Dry Eye-Induced CCR7⁺CD11b⁺ Cell Lymph Node Homing Is Induced by COX-2 Activities

Abstract

Purpose.: We aimed to determine the role of CCR7⁺CD11b⁺ cell lymph node (LN) homing and T-cell differentiation in dry eye (DE)-induced immunopathogenesis and investigate the therapeutic effects of cyclooxygenase-2 (COX-2) and prostaglandin E₂/eicosanoid-prostanoid (PGE₂/EP) inhibitors against DE.

Methods.: Six-week-old female C57BL/6 mice were housed in a controlled-environment chamber and administered topical selective COX-2 inhibitors or EP2 antagonists. Expression of major histocompatibility complex (MHC)-II^high, CD11b⁺, CCR7⁺, IFN-γ⁺, IL-17⁺, and CD4⁺ in the corneas and draining LNs was evaluated using flow cytometry. Mixed lymphocyte reactions (MLRs) with carboxyfluorescein diacetate succinimidyl ester labeling and intracellular cytokine staining were used to verify DE-induced corneal dendritic cell function. mRNA expression of COX-2, EPs, and proinflammatory cytokines in ocular surface was evaluated using quantitative RT-PCR and immunohistochemical staining.

Results.: Dry eye significantly increased MHC-II^highCD11b⁺ and CCR7⁺CD11b⁺ cells in the cornea and LNs, and MLR revealed CCR7⁺CD11b⁺ cells from DE corneas stimulated IL-17⁺CD4⁺ cell proliferation. mRNA levels of COX-2, EP2, IFN-γ, TNF-α, IL-6, and IL-17 were significantly higher in DE ocular surface but were suppressed by topical COX-2 inhibitors and EP2-specific blockers. Immunohistochemical staining showed COX-2 and matrix metalloproteinase expression in DE corneal epithelia that was diminished by both topical treatments. Furthermore, both topical treatments significantly reduced frequencies of MHC-II^high, CD11b⁺, and CCR7⁺CD11b⁺ cells in the corneas and LNs, but also IL-17⁺CD4⁺ cells in LNs.

Conclusions.: Topical COX-2/EP2 treatment reduces CCR7⁺CD11b⁺ cells on the ocular surface with inhibition of cellular LN homing and suppresses Th17 immune response, suggesting the COX-2/PGE₂/EP axis contributes to immuno-inflammatory pathogenesis on the ocular surface and may be a novel therapeutic target in DE.

Introduction

Dry eye (DE) is a multifactorial disease suggestively caused by tear film instability, tear hyperosmolarity, and inflammation of the ocular surface.¹ Although DE pathogenesis is not fully understood, previous work suggests immuno-inflammatory reactions with accompanying desiccating stress to the ocular surface are significant. Their role in DE pathogenesis is suggested by observed migration and activation of inflammatory cells, such as dendritic cells (DCs) and T cells, and induction of proinflammatory mediators such as interleukin (IL)-1, IL-6, IL-12, interferon-γ (INF-γ), tumor necrosis factor-α (TNF-α), and prostaglandin E₂ (PGE₂).^2–4 

Prostaglandin E₂ is an endogenous lipid mediator derived from arachidonic acid, which is released in response to inflammatory insults. Recent studies have shown that PGE₂, a major cyclooxygenase-2 (COX-2) metabolite, can function crucially as a tumor promoter or immunomodulator in other organs.^5–7 Prostaglandin E₂ mediates its specific effects through G protein-coupled eicosanoid-prostanoid (EP) receptors (subtypes designated EP1, EP2, EP3, and EP4), leading to the versatility of PGE₂.^8,9 Interestingly, PGE₂ interacts with highly efficient antigen-presenting DCs, which paradoxically have an important role in both immunity and tolerance.^10–12 Recent studies reported PGE₂ exerts potent immunosuppressive effects during DC maturation by inhibiting the release of proinflammatory cytokines and chemokines such as TNF-α, IL-6, CCL3, CCL4, and CXCL10 and upregulating potent anti-inflammatory cytokines such as IL-10.^12–16 In contrast, other studies revealed PGE₂ promotes DC maturation with increased expression of proinflammatory cytokines, major histocompatibility complex (MHC) class II molecules, and certain co-stimulatory molecules (e.g., CD40, CD80, and CD86).^17–19 Moreover, PGE₂ facilitates migration of myeloid DCs through chemokine signaling and matrix metalloproteinase (MMP) induction, as well as promotes chemotactic responsiveness of DCs to CCL19 and CCL21.^20–24 

