Perturbation of intestinal stem cell homeostasis and radiation enteritis recovery via dietary titanium dioxide nanoparticles

Abstract Small intestinal health and enteritis incidence are tightly coupled to the homeostasis of intestinal stem cells (ISCs), which are sensitive to dietary alterations. However, little is known about the impact of food additives on ISC pool. Here, we demonstrate that chronic exposure to low‐dose TiO2 NPs, a commonly used food additive, significantly hampers primary human and mouse ISC‐derived organoid formation and growth by specifically attenuating Wnt signal transduction. Mechanistically, TiO2 NPs alter the endocytic trafficking of the Wnt receptor LRP6 and prevent the nuclear entry of β‐catenin. Notably, dietary TiO2 NPs elicit modest chronic stress in healthy intestines and considerably impede the recovery of radiation enteritis by perturbing the homeostasis of ISCs in vivo. Our results identify a health concern of TiO2 NP exposure on ISC homeostasis and radiation enteritis recovery. These findings suggest extra precaution during the treatment of radiation enteritis and provide new insights into food additive‐ISC interaction.


| INTRODUCTION
The small intestinal epithelium is responsible for food digestion and nutrient absorption and plays a critical role in microorganism defence and immune response. 1,2 Since toxins in food can directly contact the intestine and induce intestinal epithelial cell (IEC) death, a rapid turnover of IECs is critical to maintain the integrity of small intestines. 1,3 Intestinal stem cells (ISCs), which are capable of both self-renewal and differentiation into enterocytes or secretory-lineage cells, 4,5 are indispensable for epithelial renewal. ISCs can derive single crypt to form intestinal organoid (also called mini-guts) with 3D structure in vitro. 6 Numerous studies have demonstrated that ISCs are closely related to the health and enteritis incidence of small intestines. 1,[7][8][9] Recent evidence indicates that the self-renewal and differentiation of ISCs can be controlled at least in part by dietary patterns including caloric restriction, 10 fasting, 11,12 high-fat diets, 13,14 ketogenic diets and highcarbohydrate diets as well as other nutrients. [15][16][17][18][19] However, little is known about the impact of food additive exposure on ISC homeostasis.
Nanoparticles (NPs) have brought a range of benefits to the food sector. Nanoscale titanium dioxide (TiO 2 ) is one of the most widely manufactured NPs and accounts for more than 36% of TiO 2 particles worldwide. 20 Currently, as a food additive, TiO 2 NPs have been added to more than 900 foods to enhance the opacity and brightness of products. 21,22 Candies, sweets and chewing gums have the highest content of TiO 2 NPs. 20 In addition, since TiO 2 NPs can considerably prevent microbial growth and reduce E. coli contamination of food surfaces, they have been commonly used in food packaging and storage. 23 Given that TiO 2 NPs have been widely applied in the food industry, growing concerns about the potential health risks of TiO 2 NPs on the intestines have been raised. Researchers found that foodborne TiO 2 NPs could induce strong gut microbiota dysbiosis, colonic inflammation and proteome alterations in obese mice. 24 Additionally, long-term intake of the food additive TiO 2 NPs altered the intestinal epithelial structure. 25 Our previous study also revealed that dietary TiO 2 NPs impeded the recovery of intestinal mucosal damage in colitic mice. 26 Nevertheless, studies of food additive TiO 2 NPs in ISC homeostasis are still lacking.
In clinic, radiation enteritis is inflammation of the intestines that occurs after radiation therapy, a primary treatment for malignant diseases and is commonly administered to patients with gynaecological, urological and gastrointestinal cancers. Almost all patients receiving pelvic or abdominal radiotherapy experience radiation enteritis, 27 which usually manifests as pain, nausea, bloating and diarrhoea. 28 Studies have demonstrated that ISCs are responsible for the recovery of radiation enteritis. 29 However, whether chronic TiO 2 NP intake influences the recovery of radiation enteritis remains poorly understood.
In this study, we investigated the impact of TiO 2 NPs on the homeostasis of small intestinal stem cells. Our results showed that TiO 2 NPs were mainly internalized via clathrin-mediated endocytosis pathway during intestinal epithelial cell metabolism. Analysis of primary mouse and human organoids revealed that TiO 2 NPs perturbed the homeostasis of ISCs by specifically attenuating Wnt signalling. At the molecular level, we demonstrated that TiO 2 NPs triggered noncanonical endocytic trafficking of the Wnt receptor low-density lipoprotein receptor-related protein 6 (LRP6) and prevented the nuclear entry of β-catenin to weaken the outputs of Wnt signalling. Moreover, blockade of Wnt signalling by TiO 2 NPs elicited modest chronic stress in healthy intestines and markedly impaired the recovery of radiation enteritis by inhibiting the proliferation of ISCs in both Lgr5-eGFP-IRES-creERT2 and C57BL/6 wild-type (WT) mice. Together, our work revealed the impact of chronic TiO 2 NP exposure on the homeostasis of ISCs and the recovery of radiation enteritis, and disclosed the mechanism underlying these effects. These findings give insights into the treatment of radiation enteritis, and expand our understanding of the food additive-bio effects.

