Titanium Dioxide Nanoparticles Mitigated 2,4,6-Trinitrobenzenesulfonic (TNBS) Acid Solution Induced Colitis By Downregulating NF-κB Related Transcription


 BackgroundTitanium dioxide (TiO2) with nanofractions is increasingly applied in food products as a food additive, which makes consumers under the health risks of titanium dioxide nanoparticles (TiO2-NPs) oral exposure. The recent ban of food additive TiO2 (E171) use in France aggravated public controversy on safety of orally ingesting TiO2-NPs. This work aimed to determine biological effects of TiO2-NPs (38.3 ± 9.3) oral consumption (100 mg/kg bw, 10 days) on TNBS-induced colitis mice and healthy mice, and the additional vitamin E administration was also conducted to explore the possible mechanism of TiO2-NPs on colitis development.ResultsOral consumption of TiO2-NPs exacerbated oxidative stress status in colitis mice by decreasing the colonic glutathione (GSH) and total glutathione (T-GSH) levels, however, TiO2-NPs administration repaired the dysbacteriosis of colitis mice, and downregulated the Toll-like receptors (TLRs), nuclear factor kappa-B (NF-κB) signal pathway and inflammatory factor (IL-1β and TNF-α) transcription levels in colon tissue, which finally decreased the TNF-α expression level and participated in the mitigation of colitis symptoms. Moreover, further vitamin E intervention after TiO2-NPs consumption could relieve the oxidative stress status (mainly by scavenging reactive oxygen species, ROS) and the inflammatory factor over-transcription in colonic epithelium of colitis mice, but the effect of TiO2-NPs on dysbacteriosis repair would not be further changed by Vitamin E. At last, TiO2-NPs induced oxidative stress status and increased NF-κB signal transcription level in colonic epithelium, which increased daily disease activity index (DAI) score and caused mild mucosal inflammatory cell infiltrate in healthy mice. ConclusionOur present work showed that oral TiO2-NPs administration indeed induced oxidative stress and made an adverse effect on the development of colitis, but TiO2-NPs could also downregulate the NF-κB signal transduction level by repairing gut dysbacteriosis, which made a predominant role in alleviating colitis. On the other hand, it should also be noticed that TiO2-NPs oral ingestion caused potential colonic inflammation risks in healthy mice.


Background
The in ammatory bowel disease (IBD), which includes ulcerative colitis (UC) and Crohn's disease (CD), has been considered as an expanding global health challenge, with the western lifestyle largely contributing to their pathogenesis [1]. Currently, the aetiology of IBD still remains poorly understood, but it is generally accepted that the interaction of genetic and environmental factors play a decisive role in the pathogenesis of IBDs [2], and among the environmental factors are the increased 'Western diet', which include the ingestion of dietary TiO 2 particles [3,4]. In the progression of IBD, the oxidative stress and impaired mucosal immunity are considered as main pathomechanisms and have a long-standing implication [5]. Based on the integration of a series of clinical studies, patients with IBD and colitisassociated colorectal cancer generally got extremely heavy ROS load in the intestinal mucosa, the severe oxidative stress would further activate several key signal pathways, including NF-κB and mitogenactivated protein kinase (MAPK) signaling pathways to increase release of cytokines including TNF-α, IL-1β, IL-6 and IL-8; moreover, some pro-in ammatory cytokines like TNF-α and IL-6 would form feedback loops in NF-κB signaling to augment the in ammatory response and tissue damage [6].
TiO 2 is one of the most promising engineered materials, which has been extensively used for whitening and brightening in foods, medicine and personal care products [7,8,9]. Food and pharma grade TiO 2 are usually found both in the bulk form and as nanoparticle [3]. A test of food-grade TiO 2 showed that approximately 10-36% of the particles used in food consist were TiO 2 -NPs (with diameter less than 100 nm), and the proportion of NPs could be over 43% for some sweets such as the sugar-coated chewing gum [9,10]. Indeed, the daily intake of TiO 2 was estimated to be 1.28 mg/kg bodyweight, especially the human dietary exposure dose of TiO 2 -NPs has reached 2.16 to 100 µg/kg bodyweight per day, and children or those with a sweet tooth were identi ed as having several times higher exposure values [3,11,12]. Notably, these particles could be released and swallowed by humans upon chewing, thus consumers have actually come into direct contact with TiO 2 -NPs through food products.
TiO 2 was traditionally considered as an inert and nontoxic substance for its insolubility, low absorption, and absence of toxic effects by the oral route, but recently the nanosized TiO 2 has raised wide concerns about its potential health risk on gastrointestinal tract because of the small size, large surface area and high reactivity [13]. In in vivo studies, our previous works have clari ed that TiO 2 -NPs ingestion could cause oxidative stress and induce liver, kidney as well as heart damage in young rats [14,15], affect the amino acid absorption and metabolism of adult rats [16]. Nogueira et al. [17] found that the ingested TiO 2 -NPs were detected to form dark aggregates in M-cells of Peyer's patches and in underlying macrophages of mice intestinal tract, which could further in uence the in ammatory cytokine production and induce in ammatory responses. Latterly Ruiz et al. [3] also reported that TiO 2 -NPs could active the NOD-, LRR-and pyrin domain-containing protein 3 (NLRP3) in ammasome in disuccinimidyl suberate (DSS)-induced colitis mice and worsen colonic in ammation. Except for the risk on in ammation inducement, TiO 2 -NPs was also reported to disturb the structure and function of gut microbiota as its unique antimicrobial compound, which was also suspected to in uence the host health [18,19]. On the other side, extensive in vitro evidence have shown TiO 2 -NPs could effectively induce oxidative stress in intestinal cell culture systems including Caco-2, HT29, DLD-1, SW480 and NCM460 [20,21,22], and a limited number of in vivo studies also reported that TiO 2 -NPs oral exposure could promote the cytokine levels such as IL-1β, IL-8, IL-12, TNF-α and IFN-γ in colonic mucosa of animals [17,23]. However, no in vivo study has directly focused on the relationship between the oxidative stress induced by TiO 2 -NPs oral exposure and the progression of IBD, nor has there been any attempt to explore other mechanisms such as signal pathway activation and gut microbiota impairment during this process.
In the present study, we analyzed the effect of TiO 2 -NPs on TNBS-induced acute colitis as well as the contribution of oxidative stress and NF-κB signal transcription during this process. For this purpose, we attempted to create 2,4,6-Trinitrobenzenesulfonic acid solution acute colitis model in the wild-type (WT) BALB/c mice, and TiO 2 -NPs were orally administered to the colitis mice to study its effect on colitis symptoms. Meanwhile, we carefully detected the oxidative stress status, colonic micro ecology and NF-κB signal transcription changes of mice to unravel biological mechanisms of NPs affecting colitis. Considering that antioxidant vitamin E could eliminate the ROS induced by TiO 2 -NPs, some colitis model mice received additional vitamin E administration for further argument on the role of oxidative stress during the process TiO 2 -NPs changed the disease.

