Air pollutant particulate matter 2.5 induces dry eye syndrome in mice

In this study, we explored the effects of particulate matter 2.5 (PM2.5) eye drops on the ocular surface structure and tear function in mice and established a novel animal model for dry eye research. We found that, following treatment with PM2.5, the tear volume and, the tear film break-up time showed statistical differences at each time point (P < 0.05). The FL score of the PM2.5-treated group was higher than that of others (P < 0.05). The average number of corneal epithelial layer cells in groups A and B was significantly lower than that in group C (P < 0.05). Scanning electron microscopy and transmission electron microscopy revealed that the number of corneal epithelial microvilli and corneal desmosomes was drastically reduced in group C. PM2.5 induced apoptosis in the corneal superficial and basal epithelium and led to abnormal differentiation and proliferation of the ocular surface with higher expression levels of Ki67 and a reduced number of goblet cells in the conjunctival fornix in group C. PM2.5 significantly increased the levels of TNF-α, NF-κB p65 (phospho S536), and NF-κB in the cornea. Thus, the topical administration of PM2.5 in mice induces ocular surface changes that are similar to those of dry eye in humans, representing a novel model of dry eye.

causing asthma, thrombosis, and myocardial ischemia, but also shortens the average life expectancy of individuals. However, there are few studies concerning the influence of PM 2.5 on eyes. Since the eye is one of the organs in direct contact with the outside world, PM 2.5 might have a direct impact. An investigation of 71 drivers by Torricelli et al. reported that the BUT values for these drivers were lower than that in a normal person 11 . Tatsuya et al. surveyed patients with acute conjunctivitis from May to October 2012 and found that the number of patients with acute conjunctivitis was increased with a higher level of PM 2. 5 12 . Camara et al. investigated the influence of volcanic smoke. The main components of these pollutants are PM 2. 5 , which have been found to cause some ocular symptoms, such as eye itching, foreign body sensation, tears and burning, as well as some other signs such as conjunctival congestion, increased mucus secretion, conjunctival keratoconus, swelling of the eyelids and conjunctival edema 13 . Although PM 2.5 can cause many symptoms, the mechanism by which PM 2.5 causes damage is unclear.
Phagocytosis procedure. Latex bead (LB) solution (Sigma L4530, carboxylate modified polystyrene, USA) was prepared at a ratio of 1:10 in HCE culture medium. HCE cells were plated at 1 × 10 5 cells per well in 24-well culture plates and grown until 80% confluency prior to treatment. The culture medium was removed and replaced with culture medium supplemented with LB solution. Next, the cells were washed and mounted in mounting medium containing 4, 6-diamino-2-phenyl indole (DAPI, Burlingame, CA, USA) at 0, 3, 6, 9, 12, 24, or 48 hours after treatment. Images were photographed with a Nikon TE-2000 U Eclipse epifluorescence microscope (Nikon Instruments, Tokyo, Japan).
Animal preparation. In total, 90 male specific pathogen-free (SPF) BALB/c mice (18-21 g in weight, from Laboratory Animal Center of Xi'an Jiao Tong University College of Medicine, Xi'an, China) were used in this study. No abnormality was found in the anterior segment and fundus when examined using a slit lamp microscope. The results for the Schirmer I test (SIT) were ≥10 mm/5 min. The mice were housed in a standard environment throughout the study 17 . All procedures were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the animal ethics committee of Xi'an Jiao Tong University College of Medicine (Xi'an, China).
Acquisition of PM 2.5 and preparation of eye drops. PM 2.5 samples were provided by the Xi'an environmental monitoring station. During October 1 to 31, 2015, a super station at Xi'an City acquired PM samples with sizes of 2.5 μm using the TH-16A four-channel atmospheric particulate automatic sampler (Wuhan Tianhong Instrument Ltd) and filtered them through Whatman PTFE membranes. Sampling was conducted continuously for 22 hours a day from 10:30 am to 8:30 am the next day. PTFE membranes containing PM 2.5 were cut into 1-cm × 1-cm pieces, immersed in distilled water and oscillated ultrasonically for 45 min for 3 times. After 6 layers of gauze filtration, the samples were vacuum freeze-dried and weighed 18 . The samples were then stored at 4 °C. For the preparation of PM 2.5 eye drops, PM 2.5 samples were diluted in sterile PBS to form a concentration of 5 mg/ mL and then were vortexed ultrasonically. The preservative benzyl bromide was added to two groups of eye drops (PM 2.5 and PBS) with the concentration controlled at 0.005%. The eye drops were kept at 4 °C.