Our group previously reported increased PGE₂ levels in the tears of DE patients, which correlated with their symptom grades, and we initially considered PGE₂ a chemical nociceptor activator.⁴ However, PGE₂ is a crucial proinflammatory autacoid that leads to activation and sustenance of a chronic positive inflammation loop. The overproduction of EP receptors, COX-2, and PGE₂ has been found in several cancers, autoimmune diseases, and chronic inflammatory diseases. Despite DE's chronic immuno-inflammatory nature, the roles of COX-2 and PGE₂ in its pathogenesis have not been fully determined. In this study, we investigated the contribution of the COX-2/PGE₂/EP axis in DE pathogenesis and evaluated the therapeutic effects of COX-2/EP2 inhibitors using a murine model. 

Methods

Animal Treatment and DE Induction

Six- to eight-week-old female C57BL/6 mice (Charles River Laboratory, Wilmington, MA, USA) were used in accordance with the standards in the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research. The research protocol was approved by the Institutional Animal Care and Use Committee of the Yonsei University College of Medicine. As described previously, DE was induced in the mice by placing them in a controlled-environment chamber (CEC). Seven mice were allocated to each group (control and DE-induced). To achieve maximum ocular surface dryness, mice in the CEC (with a relative humidity below 13%) were given subcutaneous injections of 0.1 mL scopolamine hydrobromide (5 mg/mL; Sigma-Aldrich Chemical Co., St. Louis, MO, USA) three times a day for the duration of the experiment. Dry eye induction and tissue preparation were performed according to a previous protocol.^3,4 Each experiment was repeated three times, and we calculated average values of expression data from more than 12 biological replicates at each time point. 

Treatment Regimen and Tissue Preparation

After 12 days of DE induction and confirming grade 3 corneal erosions or above, the mice were divided into vehicle (distilled water), 0.2 μg/mL celecoxib (190 μM Celebrex; Sigma-Aldrich), or 0.3 μg/mL AH6809 (300 μM; Cayman Chemical Co., Ann Arbor, MI, USA) control, COX-2 inhibitor, or EP2 antagonist treatment groups, respectively. To investigate the effects of celecoxib and AH6809, each substance was administered topically (8 μL, three times a day) for 1 week into eyeballs of DE mice. Mice were killed, and eyeballs and ipsilateral draining lymph nodes (LNs), including submandibular and cervical LNs, were collected. Each tissue was halved and either fixed in 3.7% paraformaldehyde for immunohistochemical (IHC) staining or stored at −70°C for quantitative real time-PCR (qRT-PCR). 

Corneal Erosion Scoring of the Mouse

Corneal erosion grading used in this study was performed as previously described.³ Briefly, at the end of treatment, 5 μL 1% fluorescein (Sigma-Aldrich) was applied to the lower conjunctival sac of the mice, and after 3 minutes, corneal fluorescein staining was examined with slit lamp biomicroscopy. Punctate staining was evaluated using the Oxford Scheme grading system, and we assigned a grade of 0 to 4. 

Periodic-acid Schiff Staining and IHC Staining

Eyeballs were harvested and analyzed by periodic acid Schiff (PAS) and IHC staining. Central vertical plane sections of 6-μm thickness were stained with hematoxylin-PAS counterstaining. Briefly, the slide was exposed periodically to the acid solution for 15 minutes. Then, after rinsing with tap water, Schiff's reagent was exposed for 15 minutes, and counterstaining with hematoxylin was performed. One masked observer analyzed the change of goblet cells in the superior or inferior conjunctiva. The IHC staining method for corneas has been described previously.^3,4 COX-2, MMP-2, and MMP-9 primary antibodies (Rabbit polyclonal anti-mouse, 1:200; Abcam, Inc., Cambridge, MA, USA) was used for IHC staining. Light microscopy (Axio Imager 2; Carl Zeiss, Oberkochen, Germany) was used for examination. 

Tissue RNA Extraction and qRT-PCR

Four to six corneas containing enough limbal area from four to six mice were included in each group (control and DE-induced). The qRT-PCR method for ocular surfaces has been described previously.³ Quantitative real time-PCR was performed using SYBR Premix Ex Taq (Takara Bio, Inc., Otsu, Shiga, Japan) with preformulated primers and StepOnePlus Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). Detailed primer information is described in Supplementary Material S1. 