TiO 2 nanoparticles and physicochemical characterization
In the present study, the food grade TiO 2 nanoparticles were obtained from Shanghai Yunfu Nanotechnology Co. Ltd, China. A detailed report of TiO 2 -NPs physicochemical characterization was included in our early publications [26,65]. TiO 2 -NPs were anatase form with a purity of 99.95%. The transmission electron microscopy (TEM, JEM-2100F, JEOL, Japan) showed that TiO 2 -NPs were nearly spherical shape, the primary size of nanoparticles ranged from 15.5 nm to 60.0 nm, and the average primary diameter was measured as 38.3 ± 9.3 nm ( Figure 10A and 10B). The SSA of TiO 2 -NPs was measured as 61.87 m 2 /g according to the Brunauer-Emmett-Teller (BET) method (ASAP 2020, Micromeritics, Norcross, US). The hydrodynamic diameter, polydispersity index values (PDI) and Zeta potential of TiO 2 -NPs in ultrapure water (H 2 O), arti cial gastric juice (AGJ) and arti cial intestinal juice (AIJ) were tested by the ZetaSizer Nano ZS90 (Malvern Instruments Ltd, Malvern, UK). As shown in Figure  10C, TiO 2 -NPs got the largest distribution (range from 122.4 nm to 3579.6 nm) after dispersed in H 2 O, but the largest average hydrodynamic diameter was found when TiO 2 -NPs were dispersed in AGJ. Moreover, TiO 2 -NPs appeared the best dispersion stability which represented by largest absolute value of Zeta potential after their dispersion in AIJ.

Animal treatment
The WT BALB/c mice (8-week-old) were supplied and bred by the Department of Laboratory Animal Science, Peking University Health Science Center. The mice were maintained under speci c pathogen-free (SPF), constant temperature (20 ± 2 °C) and humidity (50-70 %), and a 12-hour light-dark cycle condition.
All experiments were carried out in accordance with the Guiding Principles in the Use of Animals in Toxicology adopted by Society of Toxicology and the European Union Directive 2010/63/EU for animal experiments, and approved by the Institutional Animal Care and Use Committee in Peking University (approval number: LA2019216).
The TNBS solutions used for presensitization and colitis model establishment were set up based on the published studies [66]. 54 mice were randomly divided into 6 groups according to body weight (bw) after 5-day accommodation, 36 mice in 4 groups was treated with TNBS to induce acute colitis model, and the detailed report of the method used in the establishment of the mouse colitis model was included in the earlier publication by the author [65]. Brie y, after the 7-day administration of TNBS presensitization solution to the shaved back skin of mice, we conducted instillation of TNBS solutions in the colorectal tract of mice after they were anesthetized by 4% bacteria-free chloral hydrate solution (i.p., Sinopharm Chemical Reagent Co., Ltd. Beijing, China). All of mice were maintained in an ideal anesthetic state at 37 ℃ during colitis establishment, and held in a trendelenburg position to avoid solution back ow after the instillation. The timeline of the experimental protocol was shown in Figure 1E, healthy and colitis mice were separately intragastrically administrated with ultrapure water or TiO 2 -NPs dispersion (100 mg/kg bw/day, newly prepared before gavage), once a day for 10 consecutive days. After the TNBS enema, two of the colitis model groups were additionally orally treated with vitamin E dispersion (30 mg/kg bw/day), once a day and lasted for 3 days. Feces samples used for subsequent microbiome analyses were collected late in Day 10. All of mice were harvested at Day 11, serum and organs including colon, intestine, liver, kidney, spleen and mesenteric lymph nodes were collected for further analyses.

General condition assessment
In the present study, the mice symptom, mortality and body weight change of mice during treatment were carefully observed and recorded daily. The residual and supplement feed were weighed daily to calculate the food intake, and the coe cient of feeding e ciency (CFE, %) was calculated using the following equation: CFE = (total body weight gain (g) / total food intake (g)) ×100% After mice were sacri ced at Day 11, the colon, intestine, liver, kidney and spleen of animals were collected and washed in cold phosphate buffered saline (PBS, 4℃), sucked dry with normal lters, and then weighted. The organ weight coe cient were calculated using the following equation: The organ weight coe cient = organ weight (wet, g) /body weight (g) ×100% For pathological assessment, three mice were selected from every group based on the random principle and used, and all the histological examinations were performed following the standard laboratory procedure. Brie y, The excised colon and other tissues (including Jejunum, liver, kidney, spleen and mesenteric lymph nodes) were xed in 4% buffered paraformaldehyde and embedded in para n blocks, then samples were sliced into 5 μm thickness and placed onto glass slides. After slices were hematoxylin-eosin (HE) stained, the optical microscope (Olympus BX50, Moticam 2306, Japan) was used to observe slides and take photos. The identity and analysis of the pathology slides were blinded to the pathologist. Especially for the Jejunum tissue, the villi height and the crypt depth were measured based on our previous study [26] and the ratio of the villi height to the crypt depth was also calculated.

