prove fatal in immunocompromised individuals. The World Health Organization (WHO) estimates that NoV infections in developing countries may result in up to 1.1 million hospitalizations and 218,000 deaths for immunocompromised children less than 5 years of age, with an attack rate of 56% (30)(31)(32)(33). NoV contamination of food sources, including seafood, poses serious potential threats to human health and challenges to the food industry.
Due to the fecal-oral transmission of NoVs and their tropism to the GI tract, it is probable that TiO2 NPs interact with these enteric pathogens. The effect of NoV-TiO2 NP interaction on the virus cycle and host cell biology has not been reported previously. Given the lack of an established cell culture model for human NoV, we utilized murine NoV (MNV-1) infection of RAW 264.7 murine macrophages as the model system for these studies. We tested the effect of anatase TiO2 NPs, on MNV infection, host response, and imparting toxicity to the macrophages. We present evidence that while pre-incubation with TiO2 NPs has inert effects on virus replication, they do not lead to genotoxic damage, but do, however, promote differential modulation of the host response and toxicity on the host cells.

Characterization of Nanoparticles
Nanoparticle diameters were measured through imaging using a JEOL JEM-2100F, Transmission Electron Microscope (TEM). The nanoparticles were dispersed in ethanol by sonication in an ice bath, before a few drops of the suspension were deposited on a holey carbon coated copper grid. After drying for ~10 minutes, the grid was inserted into TEM for observation at 80kV. The particle diameters were obtained from measuring over 130 particles in random fields of view. EMAN1 and ImageJ software were utilized to determine the major and minor diameters of each particle, and the geometric mean of the two is defined as the average diameter. Bruker D8 Discover XRD (X-ray Diffraction) system with Histar two dimensional detector and GADDS software was utilized to obtain the XRD spectrum of the anatase nanoparticles. X-ray source used in the analysis was Cu Kα line (λ=0.154nm) at 1600 Watts output power (40kV at 40mA). Bruker's Eva software wasused to analyze the spectrum.

Cell lines, TiO2 NP incubation and MNV infection.
RAW 264.7 cells were obtained from American Type Culture Collection (ATCC), grown in DMEM supplemented with 10% fetal bovine serum and 1% Penicillin/Streptomycin (GIBCO). Murine Norovirus genotype 1 (MNV-1, ATCC PTA-5935) was propagated in RAW 264.7 cells. The cells were treated with anatase TiO2 NPs, at concentrations of 20 µg/ml and 2 µg/ml for 3 h (where indicated) and then infected with MNV-1 at a multiplicity of infection (MOI) of 5 and 0.5 particles per cell (as indicated). Two different MOI (5 and 0.5 MOI), were chosen for the viral infection, as they are reflective of the range of differential amounts of virus exposure in the gut. After 24 and 48h post incubation, infectious virus was determined by plaque assay and genome copy numbers of MNV-1 were using Real Time RT-PCR. TNF-α, IL-6, IL-12, INF-γ and TGFβ1 levels were measured in the cellular supernatants collected as described below.

Viral RNA extraction and analysis of genome copy number.
Viral RNA was extracted from 115μl of cell culture supernatant using a MagMAXTM viral RNA isolation kit and the MagMAX™ Express Magnetic Particle Processor (Applied Biosystems/ Ambion, Austin, TX). Extracted RNA samples were either analyzed immediately by RT-PCR for genome copy number or stored at −80°C until use. RNA samples were analyzed for norovirus by a real-time one-step RT-PCR, using the Rotor-Gene Multiplex RT-PCR kit (Qiagen, Valencia, CA) according to the manufacturer's instructions. The reactions were carried out in duplicate, using primers and a Cy5-labeled fluorescent probe (Integrated DNA Technology, Coralville, IA). MNV RNA was used to generate a standard curve and 18S RNA (Applied Biosystems, Foster City, CA) was used as an internal control (16). RNA from mock-infected (no viral infection) RAW cell supernatant was used as a negative control and RNAse-free water was used as a nontemplate control. The detection limit of the assay was determined to be 2 RNA copies. For the assay, 5 μl of RNA was transferred to a Qiagen Rotor-Gene strip tube containing 10 µl of 2 × Rotor-Gene Multiplex RT-PCR master mix, 800 nM of each norovirus primer, MNVF (5'-TGCAAGCTCTACAACGAAGG-3') and MNVR (5'-CACAGAGGCCAATTGGTAAA-3'), 200 nM of probe MNVP (5'-CCT TCC CGA CCG ATG GCA TC-3'-Cy5), and 0.2 µl of Rotor-Gene RT mix. The thermal cycling conditions consisted of reverse transcription at 50°C for 30 min, 95°C for 10 min, and then 50 cycles of 94°C for 30 sec and 48°C for 60 sec. Amplification was detected using a Rotor Gene Q6-plex machine from Qiagen.
Viral titers were determined by plaque assay using RAW cells (34). Briefly, virus supernatants were serially diluted in MEM, and the dilutions were inoculated on a confluent monolayer of RAW macrophages and incubated at room temperature for 1 h. Viral inoculum was removed and cells were washed twice with PBS. A 1.5% agarose overlay in DMEM was added to the cells and incubated at 37ºC. Plaques were counted at 48-72 h post incubation. Virus titers were estimated by multiplying the mean of plaque numbers in at least two wells times the reciprocal of the lowest virus dilution with visible plaques and adjusted to 1 mL volume.