Flow Cytometry

Twelve corneas and draining cervical and submandibular LNs from six mice were harvested. Single-cell suspensions of corneal and LN samples were prepared by collagenase digestion and blocked with anti-FcR mAb for 30 minutes at 4°C in 1% BSA/0.02% NaN₃/PBS. Isolated cells were then stained with the following antibodies for 45 minutes at 4°C: anti-CD11b PE-Cy5, anti-CCR7 Alexa Fluor 488, and anti-Ia/Ie PE (BioLegend, Inc., San Diego, CA, USA) and analyzed with FACSCanto and FACSAria flow cytometers (BD Biosciences, San Jose, CA, USA). All antibodies were analyzed with appropriate isotype controls. Other LN samples were prepared for intracellular cytokine staining with anti-CD4 PE-Cy7, anti-IFN-γ FITC, and anti-IL-17 Alexa Fluor 647 (BioLegend) according to the manufacturer's instructions. 

T-Cell Proliferation and Activation Assay Using Mixed Lymphocyte Reactions (MLRs)

The impact of DE-induced corneal DCs on the activation and proliferation of T cells was assessed using MLRs. After 12 days of DE induction, 16 corneas from eight mice were harvested. CCR7⁺CD11b⁺ and CCR7⁻CD11b⁺ cells of DE corneas were separately isolated with prepared antibodies by fluorescence-activated cell sorting (FACS) and used as stimulator cells. Plenty of submandibular and cervical LNs from the control group were harvested, and IFN-γ⁻IL-17⁻CD4⁺ T cells were sorted by FACS according to the manufacturer's instructions. Briefly, the sample was prepared for intracellular cytokine staining with anti-CD4 PE-Cy7, anti-IFN-γ FITC, and anti-IL-17 Alexa Fluor 647 (BioLegend). Firstly, CD4⁺ T cells were sorted by FACSAria and then analyzed with lymphocytic gating and intracellular cytokine staining. In order to obtain naïve IFN-γ⁻IL-17⁻CD4⁺ T cells, single positive cells (IFN-γ⁺ or IL-17⁺) and double positive cells (IFN-γ⁺IL-17⁺) were separated from total CD4⁺ T cells as much as possible. For confirmation of the purification, FACS was re-performed with intracellular cytokine staining. Viable cells were enumerated (CCR7⁺CD11b⁺ cells: 1.0 × 10⁴; CD11b⁺ cells: 2.0 × 10⁴; IFN-γ⁻IL-17⁻CD4⁺ T cells: 1.0 × 10⁶) by the trypan blue exclusion assay and set in 1 mL complete Roswell Park Memorial Institute (RPMI) 1640 media with 10% Fetal Bovine Serum (FBS) at 37°C with 5% CO₂. 

To measure T-cell proliferation, responder CD4⁺ T cells were labeled with the fluorescent dye carboxyfluorescein diacetate succinimidyl ester (CFSE; Molecular Probes, Eugene, OR, USA). During each round of cell division, relative fluorescence intensity of the CFSE dye was decreased by half. The dye was kept as a 0.5-mM stock in dimethylsulfoxide (DMSO) and stored at −20°C in a desiccator box. Cells were labeled by diluting the CFSE stock 1000-fold into the cell suspension and incubating them for 10 minutes at 37°C. After CFSE labeling, CCR7⁺CD11b⁺ and CD11b⁺ cells were thoroughly washed and subsequently co-cultured with CFSE-labeled CD4⁺ T cells at a 1:10 ratio (stimulator cells: responder T cells = 1.0 × 10⁴: 1.0 × 10⁵) for 4 days using U-bottom 96-well plates. 

To measure T-cell activation, CCR7⁺CD11b⁺ and CCR7⁻CD11b⁺ cells were washed and co-cultured with responder T cells at a 1:10 ratio (stimulator cells: responder T cells = 1.0 × 10⁴: 1.0 × 10⁵) for 4 days using other U-bottom 96-well plates. After incubation, the cells were stimulated with 50 ng/mL phorbol 12-myristate 13-acetate (PMA) and 1 μg/mL ionomycin (Sigma-Aldrich) for 10 hours with the addition of 0.6 μL/mL GolgiPlug (Brefeldin A; BD Pharmingen, Franklin Lakes, NJ, USA). Responder T cells were collected for subsequent flow cytometry analyses of intracellular IFN-γ and IL-17 expression according to the manufacturer's instructions with anti-IFN-γ FITC and anti-IL-17 Alexa Fluor 647 antibodies (BioLegend) as described above. All antibodies were analyzed with appropriate isotype controls. 