| The cytocompatibility of TiO 2 NPs
The physiochemical properties of TiO 2 NPs used in our study have been characterized by transmission electron microscope (TEM) and dynamic light scattering (DLS) analyses. 26 The uptake efficiency and mechanism of TiO 2 NPs were explored in the mouse colon cancer cell line CT26. There are three main types of endocytosis in nonphagocytic cells: clathrin-mediated endocytosis (CME), caveolin-mediated endocytosis (CavMe) and micropinocytosis. 30 Our TEM analysis demonstrated a rapid internalization of TiO 2 NPs in CT26 cells, while the uptake of TiO 2 NPs was significantly suppressed when cells were pre-incubated with sucrose, genistein and cytochalasin D, which are known as inhibitors for CME, CavMe and macropinocytosis, respectively ( Figure S1A). In particular, compared with other inhibitors, pre-incubation of CT26 cells with sucrose dramatically impaired the cyto-endocytosis of TiO 2 NPs (Figure S1B), suggesting that TiO 2 NPs were mainly internalized via the CME pathway. We then examined the cytotoxicity of TiO 2 NPs in CT26 cells. The results showed that TiO 2 NPs had a negligible influence on the viability of CT26 cells at concentrations between 10 μg/mL and 150 μg/mL within 48 h ( Figure S1C,D). The results suggested a good cytocompatibility of TiO 2 NPs in intestinal epithelial cells.

| TiO 2 NPs perturb ISC homeostasis in mouse intestinal organoids by specifically suppressing Wnt signalling
Although incubation with a wide range of TiO 2 NPs did not compromise the viability of intestinal epithelial cells, we continued to investigate whether uptake of TiO 2 NPs could affect the homeostasis of ISCs in mouse intestinal organoids ( Figure 1A). ISCs are characterized as Lgr5-GFP + cells and CD44 + CD24 lo CD166 + GRP78 lo/À cells in Lgr5-eGFP-IRES-creERT2 mice and C57BL/6 WT mice, respectively. 31,32 Given that Lgr5-eGFP-IRES-creERT2 mice can directly identify ISCs and that C57BL/6 WT mice are not limited by mosaic expression patterns and have ISC markers similar to those of the human intestine, we isolated crypts from both species of mice. As the results showed, ISCs are interspersed between Paneth cells in intestinal crypts ( Figure S2A). The isolated crypts initially form villus-like spherical structures with a closed-loop hollow lumen, and then the cyst buds up, differentiates into a crypt-like structure, and finally forms a mature organoid structure ( Figure S2B). Lgr5-GFP + cells were present in the crypt-like domain of the organoid ( Figure S2A). To verify the cell types in C57BL/6 WT mouse-derived organoids, the cellular markers LGR5, villin, MUC2, lysozyme and chromogranin A (CHGA) were stained to characterize the ISCs, enterocytes, goblet cells, Paneth cells and enteroendocrine cells, respectively ( Figure 1B).
Our results demonstrated that all of these markers were present in the established intestinal organoids, and granule-containing Paneth cells could also be clearly observed in the organoid crypts by using an inverted microscope ( Figure S2C). These results demonstrated that the established intestinal organoids displayed mature morphology.
We subsequently explored the impact of TiO 2 NPs on ISC homeostasis using the in vitro model and found that the addition of 50 μg/mL  Figure S3A). In addition, the presence of TiO 2 NPs considerably decreased the percentage of ISCs in Lgr5-eGFP-IRES-creERT2 and C57BL/6 WT mouse-derived organoids ( Figure 1H and Figure S3B,C), demonstrating that TiO 2 NPs affected the homeostasis of ISCs in the 3D in vitro models. Further investigation revealed that the expression of stemness markers (Lgr5 and Oct4) in C57BL/6 WT mouse-derived organoids was considerably reduced ( Figure S3D). These results thus demonstrated that TiO 2 NPs could perturb ISC homeostasis.
Since the Wnt, BMP and Notch signalling pathways play central roles in maintaining the homeostasis of ISCs. [33][34][35][36][37] We subsequently asked whether intracellular TiO 2 NPs affected the activity of these signalling pathways. Our results showed that TiO 2 NPs had little effect on the protein levels of Hes family bHLH transcription factor 1 (HES1) and p-SMAD1/5, which is well-known Notch signalling target gene and BMP signalling reporter, respectively. However, the expression of the Wnt target gene c-Myc was largely reduced in the presence of TiO 2 NPs, and the addition of TiO 2 NPs markedly reduced the protein level of β-catenin, a key transcriptional activator to be degraded in the inhibition of Wnt signalling ( Figure 1I and Figure S4). which have also been reported to regulate stem cell homeostasis 38,39 ( Figure 1J). These findings suggested that TiO 2 NPs perturbed ISC homeostasis by repressing Wnt signalling.