Animal experimental procedure. Ninety mice were divided into three groups randomly (n = 30). The right eyes of each group were treated with the following substances 4 times daily: Group A, negative control (NC); Group B, PBS; Group C, 5.0 mg/ml PM 2.5 . The frequency and doses of PM were determined previously in a pilot experiment. Before treatment, all the mice were confirmed to be free of any ocular diseases. The Schirmer test, fluorescein staining, the tear film break-up time (BUT) test, the inflammatory index, and H&E staining were performed sequentially before and 4, 7, and 14 days after treatment. All the mice were euthanized on day 14, and the eyes were harvested for histological analysis and western blotting.
Tear volume, Fluorescein, and BUT measurement. The phenol red thread tear test with phenol red-impregnated cotton threads (FCI Ophthalmics, Pembrooke, MA, USA) was used to measure the volume of tears on days 0, 4, 7, and 14 post-treatment. The tear volume, fluorescein and BUT were measured as previously described 19,20 . The average of three measurements of each eye was considered as the final readout 21 . The fluorescein score was analyzed as follows: 0, absent; 1, slightly punctate staining less than 30 spots; 2, punctate staining more than 30 spots, but not diffuse; 3, severe diffuse staining but no positive plaques; 4, positive fluorescein plaques.
Evaluation of inflammation. The inflammatory response was visualized using a slit lamp on days 0, 4, 7, and 14 post-treatment, and the inflammatory indices were evaluated as previously described 22  inflammatory index was calculated as the sum of the scores of the following parameters divided by 9: the thickness of the ciliary hyperemia (0: absent; 1: less than 1 mm; 2: 1 to 2 mm; 3: more than 2 mm); the presence of central corneal edema (0: absent; 1: present with visible iris details; 2: present without visible iris details; 3: present without visible pupil); and the presence of the peripheral corneal edema (0: absent; 1: present with visible iris details; 2: present without visible iris details; 3: present with no visible iris). Terminal deoxynucleotidyl transferase-mediated dUTP biotin nick end labeling (TUNEL). A TUNEL assay (KeyGen Biotech, China, Nanjing) was performed according to a modification of a published method 24 . Sections stained without biotinylated dUTP were used as negative controls.

Periodic Acid Schiff (PAS) and Hematoxylin and Eosin
Immunofluorescent staining of Ki67. Immunodetection of Ki67 was performed as described previously 25 . Mouse anti-mouse Ki67 antibody (Abcam, ab16667, Cambridge, MA) was used at a 1:150 dilution as the primary antibody, followed by incubation with ALEXA fluorophore-conjugated secondary antibodies (Invitrogen, USA) and counterstaining with Hoechst 33342 dye (0.5 g/mL, Invitrogen, USA). Images were obtained using a fluorescence microscope (Zeiss, Germany).  24 h at 4 °C. For TEM, the right corneas were harvested and fixed for 2 h in 2.5% glutaraldehyde and 4% paraformaldehyde in PBS (pH = 7.4). The samples were processed as previously described 26 . SEM images were acquired with a scanning electronic microscope (JSM-6330F, JEOL, Japan), while TEM images were acquired with a TEM microscope (JEM2100HC; JEOL, Tokyo, Japan) 27 .
Apoptosis of lacrimal glands. The entire lacrimal glands were dissected and fixed in formalin. Tissue sections of 4-μm thickness were stained using H&E staining and the TUNEL assay.
Image processing and Statistical analysis. Images were processed using Image-Pro Plus 6.0 software (Graphpad Prism, Inc., La Jolla, CA, USA). One-way ANOVA analysis and post hoc analysis were performed for comparisons between groups using SPSS 16.0.0 (SPSS, Chicago, IL). P < 0.05 was considered statistically significant. Data were represented as the mean ± the standard error.