Statistical Analysis

Independent samples t-tests (Student's t-tests) were performed to compare differences between the two groups using SPSS 21.0 (Chicago, IL, USA). One-way ANOVA was used to make comparisons among three or more groups, and Dunnett's test was further used to compare each treated group with the control group. Data are reported as the average ± SEM. A value of P < 0.05 indicated a statistically significant difference between data set averages. 

Results

MHC-II^highCD11b⁺ and CCR7⁺CD11b⁺ Cell Levels Increased in Both Corneas and Draining LNs in DE-Induced Mice

Cell surface expression of MHC-II⁺ and CCR7⁺ was utilized as markers for DC maturation and LN homing. Absolute cell numbers, mean fluorescence intensities (MFIs), and cell frequencies were also determined. After DE induction, the population of MHC-II^highCD11b⁺ and CCR7⁺CD11b⁺ cells increased 1.6-fold and 2.6-fold in the cornea, respectively (Fig. 1A). The cell numbers and MFI also significantly increased in the DE cornea (Figs. 1B, 1C). In draining LNs, MHC-II^highCD11b⁺ and CCR7⁺CD11b⁺ were also increased; specifically, DE led to a 2.3-fold increase in CCR7⁺CD11b⁺ cells (Fig. 1D), and cell numbers and MFI also increased (Figs. 1E, 1F). 

Figure 1

 

Flow cytometric analysis of CD11b⁺MHC-II⁺ CCR7⁺ cells in corneas and draining LNs of DE-induced murine tissue. After C57BL/6 mice were housed in a controlled environment chamber with scopolamine administration for 12 days, corneas and draining LNs (submandibular and cervical) were obtained and flow cytometry was performed. All experiments were performed with six corneas and LNs from each treatment group, and results were compared to non-DE control mice. Upper and lower panels represent cornea and LN cell analysis, respectively. Cell frequencies (A, D), absolute cell counts (B, E), and MFI fold changes (C, F) are shown. Student's t-test: *P < 0.05, **P < 0.01. Error bars indicate the SD. CTL, normal control.

Desiccating Stress Induced IL-17⁺, Not IFN-γ⁺, Expression in Draining LN CD4⁺ Cells

As migratory DCs activate naïve T cells, their responses are determined by the regional LN. Because main Th cell responses in DE-induced conditions are not well understood, we investigated Th1 and Th17 responses controlled by DE-induced mouse draining LNs. Interestingly, after DE induction, the frequency of IL-17⁺CD4⁺ cells significantly increased by 2.5-fold (Fig. 2). However, we observed no significant increase of IFN-γ⁺CD4⁺ cell frequency in DE draining LNs. 

Figure 2

 

Measurement of IFN-γ⁺CD4⁺ and IL-17⁺CD4⁺ T-cell responses in draining LNs of DE-induced murine tissue. Cells were secured from draining LNs 12 days after DE induction. Flow cytometry was performed to measure T-cell responses according to the manufacturer's protocol. All experiments were performed with more than eight LNs from four mice in each treatment condition, and results were compared to non-DE control mice. CTL, normal control.

DE-Induced Corneal CCR7⁺CD11b⁺ Stimulated CD4⁺ Proliferation With IL-17 Activation

Since we confirmed increased MHC-II^high and CCR7⁺ cell populations with upregulation of Th17 in draining LNs, T-cell proliferation and responses by DE-induced DC maturation were investigated using MLR; Figure 3A illustrates the experiments. Briefly, CFSE-labeled IFN-γ⁻IL-17⁻CD4⁺ cells were co-cultured with CCR7⁺CD11b⁺ cells, and fluorescence intensity of the CFSE was then measured with flow cytometry. CFSE intensity of CD4⁺ T cells under co-culture conditions was significantly diluted due to the progressive halving of CFSE within daughter cells following each cell division, compared to those under nonstimulated conditions. Also, T-cell proliferation was enhanced by mixed culturing with CCR7⁺CD11b⁺ cells in PMA/ionomycin stress-induced condition (Fig. 3B). With regards to cytokine expression, IFN-γ⁻IL-17⁻CD4⁺ cells shifted more into IL-17⁺CD4⁺ T-cell expression after co-culture with DE corneal CCR7⁺CD11b⁺ cells, compared to those incubated with only PMA/ionomycin/GolgiPlug or stimulated CCR7⁻CD11b⁺ T cells (Fig. 3C and Supplementary Material S2). 