Colitis activity assessment
The colitis activity was evaluated by using the daily disease activity index (DAI) score, colon mucosal damage index (CMDI) score, and colon histopathological score according to the methods described in our earlier publication [65]. The DAI score was calculated basing on scores for weight reduction and faecal characterization (including consistency and hemoccult) [65]. For CMDI assessment, the colons of mice were opened by a midline incision and rinsed with cold PBS(4℃). After the length of colon were measured, the structural damages of bowel wall were observed from luminal side under a magni er. Meanwhile, the CMDI score was also assessed based on the colon traits and ulcer numbers according to the reported method [67]. And for histopathological score assessment, 3 intact colonic mucosa was randomly chosen from each slide. After the mucosal height was carefully measured, the histopathological damage score were assessed as absence of goblet cell and crypt as well as the degree of in ammatory cell in ltration according to the earlier publication by the author [65].

In ammation-related protein transcription and cytokines expression level measurement
The mRNA transcription levels of TLRs (TLR-2 and TLR-4), NF-κB signal pathway (IKK-1, IKK-2 and p65) and in ammatory factors were measured in colon tissue (6 samples per group). About 50 mg of colon tissue was homogenized and total RNA was extracted using Trizol (Hermo Fisher Scienti c Inc, MA, USA). Concentrations of total RNA were determined with a spectrophotometer. The quality of isolated RNA was assessed by calculating the A260/ A280 ratio to ensure values between 1.8 and 2.0. SuperScript III Reverse transcriptase kit (Vazyme Biotech Co.,Ltd, Nanjing, China) was further used for the complementary DNA (cDNA) synthesis, the collected cDNA was quanti ed by quantitative real time polymerase chain reaction (PCR, Bio-Rad iCycler iQ5, Bio-Rad Co.,Ltd, the United states). The mRNA expression levels were normalized to Glyceraldehyde 3-phosphate dehydrogenase (GAPDH). Primers' sequences used in the present study were shown in the supplement information (Table S3).
The concentrations of TNF-α and IL-12 in serum and colon tissues (6 samples per group) were analyzed using ELISA Kit (Bioleaf Biotech Co.,Ltd, Shanghai, China) following the manufacturer's instructions. Brie y, colon tissues were homogenized in PBS buffer containing the protease inhibitors, then the supernatant was obtained after centrifugation at 3000 rpm for 10 min at 4℃. The tissue supernatants and serum samples were added into 96-well assay plates and incubated for 60 min at 37°C. After further incubation with horseradish peroxidase(HRP)-Avidin at 37°C, plates were detected on the microplate reader(FLUOstar Omega, BMG LABTACH, German) with the wavelength of 450 nm. The calibration curve for each cytokine was generated with a ve-parameter logistic t. The detection limits for the cytokine assays were 1 pg/mL. In order to determine the ROS production in colon tissue, we obtained single-cell suspensions of colon epithelial cells through a 48nm nylon mesh (6 samples per group). ROS kits (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) were used to measure ROS levels. According to the manufacturer's instructions, Dichloro-dihydro-uorescein diacetate (DCFH-DA; 500 nM) was added to the suspensions and incubated for 60 min at 37°C. Then cell suspensions were washed twice and resuspend with PBS. ROS-induced uorescence was detected using a microplate reader (FLUOstar Omega, BMG LABTACH, German), with an excitation wavelength of 485 nm and an emission wavelength of 528 nm.

The glutathione peroxidase (GSH-px) vitality assessment
The GSH-px activity in the colon tissue and serum was determined in the present study (6 samples per group). Brie y, 50 mg frozen colon tissue was homogenized with 5 volumes (w/v) of extraction buffer and then the mixture was centrifuged at 4000 rpm for 10min to obtain supernatants. Similarly, serum samples were also diluted with extraction buffer at a ratio of 1:2 and then centrifuged to obtain supernatants. The GSH-px activity in supernatants were assayed with commercial GSH-px assay kit (Jiancheng Bioengineering Institute, Nanjing, China) following the manufacturer's instructions, and the absorbance was measured on the microplate reader at 412 nm. The blank and background control were also set.

Glutathione (GSH) content determination
Total glutathione (T-GSH, including GSH and GSSG) levels in the colon tissue and serum were determined in the present study (6 samples per group). Brie y, 50 mg frozen colon tissue was homogenized with 5 volumes (w/v) of homogenate medium provided by the manufacturer and then the mixture was centrifuged at 3500rpm for 10min to obtain supernatants. Similarly, serum samples were also diluted with homogenate medium at a ratio of 1:4 and then centrifuged to obtain supernatants. Then T-GSH and GSSG content in the supernatant were determined by T-GSH/GSSG kits (Jiancheng Bioengineering Institute, Nanjing, China). The absorbance of the resulting yellow color was measured using a microplate reader at 420 nm. To calculate the GSH/GSSG ratio, GSH (reduced form) was obtained by subtracting the 2×GSSG values from the total glutathione values.

DNA extraction and microbial community analysis
Feces samples for microbial community analysis were collected at Day 10 by using the Bacterial DNA-Free Eppendorf Tubes (Eppendorf AG, Hamburg, Germany), and then stored at −80 °C immediately. DNA concentration and purity was monitored on 1% agarose gels, then DNA samples were diluted to 1 ng/µL using sterile water. 16S rRNA genes of distinct regions were ampli ed by using speci c primer with the barcode (including 16S V4 515F-806R / 16S V3-V4 341F-805R / 16S V4-V5 515F-907R), and the primer sequences were shown as Table S4. All PCR reactions were carried out in a 30 µL reaction based on the manufacturer's recommendations (Phusion® High-Fidelity PCR Master Mix, New England Biolabs). The same volume of 1×loading buffer (contained SYB green) was mixed to PCR products and operate electrophoresis on 2% agarose gel for detection, then mixture PCR products was puri ed with GeneJETTM Gel Extraction Kit (Thermo Scienti c). Sequencing libraries were generated using Ion Plus Fragment Library Kit 48 rxns (Thermo Scienti c) following manufacturer's recommendations. The library quality was assessed on the Qubit@ 2.0 Fluorometer (Thermo Scienti c). At last, the library was sequenced on an Ion S5TM XL platform and 400 bp/600 bp single-end reads were generated.
Quality ltering on the raw reads were performed under speci c ltering conditions to obtain the highquality clean reads according to the Cutadapt quality controlled process. Subsequently, reads were compared with the reference database using UCHIME algorithm to remove chimera sequences and obtain the effective sequence (Clean reads) for subsequent analysis. Uparse software (Uparse v7.0.1001) was used for sequences analysis, for sequences with similarity greater than 97% assigned to the same Operational Taxonomic Units (OTUs), then representative sequences with highest frequency in each OTU were annotated basing on Silva Database to obtain the taxonomic information. Community compositions of every sample were count at levels of kingdom, phylum, class. Order, family, genus and species. The Multiple sequence alignment were conducted by Using MUSCLE software (Version 3.8.31) to study phylogenetic relationship of different OTUs and the difference of the dominant species in different samples / groups. QIIME (Version1.7.0) and R software (Version 2.15.3) were used to calculate the Alpha diversity index (including Observed-species, Chao1, Shannon, Simpson, ACE, Good-coverage) and analyze complexity of species diversity. For Beta diversity index which used to evaluate differences of samples in species complexity, QIIME (Version1.7.0) was used to calculate the weighted unifrac distance of samples, and R software (Version 2.15.3) was used to conducted cluster analysis (including non-linear model NMDS method and PCA) based on the weighted unifrac distance. Moreover, we selected the rst 10 genus with the highest abundance to draw a Taxtree, and ternary plot was used to compare the relative abundance of taxons between groups at different levels. Finally, we used the Tax4Fun package (taxon for function, based on KEGG database) in R software (Version 2.15.3) to predict the metabolic capacity based on 16S results, and LefSe analysis were selected to detect signi cant differences among groups in the microbial function biomarkers (LDA Score >2).

