Attenuation of Laser-Induced Choroidal Neovascularization by Blockade of Prostaglandin D₂ Receptor 2

Abstract

Purpose: The purpose of this study was to investigate the impact of prostaglandin D₂ (PGD₂) receptor 2 (DP2) on choroidal neovascularization (CNV) formation in mice.

Methods: Using a laser-induced CNV model, the CNV size of wild-type (WT) mice treated with DP2 antagonist (CAY10471 or OC000459) was compared with that of untreated mice. Vascular endothelial growth factor (VEGF) and MCP-1 levels were also compared between the two groups. Similar experiments were performed comparing DP2 knockout (DP2KO) mice with WT mice (8 and 56 weeks old). The number of infiltrating macrophages to laser spots was also compared between the WT and DP2KO mice. We administered a DP2 antagonist to 15-methyl PGD₂ (a DP2 agonist)-stimulated ARPE-19 cells and measured VEGF secretion by enzyme-linked immunosorbent assay. Tube formation assay was performed on human umbilical vein endothelial cells with or without a DP2 antagonist.

Results: CNV sizes were significantly smaller in mice treated with CAY10471 or OC000459 than in those treated with vehicle. Similarly, the CNV size of DP2KO mice was significantly smaller than that of WT mice. The number of macrophages at laser spots in DP2KO mice was significantly lower than that in WT mice. The VEGF concentration of lasered DP2KO mice's eyes was significantly lower than that of lasered WT mice’ eyes. DP2 antagonist treatment suppressed VEGF secretion in ARPE-19 cells under 15-methyl PGD₂ stimulation. The tube formation assay suggested that lumen formation was inhibited by a DP2 antagonist.

Conclusions: DP2 blockade attenuated choroidal neovascularization.

Translational Relevance: Drugs targeting DP2 are potentially a novel treatment for age-related macular degeneration.

Introduction

Age-related macular degeneration (AMD) is a leading cause of visual impairment worldwide. Neovascular AMD is associated with choroidal neovascularization (CNV) and the proliferation of blood vessels from the choriocapillaris into the subretinal space. CNV is a result of a combination of inflammatory and angiogenic processes.^1,2 

Vascular endothelial growth factor (VEGF) plays an essential role in CNV formation in patients with neovascular AMD. Vitreous injection of anti-VEGF antibodies has been used as a first-line treatment for neovascular AMD^3,4; however, anti-VEGF therapy is expensive and requires repeated vitreous injections, which is a heavy burden for patients. In addition, some patients do not respond well to anti-VEGF treatment,^5,6 suggesting the importance of identifying new therapeutic targets. 

Cyclooxygenase (COX), known as prostaglandin-endoperoxide synthase, is an enzyme responsible for the formation of thromboxane and prostaglandins from arachidonic acid. COX is involved in the inflammatory immune response, and some studies have reported that COX2 is expressed in retinal pigment epithelium (RPE) cells and is associated with CNV formation by regulating the expression of VEGF.^7–9 Furthermore, COX2 antagonists suppress macrophage infiltration and downregulate VEGF expression in the RPE-choroid complex in a mouse CNV model.¹⁰ However, it remains unclear what kind of prostaglandin is implicated in the regulation of VEGF expression. 

Prostaglandin D₂ (PGD₂) is a COX metabolite that promotes sleep¹¹ and mediates inflammatory responses.¹² Virtue et al. reported that hematopoietic prostaglandin D synthase (H-PGDS)-produced PGD₂ polarized macrophages toward an M2 anti-inflammatory state.¹³ H-PGDS was found to be expressed predominantly in the macrophage fraction.¹³ 

In a murine macrophage cell line (RAW264.7), the COX2 inhibitor, CAY10404, inhibited lipopolysaccharide-mediated migration.¹⁴ Another report suggested that the lipopolysaccharide-mediated migration was inhibited in peritoneal macrophages of PGD₂ receptor knockout mice and H-PGDS-deficient mice.¹⁴ 

In the present study, we focused on the PGD₂ receptor, DP2. We investigated the effectiveness of DP2 antagonists in a laser-induced CNV mouse model. 

Methods

Animals

Male C57BL/6 mice (8 weeks old and 56 weeks old) were purchased from Japan SLC Inc. (Shizuoka, Japan) and used for experiments in this study in order to avoid the gender difference, as aged female mice had significantly larger CNV than age-matched males.¹⁵ DP2 knockout (DP2KO) mice on a C57BL/6 background were obtained as described previously¹⁶ and maintained in Oriental Yeast Co. (Tokyo, Japan). DP2KO mice developed normally.¹⁷ 

All animal experiments were performed according to the ethical guidelines for animal experimentation of the University of Tokyo and the Association for Research in Vision and Ophthalmology statement for the Use of Animals in Ophthalmic and Vision Research. 