Results
Effect of PM 2.5 on the proliferation and migration of HCECs. CCK-8 assays were used to measure cell viability at different dosages of PM 2.5 (0.1 mg/ml, 0.25 mg/ml, 0.5 mg/ml, 1.0 mg/ml, 2.0 mg/ml, 5.0 mg/ml or 10.0 mg/ml). There was no difference between the viability of the HCECs treated with 1% BSA and 0.1 mg/ml PM 2.5 (P > 0.05). However, PM 2.5 was shown to reduce HCEC viability with increased concentrations compared with samples treated with 1% BSA (P < 0.05, Fig. 1B). In addition, we evaluated HCEC migration when exposed to 1% BSA or 5.0 mg/ml PM 2.5 using the wound scratch method. Both groups showed incomplete healing after 8 hours, whereas cells treated with 5.0 mg/ml PM 2.5 showed less wound healing than cells treated with 1% BSA. At 12 hours after scratching, we found that the migration rate of the HCECs was significantly reduced in the 5.0 mg/ ml PM 2.5 -treated group compared with that in the 1% BSA-treated group (P < 0.05, Fig. 1A,C).

Efficiency of phagocytic uptake in HCECs.
To investigate the efficiency of phagocytic uptake in HCECs, we incubated HCECs with fluorescently labeled latex beads (2.5 µm). The beads were internalized by HCECs and were accumulated at perinuclear regions. We observed that HCECs incubated with LB exhibited fluorescent Physiological conditions of treated mice. Body weights, eyeball weights, and extra-orbital lacrimal gland weights were measured in all groups at 0, 4, 7, 10, and 14 days after inducing dry eye with PM 2.5 ( Fig. 2A). Effects of PM 2.5 on the ocular surface. The inflammatory effects of PM 2.5 or PBS (4 times/day, daily) on the ocular surface within 14 days were evaluated using inflammation scoring, the stability of tear film, and ocular surface PAS and H&E staining. Based on the preliminary pilot experiments, topical application of PM 2.5 at 5.0 mg/ ml 4 times per day for 14 days was determined as the optimal procedure for the induction of dry eye syndrome in the BALB/c mice. Severe ocular surface damage, ulceration, epithelial defects, and neovascularization were observed with higher concentrations of PM 2.5 (10.0 mg/ml), while no obvious effects were observed for PM 2.5 at 5.0 mg/ml (data not shown).

Stability of tear film and epithelial damage.
Before treatment, no significant differences were observed in BUTs and tear film/epithelial damage scores among the three groups. At day 14, the PM 2.5 -treated group showed significantly decreased BUTs compared with that in the PBS-treated group ( * P < 0.05 vs. control, v P < 0.05 vs. PBS, Fig. 3C), whereas fluorescein sodium scores (Fig. 3B) were significantly increased ( * P < 0.05 vs. control, v P < 0.05 vs. PBS). After 14 days of treatment, PBS-treated corneas did not show positive staining of fluorescein sodium (Fig. 3A, upper row of images), and the BUTs and tear film/epithelial damage scores were not changed (P > 0.05, Fig. 3B-E). The tear film/epithelial damage appeared in the PM 2.5 -treated group, probably due to the toxicity of the PM 2.5 ( * P < 0.05 vs. control, v P < 0.05 vs. PBS, Fig. 3A, lower row of images, Fig. 3B-E).
Aqueous tear volume. At day 0, there was no significant difference between the PBS-and PM 2.5 -treated groups. Compared with the vehicle group, the tear volume was decreased rapidly in the PM 2.5 -treated group after 14 days of treatment.  Fig. 3E). In addition, the PM 2.5 -treated group showed more infiltration of inflammatory cells in the central cornea and conjunctiva region than that in the PBS-treated group and negative control group (Fig. 4A). The number of epithelial layers in the central cornea and conjunctiva was significantly decreased in PM 2.5 -treated eyes than in control eyes after 14 days of treatment ( * P < 0.05 vs. control, v P < 0.05 vs. PBS, Fig. 4B,C).  Goblet cell density. We used PAS staining to examine the effect of PM 2.5 on goblet cells in the cornea and conjunctiva. Interestingly, PAS-positive cells were not detected in the cornea in all groups (Fig. 6A). However, the PAS-positive cell number in the conjunctiva was significantly decreased in the PM 2.5 -treated group compared with that in the negative control and PBS treatment groups after 14 days of treatment ( * P < 0.05 vs. control, v P < 0.05 vs. PBS, Fig. 6B).