Statistical analysis
Microbial community analysis were conducted based on methods described above. Data exhibited a normal distribution according to the Kolmogorov-Smirnov test were expressed as means ± SD, and Oneway variance (ANOVA) with least signi cant difference (LSD) test was applied to evaluate the statistical signi cance of differences between the experimental groups and the control, otherwise the Mann-Whitney U test was applied. Analysis was carried out by using the Statistical Package for Social Sciences  Figure 1A-1C, during the presensitization stage, appearance and social behavior of mice in all groups were normal, no difference in body weight changes or coe cient of feeding e ciency (CFE) was observed between groups, and animals were at slow growth status. After TNBS colorectal instillation at Day 7, we found that mice in TNBS group displayed the lowest body weight and CFE.
However, the TNBS+VE group and TNBS+TiO 2 +VE group displayed signi cant increase in body weight when compared with the TNBS group at Day 8, which indicated that vitamin E and its co-administration with TiO 2 -NPs could mitigate colitis symptom in weight loss.
Recent clinic researches have reported the appearance of extraintestinal manifestations (EIMs) in crohn's disease, and kidney, spleen, liver as well as biliary tract were the most frequently involved organs [24,25]. As shown in Table 1, the organ weight coe cients of kidney and small intestine in TNBS group were signi cantly higher than the CT group, which indicated the potential organ injury during colitis inducement. Other experimental groups displayed signi cantly decreased kidney and small intestine organ weight coe cients when compared with TNBS group, which meant that the administration of TiO 2 -NPs and vitamin E as well as their co-administration could mitigate symptoms in organ injury. At last, no obvious difference was found in the coe cients of liver and spleen. As for pathological assessment, colitis mice in TNBS group and TNBS+TiO 2 group exhibited obvious renal pathological changes shown as glomerular hyperemia and hemorrhage, which would be further improved in TNBS+VE group and TNBS+TiO 2 +VE group after colitis mice received vitamin E administration ( Figure 2). Besides, no obvious pathological change could be observed in liver, spleen and mesenteric lymph nodes ( Figure S1), which prompted the localized EIMs and no systemic in ammation was found. Especially for Jejunum, we noticed that TiO 2 group and TNBS+TiO 2 group displayed signi cantly decreased villi height and villi height / crypt depth ratio when compared with CT group and TNBS group ( Figure S1 and Table S1), indicating that TiO 2 -NPs could disturb villi structure of heathy and colitis mice, which was also consistent with ndings in our previous publications [16, 26].

Colitis activity of animals.
The DAI has been developed to assess whether or not the colitis is progressing in clinic [27]. As shown in Figure 1D, the TNBS group displayed signi cantly elevated DAI score after the TNBS colorectal instillation when compared with the CT group, which represented active status of disease in colitis mice. Moreover, it could be noticed that TNBS+TiO 2 group, TNBS+VE group and TNBS+TiO 2 +VE group displayed signi cantly downregulated DAI scores when compared with the TNBS group from Day 8 to Day 10, though still higher than the CT group, indicating the potentially supportive mitigation of TiO 2 -NPs and vitamin E as well as their co-administration on colitis activity. At last, we also found that TiO 2 group displayed higher DAI score than the CT group during the late experimental administration, which suggested that oral TiO 2 -NPs administration was potential to arouse colonic in ammation risks in heathy mice.
For further assess the mucosal injury status of colon, the gross examination was conducted to re ect macroscopic damage status. As shown in Table 2, the TNBS group displayed signi cant shortening of length and notably elevated colon mucosal damage index (CMDI) score when compared with CT group, furthermore, we noticed lower CMDI scores and larger colon length of TNBS+TiO 2 group and TNBS+VE group than the TNBS group, which meant TiO 2 -NPs and vitamin E helped to mitigate the macroscopic structure damages of acute colitis. Moreover, TNBS+TiO 2 +VE group displayed larger colon length than the TNBS+TiO 2 group, indicating that effect of TiO 2 -NPs could be further enhanced after the coadministration with Vitamin E. At last, we also noticed the organ weight coe cients of colon in TNBS+TiO 2 group and TNBS+TiO 2 +VE group were signi cantly higher than that in TiO 2 group.
We further conducted pathological investigation to re ect the mucosal injury severity. As shown in Figure 3, H&E stained sections of colonic tissue presented mild mucosal in ammatory cell in ltrates with intact epithelium in TiO 2 group, which indicated epithelial in ammation risks after healthy mice received TiO 2 -NPs oral administration. Moreover, the TNBS group, TNBS+TiO 2 and TNBS+VE group exhibited obvious in ammatory cell in ltration as well as goblet cells and crypt structure loss, but these pathological changes seemed to be recovered in TNBS+TiO 2 +VE group after colitis mice received coadministration of TiO 2 -NPs and Vitamin E. Meanwhile, according to the result of pathological score assessment and mucosal height measurement (Table 2), the TNBS group and TNBS+TiO 2 group exhibited signi cantly higher pathological score and mucosal height than CT group and other experimental groups, which was similar to the result of gross examination, however, no signi cant difference could be observed between these two groups.