Induction of CNV

CNV induction was performed as previously described.¹⁸ Mice were anesthetized by an intraperitoneal injection of a mixture of ketamine (80 mg/kg; Daiichi Sankyo Propharma Co., Ltd., Tokyo, Japan) and xylazine (16 mg/kg; Bayer, Leverkusen, Germany) in sterile saline. Laser photocoagulation was performed using a diode laser (DC-3300, Nidec Co., Ltd., Aichi, Japan) at a spot size of 50 µm, power of 150 mW, and duration of 50 ms. The laser spots were located from two- to three-disc diameters from the optic nerve avoiding the major vessels. Four to eight laser spots were generated per eye for evaluating CNV size, macrophages, and immunohistochemistry. In contrast, 12 laser spots were generated per eye for enzyme-linked immunosorbent assay (ELISA) of VEGF and MCP-1 as described below.¹⁹ 

Measurement of CNV Size and Immunohistochemistry

Seven days after laser photocoagulation, the mice were anesthetized and perfused through the left ventricle with 1 mL of fluorescein concanavalin A (FL-1001; Vector Laboratory, Burlingame, CA) diluted 10 times with phosphate-buffered saline (PBS). The mice were euthanized, and their eyes were enucleated and fixed in 4% paraformaldehyde for 20 minutes. The RPE-choroid-sclera were flat-mounted on slides with mounting medium (VECTASHIELD HardSet Mounting Medium; Vector Laboratory, Burlingame, CA) and coverslips. 

Images of CNV were captured using a fluorescence microscope, BZ-9000 (Keyence, Osaka, Japan). The area of CNV lesions was measured manually using ImageJ software (National Institutes of Health, Bethesda, MD) in a masked fashion. 

DP2 Antagonist Administration

The DP2-specific antagonists, 5-fluoro-2-methyl-3-(2-quinolinylmethyl)-1H-indole-1-acetic acid (OC000459, Cayman Chemical, #12027) and (+)-3-[[(4-fluorophenyl)sulfonyl]methylamino]-1,2,3,4-tetrahydro-9H-carbazole-9-acetic acid (Cay10471, Cayman Chemical, #10006735), were dissolved in 1:100 DMSO in PBS to a concentration of 2 mg/mL. 

Alzet pumps were injected with a solution of OC000459 or Cay10471, as well as a vehicle (1:100 DMSO in PBS) and were implanted subcutaneously in the neck at 1 day before the laser. The Alzet osmotic pump gradually released the solution of OC000459, Cay10471, or vehicle for 7 days (2 mg/mL was equivalent to 1.2 mg/day/kg). At the end of the study (day 7), we ensured that the pumps were empty. 

Immunohistochemistry

Three or 7 days after laser photocoagulation, the eyes were enucleated and embedded in an optimal cutting temperature compound (OCT, Sakura, Kobe, Japan) and frozen in liquid nitrogen. Frozen sections of 10-µm thickness were prepared and fixed with 4% paraformaldehyde for 20 minutes. After being blocked with Blocking One Histo (Nacalai Tesque, Inc., Kyoto, Japan, #06349-64), the sections were incubated with rat anti-mouse F4/80 (1:100 dilution, MCA497, Bio-Rad) and rabbit anti-mouse DP2 (1:50 dilution, ab235830, Abcam) antibodies. Sections were then incubated for 1 hour with goat anti-rat IgG (H+L) secondary antibody, Alexa Fluor 594 conjugate (1:250, Invitrogen, Carlsbad, CA, Product # A-11007), and goat anti-rabbit IgG (H+L) secondary antibody, Alexa Fluor 488 (1:250, Invitrogen, Carlsbad, CA, Product #ab150077), and then washed and examined using a fluorescence microscope (BZ-9000, Keyence, Osaka, Japan). Counterstaining was performed using DAPI (1:1000, Vector Laboratories, Peterborough, UK). 

Enzyme-Linked Immunosorbent Assay

Three days after the laser treatment, the mice were euthanized. The whole eye was homogenized in a 500 µL RIPA buffer with a protease inhibitor cocktail, and centrifuged (5000 g × 5 minutes). The supernatant was used for measurement. 

To evaluate the expression of VEGF and MCP-1, the Quantikine Mouse Immunoassay Kit (R&D Systems MMV00 and MJE00, respectively) was used according to the manufacturer's instructions. 