Apoptosis and cell proliferation. The TUNEL assay showed that apoptosis was induced in the corneal superficial and basal epithelium but not in the stroma in the PM 2.5 -treated group, while few apoptotic cells were observed in the corneal epithelium of the PBS-treated group and negative control group ( * P < 0.05 vs. control, v P < 0.05 vs. PBS, Fig. 7A-C). Compared with the control groups, the immunostaining of Ki67 revealed a drastic increase in Ki67-positive cells in both the central cornea and conjunctiva of the PM 2.5 group after 14 days of treatment ( * P < 0.05 vs. control, v P < 0.05 vs. PBS, Fig. 8A-C), and Ki67-positive cells were mainly located at the basal cell layer of the cornea and conjunctiva. Corneal Epithelial Ultrastructural changes. Corneal epithelium cells have many neatly arranged microvilli and microfolds extending outwards in the negative control (Fig. 9A,B). The corneal epithelium was intact and organized in groups A and B (Fig. 9A-D). By contrast, the epithelial cells were deformed in the cornea of the PM 2.5 -treated eyes (Fig. 9E,F). After PBS treatment for 14 days in group A, transmission electron microscopy revealed enriched regularly arranged microvilli extending from surface epithelial cells (Fig. 9B, upper line). By contrast, after 14 days of PM 2.5 treatment in group C, the number of corneal epithelial microvilli (Fig. 9G) was drastically reduced, the morphology of the microvilli was significantly different to that in the other two groups (all * P < 0.05 vs. control, v P < 0.05 vs. PBS), and most of the microvilli were much shorter and disorganized (Fig. 9E,F).
Apoptosis of lacrimal glands. Lacrimal glands from the NC groups (Fig. 10A) and PBS-treated groups ( Fig. 10D) were similar in volume and color. By contrast, lacrimal glands from PM 2.5 -treated groups (Fig. 10G) were shrunken and lighter in color. H&E staining revealed that the lobe, duct, and acinar of lacrimal glands were maintained properly in both the NC and PBS-treated groups, and no inflammatory cells were observed (Fig. 10B,E). However, irregularly arranged lobes were often found in the PM 2.5 -treated lacrimal glands (Fig. 10H). There were no apoptotic bodies observed by TUNEL assay analysis in all three groups (Fig. 10C,F,I).

Discussion
To the best of our knowledge, this is the first study to evaluate the effects of air pollutant PM 2.5 in inducing dry eye in mice. PM 2.5 , as one of the main components of atmospheric pollutants, has aroused widespread concern with regard to its impact on health. The effect of particles on the human body is related to the diameter of the particle itself 29 . Normally, large-diameter particles will be filtered through nasal cilia and mucus and cannot pass through the nose and throat. Particles with a diameter of less than 10 µm can infiltrate the lungs and bronchial alveolar structures 30 . Particles with a diameter of less than 2.5 µm have a greater penetration potential. They may infiltrate the fine bronchial wall and interfere with gas exchange in the lungs 31 . These particles can eventually enter the blood vessels, where they might impact other parts of the body through blood circulation 32 , making PM 2.5 more detrimental. BURNETT et al. surveyed eight cities in Canada and found that smaller particles cause more serious damage to the human body 33 . In summary, atmospheric particulate matter increases the occurrence rate of mortality, asthma, atherosclerosis, and diabetes [34][35][36][37] . The eye is in direct contact with the outside world, so changes in the external environment will have a major impact on the ocular microenvironment 38,39 . Dry eye is a common disease on the eye surface, and environmental factors are one of the main causes 40 . Normally, the eye surface is covered with tear film, and the stability of tear film is crucial for maintaining ocular health. The stability of tear film depends on the normal components of each layer in the tear film and normal tear dynamics 41 . Environmental factors can affect the composition of the lacrimal film, and it can also affect tear dynamics. In our experiment, we found that BUT and the corneal FL score were increased significantly, while SIT was reduced in the group treated with PM 2.5 for 14 days compared with the control treatment group and negative control group. These results indicate that the stability of the tear film is impaired, which will cause further damage to the cornea. Tiger red staining and fluorescein sodium staining showed corneal epithelial defects in the PM 2.5 -treatment group. H&E staining indicated that the number of corneal and conjunctival epithelial layers was increased drastically in the PM 2.5 -treatment group, and the arrangement was disorganized. We also found that lacrimal gland epithelial cells were reduced after PM 2.5 treatment, while no apoptosis cells were observed. Under the electron microscope, we observed finger-like projections of microvilli of the corneal epithelium. Similar to the negative control group, microvilli were increased in number and were arranged neatly after 14 days of PBS eye drops. By contrast, the corneal epithelial microvilli