| Wnt signalling blockade by TiO 2 NPs impairs primary human ISC homeostasis
Encouraged by the results observed in mouse-derived 3D models, we established human small intestinal organoids to further validate the effects of TiO 2 NPs on ISC homeostasis ( Figure 2A). The crypts of healthy human small intestine were isolated and cultured as previously described. 40 In Wnt-rich culture medium, human small intestinal crypts undergo multiple crypt fission events and finally generate cystic organoids, which comprise mainly stem cells and their highly proliferating progenitor cells. The organoids were stable in culture over passage ( Figure S5 Given that the expression of LRP6 was not affected, we further asked whether the endocytic trafficking of LRP6 was altered. It has been reported that the internalization of LRP6 with caveolin is necessary for Wnt signal transduction. [41][42][43][44] Consistently, our results showed that internalized LRP6 was primarily colocalized with caveolin rather than clathrin. However, the addition of TiO 2 NPs selectively sequestered LRP6 from caveolin-mediated endocytosis to the clathrin-dependent endocytic route ( Figure 3D-F). These results were consistent with our findings that TiO 2 NPs were mainly internalized via the clathrin-dependent pathway. Together, these data clarified that TiO 2 NPs regulated Wnt signalling by triggering non-canonical endocytosis of LRP6.

| TiO 2 NPs perturb ISC homeostasis and induce modest chronic stress in the mouse small intestine
Next, we investigated the long-term effects of dietary TiO 2 NPs, administered by drinking water, on ISC homeostasis in Lgr5-eGFP-IRES-creERT2 and C57BL/6 WT mice ( Figure 4A). FDA allows 1 wt% TiO 2 in food, and it was found that more than 36% of particles are nanoscale. 20 In our study, mice were administered 50 μg/mL TiO 2 NPs, and the dose was 70 times lower than the FDA-allowed level.
Our results showed that administration of 50 μg/mL TiO 2 NPs for 2 months had little effect on the length of intestines but induced modest chronic stress in the small intestines, as the epithelial cells of villi were moderately damaged ( Figure 4B and Figure S8A,B). In addition, dietary waterborne TiO 2 NPs significantly hampered the proliferation of ISCs, as the colocalization of Lgr5-GFP + ISCs and Ki67 was markedly reduced in the small intestine of Lgr5-eGFP-IRES-creERT2 mice ( Figure 4C). Similar results were found in C57BL/6 WT mice ( Figure S8C,D). As a result, the proportion of Lgr5-GFP + cells in Lgr5-eGFP-IRES-creERT2 mice and CD44 + CD24 lo CD166 + GRP78lo /À cells in C57BL/6 WT mice was considerably reduced by dietary TiO 2 NPs ( Figure 4D and Figure S9). The mRNA levels of ISC markers such as Lgr5, Bmi1 and Nanog were also notably weakened in C57BL/6 WT mice administered TiO 2 NPs ( Figure 4E). Next, we verified the inhibi-

| TiO 2 NPs suppress radiation enteritis recovery by inhibiting ISC proliferation
We further explored the role of TiO 2 NPs under pathological processes that require a higher threshold of ISC activity to maintain intestinal homeostasis. To address whether dietary TiO 2 NPs affect the recovery of radiation enteritis, mice were exposed to 50 μg/mL TiO 2 NPs by drinking water for 4 weeks, with 12 Gy abdominal x-ray irradiation performed in the second week ( Figure 5A and Figure S11A). Our results showed that ionizing radiation (IR) induced visible small intestine oedema and significantly impaired the villi and crypts in the small intestine on day 3.5 after IR ( Figure 5B). In addition, the percentage of ISCs was markedly blunted by IR, and dietary TiO 2 NPs did not aggravate the damage on day 3.5 ( Figure 5C and Figure S11B), indicating TiO 2 NPs did not strengthen IR-induced enteritis. However, after radiation for 14 days and 21 days, TiO 2 NP-treated mice exhibited worse small intestine histology, as the intestinal villus and crypt length were markedly reduced compared to the IR group ( Figure 5B and Figure S11C). The number of apoptotic cells in the villi and crypts in TiO 2 NP-administered mice was notably higher than that in mice treated with IR alone, indicating that dietary TiO 2 NPs suppressed the recovery of the small intestinal epithelium ( Figure 5D). Given that ISCs are indispensable for small intestine regeneration, we further analysed the impact of TiO 2 NPs on ISCs in mice after radiation for 14 days.
Surprisingly, long-term intake of waterborne TiO 2 NPs considerably inhibited the proliferation of ISCs and decreased the proportion of ISCs in radiated mice ( Figure 5C,E). Since Wnt signalling was activated to stimulate ISC regeneration after ionizing radiation, 45