Figure 3

 

Contribution of CCR7⁺CD11b⁺ cells to naïve T-cell proliferation and activation. (A) Schematic illustration for MLRs used in this study. After DE induction, 16 corneas from DE-induced group were secured, and CCR7⁺CD11b⁺ and CCR7⁻CD11b⁺ cells were isolated by cell sorting FACS analysis. IFN-γ⁻IL-17⁻CD4⁺ T cells were harvested from submandibular and cervical LNs of normal condition group and sorted by FACS analysis. To measure T-cell proliferation, responder CD4⁺ T cells were labeled with CFSE, and each CCR7⁺CD11b⁺ and CCR7⁻CD11b⁺ cell was thoroughly washed and subsequently co-cultured with CFSE-labeled CD4⁺ T cells. To measure T-cell activation, other samples after 5 days of incubation were stimulated with PMA/ionomycin for 10 hours with GolgiPlug. Responder T cells were collected for subsequent flow cytometry analyses of intracellular IFN-γ and IL-17 expression. (B) Proliferating CD4⁺ T cells were measured by flow cytometry. (C) After permeabilization, intracellular cytokine levels were measured with anti-mouse IL-17-FITC and anti-mouse IFN-γ-PE antibodies.

COX-2, EP2, and Proinflammatory Cytokines Are Overexpressed in DE Ocular Surface

Because COX-2/PGE₂ induces DC maturation in many pathological conditions,^8,25–27 we investigated COX-2 expression in ocular surface of DE-induced mouse. mRNA expression of COX-2 and EP2 was significantly elevated in ocular surface of DE mice compared to normal mice (Figs. 4A, 4B), and EP3 expression was slightly elevated, although not significantly. However, EP1 and EP4 expression was not upregulated. With additional qRT-PCR assays for proinflammatory cytokine expression, we observed upregulation of IFN-γ, TNF-α, IL-6, and IL-17 in DE ocular surface (Fig. 5). 

Figure 4

 

Quantitative RT-PCR analysis of COX-2 and EP receptors in DE-induced murine ocular surface. After 12 days of DE induction, six to eight corneas containing enough limbal area from each group were secured and prepared for qRT-PCR, and mRNA levels of (A) COX-2 and (B) prostaglandin E receptor (EP) subtypes were determined. (C, D) After DE induction, topical COX-2 or AH6809 was administered three times a day, and measured mRNA levels of COX-2 and EP receptors were measured again. One-way ANOVA: *P < 0.05, **P < 0.01. Error bars indicate the SD. AH6809, EP2 antagonist; CLX, Celecoxib; CTL, normal control; V, vehicle.

Figure 5

 

Quantitative RT-PCR analysis of proinflammatory cytokine expression in DE-induced murine ocular surface after topical COX-2 inhibition or EP2 blocking treatment. After 12 days of DE induction, 0.2 μg/mL (190 μM) celecoxib or 0.3 μg/mL (300 μM) AH6809 was topically administered to the ocular surface. mRNA was extracted six to eight corneas containing enough limbal area from each group after enucleation and used for qRT-PCR of cytokines IFN-γ, TNF-α, IL-1β, −6, −12, and −17. One-way ANOVA: *P < 0.05, **P < 0.01. Error bars indicate the SD. AH6809, EP2 antagonist; CLX, Celecoxib; CTL, normal control; V, vehicle.

To investigate the inhibitory role of a COX-2 antagonist on the expression of EPs, a selective COX-2 inhibitor (celecoxib) and EP2 antagonist (AH6809) were administered topically to murine DE ocular surface. Celecoxib treatment reduced COX-2 and EP2 receptor mRNA levels (Fig. 4C). Unexpectedly, topical AH6809 treatment increased EP4 expression 1.8-fold compared to the control (Fig. 4D), and mRNAs for proinflammatory cytokines, which may contribute to DC maturation, were suppressed after either topical treatment (Fig. 5). 