were reduced in number and became shorter and disordered after treatment with PM 2.5 for 14 days. Mucus protein secreted by goblet cells is one of the components of tear film and a very important factor in maintaining the stability of tear film 42 . In dry eye patients, decreased conjunctival goblet cells result in decreased secretion of mucin, thereby reducing the stability of tear film 43 . This change was verified in our experiments. Similar to those with dry eye, conjunctival goblet cells were significantly fewer in PM 2.5 -treated eyes compared with the PBS-treated group and negative control. PM 2.5 not only affects the stability of tear film but also introduces many toxic and harmful substances 44 . In vitro results have suggested that PM 2,5 can directly impact corneal cellular health because exposure to PM 2,5 affects HCE cell viability. With the increase in the dosage of PM 2.5 , the survival rate of HCE decreases. According to our in vitro experiments, we explored the in vivo effects of PM 2.5 of different concentrations. We found that 0.5-mg/ml PM 2.5 administration did not cause any changes in tear secretion as shown by in situ staining (data not shown). It is possible that PM 2.5 eye drops in animal studies are diluted quickly due to tear secretion and blinking. By contrast, 10 mg/ml PM 2.5 was too irritating in mice. Corneal angiogenesis occurred at approximately 2 weeks (data not shown). Thus, we chose 5 mg/ml as the optimal PM 2.5 eye drop concentration to induce dry eye syndrome in the present study. It would be interesting to investigate the ocular surface damage induced by different PM 2.5 concentrations in future toxicity studies. In addition, the uptake of 2.5-µm diameter latex beads by HCECs suggests that PM 2.5 can be phagocytized into HCECs and intoxicate cells. The results of the scratch experiment showed that PM 2.5 could inhibit the migration and proliferation of HCE. These may be additional mechanisms of PM 2.5 -induced dry eye.
A recent study reported that NF-κB can regulate the gene expression of various cytokines and adhesion molecules involved in the inflammatory response and is closely related to the occurrence of inflammation 45,46 . In addition to its role in inflammation, the NF-κB pathway also affects cell proliferation and anti-apoptosis 47 . Moreover, the activation of the NF-κB pathway includes the involvement of upstream and downstream genes, and the interaction between apoptosis and anti-apoptotic genes 46 .
Existing studies have found that genes involved in the inflammatory response of dry eye are mostly target genes of NF-κB. In a rabbit dry eye model, there is increased expression of NF-κB P65 in the cornea, conjunctiva, and lacrimal gland tissue, indicating the activation of NF-κB. Activated NF-κB enters the nucleus, where it binds to κB (GGGACTTTCC) of the NOS2 target gene promoter to induce transcription and promote the synthesis of Figure 10. Images of lacrimal glands from all three groups (n = 3 in each group). Lacrimal glands from the NC groups (A) and PBS-treated groups (D) maintained the same volume and color. By contrast, lacrimal glands from the PM 2.5 -treated groups (G) were shrunken and became lighter. (H,E) Staining showed similar images of the maintenance of the complete structure of the lobe, duct, and acinar of lacrimal glands in both the NC and PBS-treated groups, and no inflammatory cells were observed (B,E). However, in the PM 2.5 -treated groups, irregularly arranged lobes were found (H, red arrow head). TUNEL assay analysis detected no apoptotic bodies among all three groups (C,F,I).
Scientific RepoRTs | (2018) 8:17828 | DOI:10.1038/s41598-018-36181-x target genes (e.g., TNF). Therefore, the activation of NF-κB may be one of the initiation mechanisms of dry eye 48 . Apoptosis is an important mechanism in the development of dry eye. The apoptosis index of epithelial cells in dry eye is increased 49 . Dry eye is an immune disease, and apoptosis is involved in the development of immune cells, immune regulation, immune effect, and many other physiological and pathological processes 50 . In our experiment, after PM 2.5 treatment, the apoptotic cells in the corneal epithelial cells were significantly increased compared with those in the PBS treatment group and negative control group. Western blotting revealed that the PM 2.5 -treatment group showed higher levels of pNF-κB expression than the other two groups. The ratio of pNF-κB to pan NF-κB was higher than that in the two control groups, indicating that NF-κB is activated, which is further validated by the increase in TNF-α expression. Thus, we hypothesize that the mechanism of PM 2.5 -induced dry eye may be mediated by the activation of NF-κB. It will be of great value to verify the potential relationship between NF-κB activation and the progression of dry eye, which might promote the development of dry eye therapy.
One major caveat of the study is that the environmental exposure of PM2.5 may be different from direct PM2.5 topical administration. We will modify the concentration and the mode of contact in future studies to better simulate PM2.5 environmental exposure.