Each absorbance was measured with a microplate reader (2030 ARVO X3, Perkin Elmer Japan, Kanagawa, Japan). The results are expressed as pg/mL protein. 

Measurement of Macrophage Recruitment into CNV

To investigate the influence of DP2-deficiency and DP2 antagonists on macrophages, which are believed to play an important role in CNV formation, macrophage recruitment into CNV lesions was measured in wild-type (WT) and DP2KO mice, as previously described^20,21 with some modifications. The right eyes of male adult (8 weeks old) WT and DP2KO mice were treated with laser, as described above. Four to six shots of laser photocoagulation were delivered to each eye. Three days after the laser photocoagulation, all mice were euthanized. Their eyes were enucleated and fixed in 4% paraformaldehyde. The RPE–choroid–sclera complex flatmounts were made and immunostained with 1:100 rat anti-mouse F4/80 antibody (1:100 dilution) and rabbit anti-mouse DP2 antibody (1:50 dilution), as described above. F4/80-positive cells inside or around each laser photocoagulation site were counted using ImageJ (National Institutes of Health, Bethesda, MD). The laser spots were excluded when their sizes were five times the mean CNV size or when their size was one-fifth the mean CNV size, as previously described.²² 

Expression of DP2 in ARPE-19 and Human Umbilical Vein Endothelial Cells by Reverse Transcription Polymerase Chain Reaction

The expression of DP2 in ARPE-19 and human umbilical vein endothelial cells (HUVECs) were examined by reverse transcription polymerase chain reaction (RT-PCR). Total RNA was extracted from ARPE-19 and HUVECs using TRI Reagent (Cosmo Bio, Carlsbad, CA, #TR118). The cDNA was synthesized using random hexamer primers and reverse transcriptase ReverTra Ace real-time quantitative polymerase chain reaction (RT-qPCR) master mix with gDNA remover (Toyobo Co., Ltd., Osaka, Japan). The synthesized cDNA was amplified using Takara Ex Taq (Takara Bio) and various primers (human DP2 forward: TGGACACGTGGTGCATTTTG, backward: TGCATACAGGCACAATCCTAGG, size 124 bp human GAPDH forward: AATTCCATGGCACCGTCAAG, reverse: ATCGCCCCACTTGATTTTGG, size 104 bp) in a Thermal Cycler Dice Real System II TP900 (Takara Bio Inc.) under the following reaction conditions: 

• 1) 95°C for 1 minute as 1 cycle
• 2) 35 cycles of 98°C for 30 seconds, 60°C for 30 seconds, and 72°C for 30 seconds
• 3) One cycle of 72°C for 5 minutes

The amplified cDNA fragments were then electrophoresed in 2% agarose gel for 20 minutes, and their expression was examined. 

ARPE-19 Cell Culture and Cytokine Measurement

ARPE-19 cells were cultured in DMEM/F-12 (Sigma–Aldrich, St. Louis, MO) supplemented with 10% fetal bovine serum (Life Technologies Corporation, Carlsbad, CA) and 1% penicillin-streptomycin solution (Gibco; Thermo Fisher Scientific, Inc.) at 37°C in 5% CO2/95% air. The culture medium was replaced every 48 hours. In the experiments, cells were plated on 6-well plates (Nunc, Rochester, NY) at a density of 100,000 cells/well in 2 mL of serum-containing medium and incubated for 5 days. Confluent cell cultures were washed with serum-free DMEM/F-12 medium and treated with or without 15 µM CAY10471 in a serum-free medium and after 2 hours of incubation in the presence or absence of 15 µM 15(R)-15-methyl PGD₂, a stable DP2-selective agonist (Cayman Chemical, #12720). After 22 hours of incubation, the medium was collected for cytokine measurements. The concentrations of VEGF and MCP-1 were measured using ELISA kits (R&D Systems, DVE00 and DCP00, respectively) according to the manufacturer's instructions. 

HUVEC Culture and Tube Formation Assay

To investigate the effect of the DP2 antagonist on the differentiation and proliferation of vascular endothelial cells, we performed a tube formation assay of HUVECs. The HUVECs (p2) were placed in 96-well plates with polymerized BME at the bottom, at a density of 2 × 10⁴ cell/50 µL PBS/ well. Vehicle (DMSO), 1 µM of CAY10471, 10 µM of CAY10471, and 100 µM of CAY10471 were administered at 50 µL/well each. 

Cells were incubated at 37°C in 5% CO2/95% air, and the number of branch points was measured after 18 hours. 

Statistical Analysis

All data are presented as the mean ± standard error of the mean (SEM). 