Topical COX-2 Inhibition and EP2 Blocking Reduced MHC-II^high and CCR7⁺ Expression in CD11b⁺ Cells During DE

We explored the role of COX-2 expression in DC maturation during DE by measuring MHC-II^highCD11b⁺ and CCR7⁺CD11b⁺ cells after celecoxib or AH6809 treatment of the ocular surface and draining LNs. After topical treatment, frequencies of MHC-II^highCD11b⁺ and CCR7⁺CD11b⁺ cells were significantly decreased in the cornea (Fig. 6A), as well as their absolute cell numbers and corresponding MFIs (Figs. 6B, 6C). Interestingly, the MFI of CCR7 was reduced to approximately 50% of control levels by celecoxib or AH6809 treatment. Next, we measured the same cell populations draining LNs. After topical celecoxib treatment, the frequency of MHC-II^highCD11b⁺ and CCR7⁺CD11b⁺ cells were significantly decreased in LNs, and specifically, CCR7⁺CD11b⁺ cell frequency was downregulated to 50% of the vehicle treatment group frequency. We compared the inhibitory effects of topical celecoxib and AH6809 treatment and observed that celecoxib, but not AH6809, induced statistically significantly reduced cell frequencies, numbers, and MFI related to MHC-II^high and CCR7⁺ expression in draining LNs (Figs. 6D, 6E, 6F). 

Figure 6

 

Flow cytometric analysis of CD11b⁺, MHC-II⁺, or CCR7⁺ cells in corneas and draining LNs in DE-induced murine corneas after topical COX-2 inhibition or EP2 blocking treatment. Corneas and ipsilateral draining LNs (submandibular and cervical) were secured after 12 days of DE induction and 7 days consecutive topical treatment with celecoxib or AH6809. Flow cytometry was performed according to the manufacturer's protocol. Upper panels (A, B, C) represent corneal cell assay data, and lower panels (D, E, F) represent LN cell assay data. All experiments were performed with six corneas and LNs from each treatment group, and results were compared to the vehicle group. Cell frequencies (A, D), cell numbers (B, E), and MFI values (C, F) are presented. One-way ANOVA: *P < 0.05, **P < 0.01. Error bars indicate the SD. AH6809, EP2 antagonist; CLX, Celecoxib; V, vehicle.

Topical COX-2 Inhibition and EP2 Blocking Reduced IL-17⁺CD4⁺ Cells Infiltration in Draining LN Against DE

We investigated the immunosuppressive effect of topical COX-2 inhibitor and EP2 antagonist by measuring lymph nodal T-cell activation. Desiccating stress led to a 2.3-fold increase in IL-17⁺CD4⁺ cells of draining LN, but topical celecoxib treatment significantly downregulated to the control level. Topical AH6809 also decreased the frequency of IL-17 expressing cells as compared with vehicle treated group. IFN-γ⁺CD4⁺ cell frequency in DE-induced LN had no significant change with both treatments (Fig. 7). 

Figure 7

 

Measurement of IFN-γ⁺CD4⁺ and IL-17⁺CD4⁺ T-cell responses in draining LNs of DE-induced murine tissue after topical COX-2 inhibition or EP2 blocking treatment. Draining LNs (submandibular and cervical) were secured after 12 days of DE induction and 7 days consecutive topical treatment with celecoxib or AH6809. Flow cytometry was performed to measure T-cell responses according to the manufacturer's protocol. All experiments were performed with eight LNs from four mice in each treatment group, and results were compared to the vehicle group. AH6809, EP2 antagonist; CLX, Celecoxib; CTL, normal control; V, vehicle.

Topical COX-2 Inhibition and EP2 Blocking Improved Ocular Surface Deterioration in DE

Lastly, we examined ocular surface changes and histological findings related to COX-2 inhibition. The increased corneal erosion score induced by DE was significantly reduced by both celecoxib and AH6809 treatments (Fig. 8A). As observed in our previous study,⁴ COX-2, MMP-2, and MMP-9 were strongly expressed during DE. Interestingly, areas of MMP-2 and MMP-9 expression were co-localized with COX-2 expression areas (Fig. 8B). However, IHC analysis revealed treatment with either celecoxib or AH6809 reduced COX-2, MMP-2, and MMP-9 expression in peripheral cornea. Additionally, DE-induced loss of conjunctival goblet cells was reversed, and corneal erosion was improved by each topical treatment (Figs. 8A, 8B). 