Trypan Blue Exclusion Assay.
Macrophages treated under indicated experimental conditions were harvested by scraping, and single suspensions were prepared in cold phosphate buffered saline (1x, pH 7.4, GIBCO). Trypan Blue (Ambion) 0.2% was diluted 1:2 in 1x PBS, and a 1:10 dilution of cell suspension was made using the diluted trypan blue solution. Viability was measured using Cellometer Auto T4 Manual (EMD Millipore, Germany).
Here, we analyzed DNA damage and cell viability in RAW cells following MNV-1 infection (at 5 MOI) alone, or in combination with exposure to 20 µg/ml of anatase

Statistical analyses.
All data were plotted and analyzed; statistical analysis was performed using Graph Pad Prism software, Microsoft Excel (Microsoft Corporation, Redmond, WA) and/or SigmaPlot (Systat Software, Inc., San Jose, CA). Two tailed T-tests were used to analyze statistical significance between the experimental groups.

Anatase TiO2 NPs decrease cell viability of murine macrophages at higher doses.
Prior to analyzing the effects of TiO2 toxicity and modulation of host response in murine macrophages in the context of MNV infection, we first examined whether the Anatase TiO2NPs alone exerted any effects on cell morphology and cell viability to the murine macrophages. We dosed the macrophages at two concentrations, 20 µg/ml and 2 µg/ml of Anatase NPs, and analyzed viability and morphology of RAW cells. Importantly, our dose range was based on our extrapolation of data indicative of the amount of human daily NP consumption. When RAW cells were incubated with 20 µg/ml of anatase NPs and monitored for 48 h, we did not observe any cytopathic effects (cell rounding, detachment and floating) compared to the mock-infected controls (Fig. 1A). Although there were no visible cytopathic effects, pre-incubation with 20 µg/ ml anatase caused a 30% reduction in cell viability, (Fig. 1B), as measured by a Trypan Blue exclusion assay. Pre-incubation with 2 µg/ml anatase did not alter morphology or viability of RAW macrophages as compared to mock controls (data not shown).

MNV Replication is altered in cells exposed to TiO2 NPs
We next analyzed whether pre incubation of RAW cells with high (20 µg/ml) or low dose (2 µg/ml) anatase NPs altered MNV replication and/or the production of infectious virus particles. The MNV-induced cytopathic effect observed at 48 h post infection was used a surrogate marker for viral replication, as MNV has been shown to cause cell rounding, shrinking and cell death following virus replication (37). At 24 h post infection, MNV-1 replicated to reach peak viral titers of 1x10 7 pfu/ml, when infected with 5 MOI, and ~6x10 6 pfu/ml, when infected with 0.5 MOI (Fig.  2A, left and right panels, MNV control). Replication of MNV-1 at 5 MOI was accompanied by rounding, shrinking and floating of cells in the monolayer, demonstrating the cytopathic effect due to MNV infection (Fig. 3A). These cytopathic effects resulted in a significant reduction in cell viability by 50% at 24 h and 75% at 48 h post infection, respectively, as measured by a Trypan Blue exclusion assay (Fig. 3C). Infection with 0.5 MOI MNV-1 had similar cytopathic effects (data not shown).