TiO 2 -NPs downregulated NF-κB transcription and decreased cytokine levels in colitis mice
Generally, nuclear factor κB (NF-κB) controlled transcription of in ammation genes in colonic epithelium.
Classic activation of NF-κB could be initiated by a broad panel of different stimuli, including abnormal intestinal pathogen through TLRs or colonic tissue damage from oxidative stress [6]. On activation, NF-κB was rapidly released from cytoplasmic inhibitor (inhibitor of NF-κB, IκB) with the function of IKKs, transmigrated into the nucleus, and bound to DNA response elements in gene promoter regions [28,29]. In the present study, we detected mRNA transcription levels of the NF-κB signal pathway in the colonic tissue of animals. As shown in Figure 4 and Figure S2, oral TiO 2 -NPs exposure signi cantly increased mRNA expression of P65 (subunit of NF-κB) in healthy mice. Meanwhile, the TNBS group showed signi cantly increased transcription levels of TLR-2, TLR-4, IKK2 and NF-κB P65 when compared with the CT group, which suspected an active NF-κB signal transcription in colon tissue of colitis mice. However, the transcription level of TLR-2, IKK2 and NF-κB P65 signi cantly decreased in the TNBS+TiO 2 group after colitis mice received additional TiO 2 -NPs administration, and the transcription level of NF-κB P65 also downregulated in the TNBS+VE group, which meant that both TiO 2 -NPs and vitamin E administration could downregulated the NF-κB signal transcription level in colon tissue of colitis mice.
Among the downstream signaling events of NF-κB signal pathway, the in ammatory molecules and biomarkers transcription were believed to participate in immune-mediated injury in the gut and promote the progression of IBD [6]. In the present work, we found that healthy mice in TiO 2 group displayed higher TNF-α mRNA transcription level than CT group. Meanwhile, colitis mice in the TNBS group showed signi cantly higher IL-1β, IL-10 and TNF-α transcription level in colon tissue when compared with healthy ones in the CT group, and the transcription level of HSP-90 also signi cantly decreased at the same time. Moreover, we noticed that the elevated IL-1β and TNF-α transcription level in TNBS group could be downregulated after colitis mice received TiO 2 -NPs administration, and the transcription level of antiin ammatory cytokine IL-10 could also be further upregulated, indicating that TiO 2 -NPs administration could participate in reducing in ammation level of TNBS-induced acute colitis. We also detected the expression level of typical in ammatory factors including TNF-α and IL-12 by Enzyme-linked immunosorbent assay (ELISA). As shown in Table 3, colitis mice in the TNBS group showed signi cantly higher TNF-α secretion level in colon tissue when compared with the CT group, and the TNBS+TiO 2 group, TNBS+VE as well as TNBS+TiO 2 +VE group further displayed lower TNF-α level than the TNBS group, which was consistent with changes of TNF-α mRNA transcription level. Besides, the serum TNF-α and IL-12 contents were also detected to measure systemic in ammation levels of animals, no signi cant change was found in serum contents of cytokines in the TNBS group when compared with the CT group, which meant the localized in ammation after acute colitis inducement. However, we noticed that TiO 2 group showed higher serum IL-12 level when compared with the CT group, which prompted TiO 2 -NPs administration could induce potential in ammation risk in healthy mice.
The body has antioxidant defenses mainly consisted of intracellular enzymatic antioxidants and nonenzymatic antioxidant glutathione to counteract the effects caused by ROS, and the uncontrolled oxidative stress was destructive to the GI tract and has been proved as an important pathogenic mechanism of IBD [3,5]. As shown in Table 4, we found that the glutathione peroxidase (GSH-px) vitality, T-GSH, oxidized glutathione (GSSG) as well as GSH content, but not the GSH / GSSG ratio, were signi cantly decreased in colon tissue of TiO 2 group after healthy mice received TiO 2 -NPs administration.
Meanwhile, the signi cant ROS accumulation and decreased GSH-px vitality, GSH, T-GSH content and GSH/GSSG ratio could be observed in the TNBS group when compared with the CT group, indicating severe oxidative stress in colon tissue of colitis mice. Furthermore, the GSH and T-GSH content in TNBS+TiO 2 group was lower than those in TNBS group, which meant that TiO 2 -NPs administration were potential to increase the antioxidant consumption and aggravate oxidative stress status in colon tissue of colitis mice. At last, the TNBS+VE group and TNBS+TiO 2 +VE group displayed signi cant decreased ROS accumulation in colon tissue when compared with colitis mice in TNBS group, but the GSH content and GSSG / GSH ratio were not obviously changed, which was in consistent with vitamin E's capability to scavenge oxygen free radical and prompted its effect on balancing oxidative stress in TNBS-induced acute colitis [30].
On the other hand, we also detected the GSH-px vitality and glutathione content in serum to re ect the systemic oxidative stress status of animals. As shown in Table 5 Previous studies have pointed out that compositional and metabolic changes in the gut microbiota were major driver of IBD [31], thus microbial community pro les were created by clustering 16S rRNA sequences into operational taxonomic units. The rarefaction and rank abundance curves exhibited the plateau and indicated the vast majority of bacterial species in the samples had been covered ( Figure S3). Table 6, we observed signi cant increase of the ACE and Chao1 richness estimator in the TNBS group when compared with the CT group, which indicated the imbalance of gut microbiome richness in colitis mice. Moreover, further downregulation of Chao1 index could be found in the TNBS+TiO 2 group than TNBS group after colitis mice received administration of TiO 2 -NPs, indicating that imbalance of gut microbiome richness in colitis mice would be partly repaired by TiO 2 -NPs. Similar changes could also be observed on other richness estimators like Shannon, Simpson, observed species and PD whole tree index, though without signi cance.