Figure 8

 

Corneal erosion and histological findings of the ocular surface after treatment with celecoxib or AH6809. (A) Representative micrographs of fluorescein stained-DE mouse corneas with mean erosion scores for each group shown on the left. (B) Representative micrographs of PAS staining of tissues from each treatment group and immunohistochemical staining of COX-2, MMP-2, and MMP-9 on the ocular surface. One-way ANOVA: *P < 0.05, **P < 0.01. AH6809, EP2 antagonist; CLX, celecoxib; CTL, normal control; V, vehicle.

Discussion

The novel findings of the present work are the following: (1) CCR7⁺CD11b⁺ and MHC-II^highCD11b⁺ cells are expressed in DE induced ocular surface, migrate into draining LNs, and activate Th17 response. (2) COX-2/PGE₂/EP2 activation was an important role in activation of DE induced DCs (e.g., CCR7⁺CD11b⁺). Additionally, inhibition of COX-2 or EP2 activation effectively reduces MHC-II^highCD11b⁺ and CCR7⁺CD11b⁺ cells in both DE cornea and LNs, improves proinflammatory cytokine milieu in cornea, and suppresses Th17 immune response, resulting in amelioration of ocular surface deterioration in murine DE model. 

Acquisition of CCR7 on DCs on the Ocular Surface in Murine DE Model

CCR7 confers distinct migratory properties on leukocytes under both homeostatic and inflammatory states.^28,29 Interactions between the receptor and its ligands CCL19 and CCL21 interaction is now recognized for its role in promoting migration of DCs from the affected tissue to the LN paracortex.³⁰ Specifically, CCR7 has been implicated in the spontaneous development of Sjögren's syndrome-associated ocular changes seen in thrombospondin-1-deficient mice.³¹ Also, Deutsch et al.³² reported CCR7 expression was responsible for B cells homing to secondary lymphoid tissue in parotid gland of Sjögren's patients. However, in non-Sjögren's DE (the most common type of DE), no published studies have explored the role of CCR7 in DE pathophysiology. 

The cornea is populated with nonlymphoid DCs, although their lineage and precise role are unclear. Hattori et al.³³ reported that Langerhans cells are found in the corneal epithelium, whereas nonlymphoid CD11b⁺ DC subsets are in the corneal stroma in normal condition.^2,30,34–36 Relevant to DE, there are multiple evidences supporting a role for corneal DCs, such as CD11b⁺ or CD11c⁺ cell.^36–38 CD11b⁺ DCs are specialized in presenting soluble antigen to T cells and inducing proinflammatory cytokines.³⁵ Meanwhile, CD11c is a cell surface molecule with a broad expression, since it is widely expressed and inducible during inflammation in macrophages, granulocytes, DCs, and T cells in the mouse.³⁹ With this knowledge, we evaluated the change of migratory CD11b⁺ cell population in both cornea and LNs with DE. We observed 18.5% of CD11b⁺ cells in corneas of DE-induced mice co-expressed CCR7 and found an associated increase of CD11b⁺, MHC-II^high, and CCR7⁺ cells in the LNs of these mice. However, 44.9% of CD11b⁺ cells showed MHC-II expression, and only half of these cells expressed CCR7. Although CCR7 plays a role for cells to enter lymphatic vessels, only a small fraction of CCR7⁺ cells may actually home to LNs and activate T cells. The mechanism for CCR7 acquisition in DE-induced DCs should be further investigated. 

LN Homing of DE-Induced Ocular Surface CCR7⁺CD11b⁺cells Promotes Th17 Responses

DCs that populate the ocular surface are important contributors to DE immunopathogenesis,³⁰ which is supported by the observation that pathogenic T cells promote development and perpetuation of DE disease.^40,41 Numerous reports have also demonstrated T cells' central role in pathogenesis in mice under desiccating induced-stress.^42–44 However, although many studies have documented DC maturation by showing increased expression levels of MHC class II and co-stimulatory molecules, no studies have explored the interaction between DE-induced CCR7⁺CD11b⁺ cells and T cells in DE.^2,45,46 