| DISCUSSION
The small intestinal epithelium is a rapidly renewing tissue. The rapid renewal is maintained by ISCs, which are tightly coupled to intestinal health and enteritis. 2 The imbalance of small ISC homeostasis will compromise the integrity of mucosal barrier and inhibit the recovery of inflammation. In clinic, almost all patients receiving pelvic or abdominal radiotherapy experience radiation enteritis. Recent evidence indicates that the self-renewal and differentiation of ISCs can be controlled at least in part by diet and nutritional status. [10][11][12][13][14][15][16][17][18][19] However, studies of food additives in ISC homeostasis are still lacking.
NPs have been widely used in the food industry for their broad benefits, and good biocompatibility of NP is a prerequisite for foodbased applications. [46][47][48] As one of the most manufactured nanoparticles, TiO 2 NPs have been commonly used as food additive. However, the impact of chronic NP exposure on ISC homeostasis is still unexplored. Additionally, the doses of TiO 2 NPs employed in many studies were too high, and TiO 2 NPs were usually administered to mice by oral gavage, a technique that has been shown to cause stress responses in the gut. Here, we limited the dose of TiO 2 NPs to 70 times lower than the FDA-allowed level, and administered TiO 2 NPs as a part of mouse daily water to minimize the handling stress to assess the impact of TiO 2 NPs on ISC homeostasis. We found that the also marks stem cells, 31,51 indicating that TiO 2 NP-induced abnormalities in Lgr5 + ISCs may also occur in other tissues. In particular, given that LGR5 is also expressed in scattered cells in pre-malignant mouse adenomas, 52-54 TiO 2 NPs could potentially be applied to inhibit the progression of diseases in which Lgr5 + cells play essential roles.
Taken together, our findings reveal the impact of long-term TiO 2 NP exposure on ISC homeostasis and the recovery of radiation enteritis, and disclosed the underlying mechanism. These findings suggest extra precaution during the treatment of radiation enteritis, and provide a better understanding of the relationship between food additives and ISCs.

| Cell culture
The mouse colon cancer cell line CT26 was obtained from the Cell Bank of Chinese Academy of Sciences. The cells were cultured at 37 C in a humidified incubator with 5% CO 2 . RPMI 1640 medium, foetal bovine serum (FBS), streptomycin and penicillin were purchased from Gibco.

| Histological analysis of organoids
Organoid in Matrigel were resuspended with 10% neutral formalin and incubated for 1 h on ice. After washing with PBS to remove Matrigel, organoids were then fixed overnight at 4 C. Paraffin sections were processed with standard techniques. Haematoxylin and eosin (H&E) and the indicated immunofluorescence staining were performed to characterize the morphology and composition of the organoids. All fluorescence was imaged with a ZEISS Apotome 3.

| Statistical analysis
To determine the colony-forming efficiency, crypts were cultured in 24-well plate. Seven days after plating, spheres were counted and the colony-formation efficiency was calculated (number of organoids formed/number of crypts seeded Â 100%). The relative organoid area was counted with ImageJ and the budding events in each well were counted to calculate the relative budding events. Student's t-test was used for comparison of two samples with unequal variances.

| Confocal imaging
To investigate the endocytic trafficking of the Wnt receptor LRP6, CT26 cells were treated with or without 50 μg/mL TiO 2 NPs. After 48 h of incubation, the cells were fixed in 4% PFA for 20 min at room temperature, washed with PBS and blocked with goat serum for 1 h.
Cells were incubated with the indicated primary antibody at 4 C overnight and the corresponding fluorescent secondary antibody for 2 h at room temperature. After washing, 4 0 ,6-diamidino-2-phenylindole (DAPI) was stained at room temperature for 8 min. All fluorescence was imaged with a confocal laser scanning microscope (ZEISS). The colocalization of clathrin or caveolin with LRP6 was quantified with ImageJ.

| Fluorescence activated cell sorting (FACS) analysis
To investigate the impact of TiO 2 NPs on the percentage of ISCs, mouse intestines or cultured mouse/human organoids in each group were obtained and mechanically dissociated and centrifuged at 290 g for 5 min to isolate intestinal crypts. Two millilitres of TrypLE Express