As shown in
Venn's diagram in Figure S4 showed that all of the groups shared the compositional overlap of 377 core microbiota, these overlapping phylotypes contributed to majority of the CT group and TiO 2 group, which shared proportion of 99.4% and 98.4% respectively. Meanwhile, more unique phylotypes were observed in colitis mice, among which the TNBS+TiO 2 group displayed the highest proportion of 10.8%. As shown in Figure 5A, the histogram generated by clustering analysis at phylum level divided the mouse feces samples into four major groups, the CT group, the TiO 2 group and TNBS group, the TNBS+VE and  Figure 5C) also showed signi cantly different microbial composition among groups (Amova, p <0.001). Compared with CT group, the obvious dissimilarities of gut microbiota composition were found in TNBS group, TNBS+VE and TNBS+TiO 2 +VE group. And after colitis mice received the single administration of TiO 2 -NPs or the coadministration of TiO 2 -NPs and vitamin E, obvious separations of gut microbiota composition could be further observed in the TNBS+TiO 2 group and TNBS+TiO 2 +VE group when compared with the TNBS group.

Microbial phylotype composition in mice
The classi cation of speci c taxonomy groups of species was conducted and shown as Figure 6, Firmicute (27.50%), Bacteroidetes (16.09%), Proteobacteria (7.13%) and Deferribacteres (0.91%) were four major phylum in feces samples of mice. Meanwhile, Proteobacteria phylum was primarily composed of Gammaproteobacteria-Stenotrophomonas strain (7.13%), Deferribacteres phylum mainly consisted the Deferribacteres-Mucispirillum strain (0.91%), and the TNBS+TiO 2 group shared the highest relative abundance proportion of these two strains. Moreover, Firmicutes phylum was primarily composed of Lactobacillaceae (15.69%), Enterococcaceae (1.32%), Ruminococcaceae (8.22%) and Lachnospiraceae (2.27%) at the family level, and it could be noticed that the TiO 2 group, TNBS group as well as the TNBS+TiO 2 +VE group shared high relative abundance proportion of the rst three taxonomy, respectively (Table S2). Then for other taxonomy at different levels, the classi cation of bacteria tended to be dispersed in groups and it would be hard to determine the dominant.
Furthermore, Ternary plot analysis was conducted to determine the predominant taxonomy among groups at the classi cation from phylum to genus. As shown in Figure 7

Gut-associated functional category of gut microbiota in mice
It was known that changes in gut microbiota structure could affect the gut-associated function by providing modi cations to gut microbiota that would affect molecule, enzyme, metabolism, biosynthesis and signal pathway. Thus the Tax4Fun analysis as well as LDA effect size (LEfSe) analysis were used to predict the functional pro les of gut microbiota based on the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. When compared with the CT group, we observed that functional category "ovarian_Steroidogenesis" and "Secretion_systerm" were signi cantly enriched in the TiO 2 group, while "Enzyme_families", "amino_sugar_and_nucleotide_ sugar_metabolism" and "lipid metabolism" were also diminished at the same time ( Figure 8A). Moreover, enhanced amino acid biosynthesis function (including "Arginine_biosynthesis", "Phenylalanine_tyrosine_and_tryptophan_biosynthesis" and "Valine_leucine_and_isoleucine_biosynthesis"), "Butanoate_metabolism", "Protein_folding_and_associated_processing", "Methane_metabolism", "Nitrotoluene_degradation" and "AMPK_signaling_pathway" could be observed in the TNBS group when compared with the CT group, but functional category "Galactose_metabolism" were diminished ( Figure 8B and 8C). Meanwhile, oral TiO 2 -NPs administration signi cantly enriched the category "Phenylalanine_metabolism" of colitis mice in TNBS+TiO 2 group ( Figure 8B), while vitamin E administration enriched the category "Pathogenic_Escherichia_coli_infection" in TNBS+VE group ( Figure 8C). At last, no functional category was signi cantly changed after the co-administration of TiO 2 -NPs and vitamin E.