To address this question, we performed T-cell proliferation and activation assays using MLR. We observed that DE-induced corneal CCR7⁺CD11b⁺ cells effectively activate IFN-γ⁻IL-17⁻ T cells and stimulate their differentiation into Th17, rather than Th1 cells. In comparison with T-cell differentiation of CCR7⁻ DCs in MLR, CCR7⁺ DCs have a certain role in generation of Th17 immune response in DE. Chen et al.⁴⁵ demonstrated desiccation stress imparted greater Th17 cell activity than acetylcholine receptor blocking, and additional studies by De Paiva et al.⁴⁷ and Chauhan et al.⁴⁸ suggested the Th17 response plays a crucial role in disrupting the corneal epithelial barrier in DE and that desiccation stress stimulated an influx of IL-17⁺ cells in peripheral corneas with Th17-inducing factors.⁴⁷ These reports highlight the important role of CCR7⁺CD11b⁺ DCs in Th17 activation during desiccation-induced DE. Therefore, blocking CCR7-induced DC maturation and migration and COX-2/PGE₂/EP2 upregulation is a critical step for inhibiting Th17 pathway initiation. 

COX-2/PGE₂/EP2 Axis Enhances DC Maturation With CCR7 Expression in DE

COX-2 and PGE₂ are well-known inducers of immunological and inflammatory diseases. Rieser et al.¹⁸ and Landi et al.²⁵ reported that PGE₂ can hasten DC maturation, and other studies showed PGE₂/EP receptors can act as proinflammatory cytokine amplifiers.^26,27 We found that COX-2 inhibition improved DE-induced ocular surface damage and led to reduced inflammatory cytokine expression. Topical COX-2 inhibition or EP2 blocking also effectively suppressed CCR7⁺CD11b⁺ cell recruitment and LN homing and caused a significant reduction in cytokine expression. Kalinski⁴⁹ suggested PGE₂ may enhance DC migration by upregulating chemokine receptor CCR7 expression on DCs, and another study showed PGE₂ enhances DC surface CCR7 expression but suppresses CCL19 production.²³ These data support our findings and explain the important role of COX-2/PGE₂ in DE-induced DC activation and maturation, as well as enhanced inflammatory cytokine expression during T-cell responses. Therefore, COX-2 and PGE₂ may be promising ocular surface targets for therapeutic agents against DE. 

Interestingly, suppression of CCR7⁺CD11b⁺ cells on the ocular surface was observed after either celecoxib or AH6809 treatment given at similar molarities, suggesting EP2 may be the most important receptor subtype for CCR7⁺ cell recruitment and LN homing in the COX-2/PGE₂ activated pathway. This finding supports our proposal that EP2-specific blocking is a good alternative target for treating DE, in addition to COX-2 inhibition. However, unexpectedly, we found topical EP2 antagonist treatment increased mRNA levels of EP4 in DE-induced corneas. Millard et al.⁵⁰ demonstrated a decreased maximum cAMP response of EP4 after inhibiting the accumulation of cAMP from EP2 activation by using an EP2 antagonist. Receptor EP4 mRNA levels may increase due to a reduced cAMP threshold and a positive feedback loop. Elevated EP4 can lead to increased vascular endothelial growth factor receptor-2 (VEGFR2) gene transcription, as EP2 and EP4 agonists stimulate VEGF pathway activity in murine cochleas.⁵¹ Therefore, EP2 antagonists may have insufficient immunosuppressive effects compared to COX-2 inhibition due to EP4-VEGFR2 upregulation during lymphangiogenesis. 

In conclusion, COX-2 and PGE₂ are promising targets for DE treatment because their activation contributes to CCR7⁺ cell infiltration with Th17 response and proinflammatory cytokine expression on the ocular surface and draining LNs following desiccating stress. Thus, their inhibition could reduce DE pathogenesis and symptom severity. However, the origin and recruitment of CCR7⁺ cells, the mechanism of CCR7 acquisition, and cell interactions in the lymphatic vessels remain unclear. These events, as well as the cellular characteristics that induce a Th17-shifted response during DE, should be investigated in the future. 

Acknowledgments

Supported by Advanced Science Research Program (Grant NRF-2012R1A2A2A02009081) through the National Research Foundation of Korea (NRF), funded by the Ministry of Education, Science, and Technology and partially by a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (Grant HI13C0055). The authors alone are responsible for the content and writing of the paper. 

Disclosure: Y.W. Ji, None; Y. Seo, None; W. Choi, None; A. Yeo, None; H. Noh, None; E.K. Kim, None; H.K. Lee, None 

Footnotes

 YWJ and YS contributed equally to the work presented here and should therefore be regarded as equivalent authors.