Viral copy numbers in the supernatant are altered upon exposure to different doses of NPs.
We next analyzed whether MNV-1 genome copy numbers in the cell culture supernatant are altered due to exposure of the RAW cells to NPs followed by MNV infection. While pre-incubation of cells with 20 µg/ml anatase followed by 5 MOI did not result in alterations of viral genome copy numbers (as compared to controls), pre-incubation with 2 µg/ml resulted in a non-significant reduction (non-significant P>0.05), in genome copy number at 24 h and 48 h post infection, respectively (Fig. S1A, left panel). Similarly, pre-incubation of cells with either 20 µg/ml or 2 µg/ ml of anatase followed by infection with 0.5 MOI did not affect viral genome copy numbers when compared to controls alone (Fig.  S1A, right panel). A.

Secretion of cytokines by RAW macrophages is modulated by TiO2 NPs.
Macrophages secrete several pro-inflammatory cytokines including Interleukin 6 (IL-6), Interferon Gamma (IFN γ), Tumor Necrosis Factor Alpha (TNFα), Interleukin 12 (IL-12), and Transforming Growth Factor Beta 1 (TGFβ1) in response to infection by microorganisms or antigenic stimulation. Since TiO2 NPs stimulate the immune system due to their foreign nature, we hypothesized that their presence during MNV-1 infection would result in a synergistic increase in host immune response and increased production of pro inflammatory cytokines. We therefore analyzed the secretion of pro-inflammatory cytokines during high and low MOI virus infection, in the presence of high and low doses of TiO2 NPs. Results from cytokine analysis are grouped in Fig. 4 (anatase 20 µg/mL and 2 µg/mL) under conditions of low and high MOI MNV-1 infection.

Assessment of DNA Damage
Literature studies have shown that NPs, including silica NPs, are capable of inducing single cell DNA damage consistent with genotoxicity that can be deciphered via the alkaline Comet assay (38). While viabilities ranged among cell sample preparations, there was a general reduction in viability as a result of either MNV-1 infection or NP exposure, consistent with known cytotoxic effects of these treatments. In contrast to cells exposed to Methyl Methane Sulfonate MMS (positive control), no statistically significant differences in DNA damage were observed for anatase TiO2 NP in the presence or absence of virus infection, when compared to the responses in the vehicle control group (Fig. 5).
The RAW macrophages were susceptible to single cell DNA damage, consistent with genotoxicity, after exposure to MMS, a known genotoxicant.