TiO 2 -NPs mitigated TNBS-induced acute colitis by decreasing NF-κB signal transcription
In the present study, we focused on the effect of oral TiO 2 -NPs administration on development of TNBSinduced colitis. As shown in the Additional File j, the TNBS-induced acute colitis mice displayed dysbacteriosis and oxidative stress status in the colonic epithelium, as a result, the upstream activators (including TLRs and IKKs) of NF-κB-p65 would pervasively transcript and active the NF-κB signaling pathway, which nally increasing the in ammation levels in colonic epithelium and inducing disease activity in animals. Meanwhile, we noticed that TiO 2 -NPs administration seemed to be able to partly reverse the whole process in colitis mice (Figure 9 and Additional File k). TiO 2 -NPs administration still triggered oxidative stress status in the colonic epithelium of colitis mice, but also decreased the relative abundance of some pathogenic strains to repair the dysbacteriosis, further downregulated the TLRs and NF-κB signal transcription and in ammatory factor levels in colonic epithelium, and nally participated in the mitigation of TNBS-induced acute colitis.
The precise etiology of IBD remained unclear, present clinic studies have con rmed that the uncontrolled oxidative stress status would be a mechanism underlying the pathophysiology of IBD [30]. The oxidative stress status in colonic epithelium could activate nuclear factor-κB signal pathway transcription and promote release of pro-in ammatory cytokines [32,33]. It was found in previous researches that TiO 2 -NPs could induce decreased antioxidant enzyme activity and increased ROS accumulation in the colon of mice, which further leading to adverse consequences such as in ammatory cell recruitment, proin ammatory factor expression and abnormal proliferation of epithelium [34,35], and it was also suspected that TiO 2 -NPs might exacerbate in ammation risks of colitis by inducing excessive oxidative stress [32]. In the present study, we did nd that TiO 2 -NPs administration signi cantly downregulated T-GSH and GSH content and partly upregulated ROS accumulation (p=0.13) in colonic epithelium of colitis mice. As the most abundant low molecular weight thiol compound synthesized in cells, GSH could play a critical role in protecting cells from oxidative damage and xenobiotic electrophiles toxicity, as well as help to maintaining the redox homeostasis [36], thus we agreed that TiO 2 -NPs induce antioxidant system disorders and aggravate oxidative stress in colonic epithelium of colitis mice. However, colitis mice displayed downregulated in ammatory factor levels and disease activity index after oral TiO 2 -NPs administration, which indicated the mitigation effect of TiO 2 -NPs on colitis and this should be independent of oxidative stress aggravation effect of TiO 2 -NPs. For better demonstration on this view, some colitis mice in the present study also received co-administration of TiO 2 -NPs and vitamin E. As an effective free radical scavenger, vitamin E has been well established to exert excellent antioxidant effect through hydrogen atom transfer reaction to yield non-radical product [37,38]. According to present results, colitis mice received co-administration of TiO 2 -NPs and vitamin E displayed signi cantly lower ROS accumulation in colonic epithelium when compared with those only received TiO 2 -NPs, thus it could be conducted that vitamin E effectively scavenged the ROS and antagonized with the oxidative stress status induced by TiO 2 -NPs. We also noticed that vitamin E helped to enhance the mitigation effect of TiO 2 -NPs on colitis, suggested that the TiO 2 -NPs induced oxidative stress made an adverse effect on the development of colitis. Hence, we believed that TiO 2 NPs alleviated colitis through other pathways which made promoting effect, and the signi cance was far greater than the adverse effect of oxidative stress. Further effort should be taken to fully explain the mechanism.
On the other side, the inappropriate response of dysfunctional mucosal immune system to the resident microbiota and other noxious antigens was also central to the development of IBD [28]. Previous studies also proved that TiO 2 -NPs could possess excellent antimicrobial activity for its nanostructure, large speci c surface area (SSA) and high surface activity, which was suspected to change the microbial composition in in vivo and in vitro [34,35,39]. In the present work, we found that TiO 2 -NPs signi cantly corrected the imbalanced richness as well as evenness of gut microbiota, and changed predominant taxonomy in colitis mice, including decreased relative abundances of Enterococcus_faecalis and Bacteroides_acidifaciens, and increased relative abundance of Mucispirillum_schaedleri . On the basis of microbiology research, increased abundance of Enterococcus_faecalis and Bacteroides_acidifaciens are both prominent features in patients with CD and UC [40,41], among which Enterococcus_faecalis could release metalloprotease to compromise epithelial barrier, further induce the overexpress of proin ammation cytokines (including IL-6-IL-17α and IL-23α) and clinically active disease in patients [42,43], besides, Bacteroides_acidifaciens was found to produce signi cant starting point to in ammatory cascades and eventually lead to development of IBD [44]. According to the correlation analysis ( Figure   S6A and S6C), we found that relative abundance of Enterococcus_faecalis and Bacteroides_acidifaciens were positively correlated with NF-κB signal transcription level in colonic epithelium (including IKKs, NF-κB-p65 and IL-1β mRNA level) and colitis symptom indicators (including elevated DAI and CMDI score as well as decreased body weight). Except for changes of pathogenic strains, model research also pointed out that Mucispirillum_schaedleri could help to improve the functional mucus layer in the gut and conferred protection against colitis in mice [45], our correlation analysis also showed that relative abundance of Mucispirillum_schaedleri was negatively correlated with TNF-α expression level in colon tissue. Collectively, considering the shift of dominantance from pathogenic strains to probiotic in gut microecology of colitis mice after TiO 2 -NPs oral administration, we believed it could be an important mechanism that TiO 2 -NPs participated in the mitigation of TNBS-induced colitis.
Furthermore, as the most common pattern-recognition receptors, TLRs could sense and recognize pathogen-associated molecular patterns (PAMPs) [46]. Upon their activation under stress of gut dysbacteriosis, TLRs could mediate the phosphorylation of IκB to activate NF-κB signal transduction cascade and induce the overexpression of in ammatory cytokines [28]. Hausmann et al. [47] detected biopsy samples from patients with active IBD and found that the colonic epithelial layer showed signi cant positive staining for TLR2 and TLR4, meanwhile the mRNA transcription of pro-in ammatory cytokines such as IL-1β was also signi cantly upregulated. By now few previous work had explored the effect of TiO 2 -NPs on gut TLRs, but present work suggested that the decreased relative abundance of pathogenic strains and gut dysbacteriosis stress after TiO 2 -NPs administration could interact with the decreased TLRs transcription level in colitis mice, which would further induce the lower transcription of IKK2 and NF-κB p65, nally contribute to the downregulation of IL-1β and TNF-α as well as the in ammation suppression. Except for the changes of pro-in ammatory signal transcription, we also found that TiO 2 -NPs administration would increase IL-10 mRNA transcription in colitis mice. Previous work reported that chronic food-grade TiO 2 administration could signi cantly increased IL-10 expression in rat colon [23]. As an important anti-in ammatory cytokine, IL-10 was proved to block IKK activation and directly inhibit the nuclear localisation of the NF-κB p65/p50 heterodimer, which participated in resolution of many major in ammatory response and contributed to wound healing [48, 49,50]. Thus we believed that the biological effect of TiO 2 -NPs on increasing IL-10 mRNA transcription could also participate in the mitigation of TNBS-induced colitis, and the mechanism could also be related with the effect of TiO 2 -NPs on increasing relative abundance of Mucispirillum_schaedleri ( Figure S6C).
Taken together, TiO 2 -NPs could aggravate oxidative status in colonic epithelium of colitis mice and repair the gut dysbacteriosis at the same time, but it was the latter rather than former made the predominant role and downregulated the NF-κB signal transcription and in ammation level, which nally contributed to the mitigation TNBS-induced acute colitis. Due to its unique properties, there has been an increase in biomedical applications of TiO 2 -NPs such as drug delivery, bioimaging, photoablation therapy and biosensors [51]. Recently, the application of TiO 2 -NPs conjugated with speci c monoclonal antibody has been proved to photokill human metastatic colon adenocarcinoma (LoVo) cancer cells effectively [52].
However, the safety of TiO 2 -NPs in the oral drug delivery system has long been concerned [53]. Our research provided a new understanding about oral intake of TiO 2 -NPs, this material actually exhibited mitigation effects for individuals with colitis, and it could be an important inspiration for the safe biomedical application of TiO 2 -NPs, though further effort on clinic research was still needed.
2. TiO 2 -NPs induced oxidative stress and colonic in ammation risks in healthy mice.
Except for the effect of TiO 2 -NPs on acute colitis, the present study also investigated the in ammation risk of healthy mice after ingesting TiO 2 -NPs. As shown in Figure S7, TiO 2 -NPs induced oxidative stress status and increased NF-κB signal transcription level in colonic epithelium, meanwhile, TiO 2 -NPs in uenced the prediction of gut microbiota functions, though not be able to change predominant taxonomy in healthy mice.
The disorder of glutathione antioxidant system in colon tissue showed as downregulated GSH and T-GSH content as well as the decreased GSH-px activity. As the most effective antioxidant against hydroperoxides in cells, GSH-px provides the most vital defense against the peroxidative damage of biological membranes in mammalian cells [54]. Previous clinic researches reported a signi cantly decreased serum GSH-px activity in IBD active patients when compared with healthy controls [55], and in vivo studies also found that oral TiO 2 -NPs exposure could induce decreased GSH-px activity in liver, kidney, brain and serum of rats and mice, which further lead to organ in ammation as well as functional disorders [56,57]. Li et al. [58] and Hugo et al. [59] pointed out that the antioxidant activity of GSH-px played an important role in inhibiting IKKs activity and NF-κB signaling pathway activation. In the present work, the NF-κB signaling transcription level (including NF-κB-P65 and TNF-α) in colonic epithelium of healthy mice was signi cantly upregulated after oral TiO 2 -NPs ingestion, but the transcription level of IKKs were not obviously changed, thus only limited evidence support that the oxidative stress status induced by TiO 2 -NPs administration could induce in ammation by activing NF-κB signal transcription.
Moreover, we also found that TiO 2 -NPs administration changed prediction of gut microbiota functions in healthy mice, such as the reduced function in amino_sugar_and_nucleotide_sugar metabolism as well as lipid metabolism. Metabolic studies also pointed out these functional category changes were potential to disturb lysine acetylation and SCFA synthesis, which could in uence the colonic barrier function in host and result in in ammation risks [60, 61]. On the other side, the ternary plot analysis suggested that no signi cant dominant distribution of taxonomy was reserved in the TiO 2 group when compared with CT group and TNBS group, which meant the effect of TiO 2 -NPs on microbial phylotype composition could be scant and not be able to induce obvious gut dysbacteriosis, especially when compared with the TNBScolitis inducement. Meanwhile, considering that the transcription levels of TLRs were also not signi cantly changed, we didn't think TiO 2 -NPs administration could upregulate the NF-κB signal transcription by improving TLRs transcription and recognizing PAMPs caused by gut dysbacteriosis stress.
Except for the upregulated mRNA transcription of pro-in ammation factors in colonic epithelium, we also noticed that the serum IL-12β level increased after healthy mice orally exposed to TiO 2 -NPs. Previous studies found that TiO 2 -NPs exposure by intratracheal instillation could signi cantly increase IL-12β serum level in a dose-dependent manner in mice, while oral exposure to TiO 2 -NPs also induced overexpression of IL-12β in liver and intestine of mice [17,62]. As an important pro-in ammatory cytokine, IL-12β participated in inducing naive CD4+T cells to differentiate into Th1 cells, which could be implicated in the etiology of a number of human autoimmune diseases [63]. Clinic researches also reported that serum IL-12β levels were positively correlated with clinical and endoscopic disease activity in IBD patients, which suggesting it as a biomarker for assessing disease activity [64]. In present study, histological examination displayed mild in ammatory cell in ltrates in colonic mucosa of healthy mice after oral TiO 2 -NPs administration, which could be strong mutual corroboration with serum IL-12β changes and suggested the odds of in ammation risks.
In short, the present study prompted that TiO 2 -NPs ingestion induced in ammation risks in the colonic epithelium of healthy mice, the oxidative stress status or functional category changes caused by TiO 2 -NPs might be interrelated with the whole process but should not be the driving factors, and further efforts were also recommended to investigate the possible mechanism.

Conclusion
In summary, ndings in our present study showed that orally administered TiO 2 nanoparticles could mitigated the TNBS-induced acute colitis through the repairs to gut dysbacteriosis and downregulation of NF-κB signal transcription. TiO 2 -NPs indeed exacerbated oxidative stress status in colonic epithelium of colitis mice, but the co-administration with vitamin E also demonstrated that the mitigation of TiO 2 -NPs on colitis was independent from their effect on oxidative stress inducement. It could be suspected that the promoting effect of ora imbalance repair was far greater than the adverse effect of oxidative stress induced by TiO 2 -NPs on colitis mice. Considering the prospective bio-effect of TiO 2 -NPs, we strongly recommended more chronic colitis experiments to be carried out in the future to provide an insight into the potential role of TiO 2 -NPs in alleviating the disease.
Meanwhile, we also noticed that TiO 2 -NPs induced upregulated NF-κB signal transcription and potential in ammation risks in healthy mice. Since the use of TiO 2 -NPs as food additive and medical adjuvant is on the rise, more cautious use of these particles was suggested before we got biologically relevant exposure information to facilitate interpretation of their safety.       Figure 1 Effect of TiO2-NPs on TNBS-induced mice acute colitis. Body weight (A), coe cient of feeding e ciency (CFE, %) (B), total CFE (C) and disease activity index (DAI) score (D) of animals after the TNBS-colitis inducement and TiO2-NPs or vitamin E treatment (n=9). For CFE, mice were fasted at Day 6 for the TNBS colorectal instillation, thus the CFE of mice at Day 6 was not included. Signi cant difference was found between the CT group and the experimental groups (** p < 0.01, * p < 0.05). Signi cant difference was also found between the TNBS group and other experimental groups ( aa p < 0.01, a p < 0.05). E. Timeline of the experimental protocols, TTS was for the trandermal thrapeutic systems.  TiO2-NPs and Vitamin E administration changed gene transcription in colonic epithelium of mice (n=6).

Figures
Per gene, treatments groups that do not share any letter are signi cantly different according to a one-way ANOVA with LSD (p < 0.05).