B.
A.

DISCUSSION
The use of TiO2 NPs in food additives and cosmetics results in accumulation of these NPs in the GI tract, which may allow interaction with enteroviruses. While several studies (16,19,20) have investigated the safety and toxicity of TiO2 NPs on intestinal cell lines, the direct effects of TiO2 NP accumulation on enteric virus replication, immune response of the host and genotoxic insult have not been studied before. Since noroviruses (NoVs) infect the GI tract by fecal-oral transmission and represent the leading cause of viral gastroenteritis (39,40), we investigated the effect of TiO2 NPs accumulation on NoV replication and immune response of the host. Because MNV is the only NoV cultivable in vitro, we utilized MNV-1 and RAW macrophages as the host model system to analyze the safety hazard impact of TiO2 NPs in the context of MNV-1 infection. Although previous studies (41) have analyzed the effects of TiO2 NPs on MNV replication, those studies involved pre-incubation of MNV with TiO2 NPs followed by analysis of virus infectivity in RAW macrophages. In contrast, our study is the first to analyze the effects of pre-incubation of RAW macrophages with different doses of anatase TiO2 NPs on MNV replication, host immune response and genotoxicity.
Before viral infection, we analyzed the effects of anatase TiO2 NPs on morphology and cell viability in RAW macrophages.
Although there were no distinct visual changes in cell viability, pre-incubation with 20 µg/ml of anatase resulted in ~30% decrease in viability. This effect of anatase was pronounced with high MOI MNV-1 infection, where there was a significant reduction in cell viability. Alkaline Comet assay analysis indicated there was no significant genotoxic insult imparted by 20 µg/ml of anatase alone or in combination with 5 MOI MNV, indicating that the observed cell death likely was mediated by mechanisms other than subtle DNA damage. Cytokine analysis showed an increase in proinflammatory cytokine IL-6 upon anatase incubation alone or in combination with MNV-1 infection. This elevation in IL-6 may have contributed to the reported decreases in cell viability, as cell death due to enhanced interleukin production stimulated by NPs in soil dust has been reported previously (42).
Pre-incubation of RAW cells with 20 µg/ml of the anatase form of TiO2 NPs had no effect on infectious virus production at 5 and 0.5 MOI of MNV-1 infection, whereas a tenfold reduction in dose (2 µg/ml) caused a significant reduction in infectious virus titers at 5 MOI but not 0.5 MOI (Fig. 2A). This was accompanied by a large increase (~6 fold) in IL-6 levels, further indicating that the increased expression of the pro-inflammatory cytokine could have contributed to the observed reduction in virus replication. Reduction of influenza virus and hepatitis B virus replication by increased IL-6 levels has been previously documented (43,44).
The effects of TiO2 NPs and viral infection on cytokine secretion, analyzed by ELISA, yielded several interesting observations. Pro-inflammatory cytokines, including IL-6, IFN-γ and TNFα, were increased by both doses of anatase during virus infection. Anatase NPs have been shown to induce the inflammatory response by upregulating cytokines, in many human cells (48,49).
Overall, anatase NPs evoked a pro-inflammatory response in the presence of MNV infection, as shown by the increase in indicated cytokines. Previous studies have documented the induction of proinflammatory response in macrophages by TiO2 NPs (48,50), but our study is the first to report the effect of NPs during MNV infection in a macrophage model. Pre-incubation of RAW macrophages with titanium NPs resulted in enhanced inflammation, which was exacerbated by virus infection. The increase in pro-inflammatory cytokines did not have a high impact on virus replication, although several of the cytokines, including IL6, IFN-γ, and TNFα, have an antiviral effect in many systems tested (51,52). Rather, this increase could facilitate the establishment of persistent or chronic virus infection, resulting in enhanced damage to the intestinal system. A recent report have shown an increase in inflammation, resulting in enhanced pneumonia, in mice inoculated with TiO2 NPs followed by infection with RSV virus (53).
Our rationale for using MNV-1 in this study was to confirm our in vitro findings using a mouse model, because MNV causes disease in mice. Studies investigating morbidity, virus replication and enhancement of disease upon pretreatment with TiO2 NPs are underway in mouse models to confirm our in vitro findings. Nevertheless, our study is the first to indicate that such phenomena could also occur in the GI tract, resulting in enhanced inflammation and damage due to interaction of NPs and enteric pathogens.
We used the alkaline Comet assay to investigate the genotoxicity of TiO2 NPs and of MNV infection. Although we used doses that were consistent with decreases in overall global viability suggestive of cytotoxic effects, even at the highest viral MOI or TiO2 NP exposure conditions, none of these groups induced DNA damage. These negative responses contrasted with the strong positive responses produced by MMS, consistent with previous reports using this chemical (38). The absence of a detectable response in the alkaline Comet assay in the presence of known global cytotoxic damage on the corresponding cellular suspensions was unexpected. The DNA damage may have occurred in a way that did not generate strand breaks detectable by the alkaline Comet assay, or perhaps DNA damage and repair took place outside of the time frame of our observations. To examine the likelihood of the latter possibility, we examined cells at a later time point after MNV infection (48 h). We again saw no difference in DNA damage between mock and MNV infection by the alkaline Comet assay. Our viability and global DNA damage data potentially suggest that large-scale DNA damage and apoptosis occur on such a scale that does not allow for retention of embeddable single cells and resulting genotoxic events. Although we did not detect genotoxicity in our current study, the effects of TiO2 NPs and MNV infection on cell viability, and thus likely on cytotoxicity, warrant further investigation of the safety hazards of TiO2 NPs.

Disclaimer
The findings and conclusions in this article are those of the authors and do not necessarily represent the official position of the US FDA.

Declaration of Interest
The authors declare no potential conflict of interest.

Author Contributions
SA, TT, and MA designed experiments, performed analysis and wrote the manuscript. LM performed the experiments. FW, TM and AB prepared and characterized NPs, while MM originally developed and optimized the Comet assay described in this study.