Differences between the two groups were compared using the Student's t-test. Differences among the three groups were analyzed using 1-way ANOVA followed by Tukey's post hoc test. Statistical significance was set at P < 0.05. All statistical analyses were performed using the statistical programming language “R” (R version 3.5.1; The R Foundation for Statistical Computing, Vienna, Austria). 

Results

Frozen sections and flatmounts of eyes at 3 days after laser photocoagulation were immunostained with antibodies against DP2 (green), F4/80 (red), and DAPI (blue; Fig. 1). Co-localization of DP2 and F4/80 (yellow in merged views) was observed in many macrophages, suggesting that DP2 is expressed in macrophages. 

Figure 1.

 

Immunostaining of laser coagulation sites at 3 days after CNV induction with DP2 (green), F4/80 antibody (red), and DAPI (blue) on eye flatmounts (A) and cryosections (B). Colocalization of DP2 and F4/80 staining (yellow in merged views) indicated the presence of macrophages expressing DP2. Arrows showed the center of laser photocoagulation. Arrow heads showed DP2 and F4/80 positive macrophages.

Using laser-induced CNV model, the CNV size was compared among CAY10471 (2 mg/L, n = 49 spots of 9 eyes), OC000459 (2 mg/mL, n = 79 spots of 16 eyes), and vehicle groups (n = 74 spots of 17 eyes). We found that CNV sizes in the CAY10471 and OC000459 groups were significantly smaller than those in the control group (*P < 0.05 and **P < 0.01, respectively; 1-way ANOVA followed by Tukey's post hoc test; Fig. 2). 

Figure 2.

 

The CNV sizes in vehicle, CAY10471 (2 mg/L), or OC000459 (2 mg/mL) administered mice at 7 days after laser photocoagulation. The CNV sizes are shown in 3 groups with subcutaneous administration of vehicle or DP2 antagonists using Alzet pumps from 1 day before CNV induction. Lower photographs show typical fluorescent images of CNV in single laser photocoagulation locus of each group, in which the CNV area was circled by a thin white dotted line.

Furthermore, CNV size was compared between 8-week-old DP2KO and WT mice. CNV sizes were 20,534.7 ± 1621.0 µm² (n = 83 spots of 21 eyes) in DP2KO mice and 29,275.3 ± 2870.9 µm² (n = 74 spots of 17 eyes) in WT mice. There was a significant difference in CNV size between the two groups (**P < 0.01; Student's t-test; Fig. 3A). A similar result was observed when comparing 56-week-old DP2KO and WT mice. The CNV size in DP2KO mice was significantly smaller than that in WT mice (27,615.3 ± 4981.5, [n = 32 spots of 10 eyes] vs. 69,557.3 ± 8803.8 µm² [n = 26 spots of 10 eyes], ***P < 0.001; Student's t-test; Fig. 3B). The CNV size of aged WT mice was 2.3-fold larger than that in 8-week-old animals, whereas the CNV size of aged DP2KO mice was only 1.3-fold larger than that of 8-week-old animals, suggesting that DP2 is also involved in the age-related exacerbation of CNV. 

Figure 3.

 

The CNV sizes in 8-week-old WT and DP2KO mice (A) and 56-week-old WT and DP2KO mice (B). Lower photographs show typical fluorescent images of CNV in single laser photocoagulation locus of each group, in which the CNV area was circled by a thin white dotted line.

The number of infiltrating macrophages in the laser spot (cells/spot) at 3 days after photocoagulation was significantly lower in DP2KO mice than in WT mice (24.2 ± 0.83 [n = 76 spots of 12 eyes] vs. 41.1 ± 1.6 [n = 93 spots of 12 eyes] counts/photocoagulation site, ***P < 0.001, respectively; Student's t-test; Fig. 4). 

Figure 4.

 

The number of macrophages infiltrating to the laser spot (cells/spot) at 3 days after photocoagulation of WT and DP2KO mice. (A) Macrophages were immunostained with F4/80 antibody and counted. The representative micrographs of 3 days after laser spots immunostained by F4/80 in WT mice (B) and DP2KO mice (C) were shown.

In addition, the concentrations of VEGF and MCP-1 were measured by ELISA twice using 8-week-old mouse whole eye lysate at 3 days after laser photocoagulation. The VEGF level in DP2KO mice was significantly lower than that in WT mice (79.5 ± 3.5 [n = 10 eyes] vs. 94.2 ± 4.5 [n = 10 eyes] pg/mg, *P < 0.05; Student's t-test; Fig. 5A). On the other hand, there was no significant difference between the concentration of MCP-1 in DP2KO mice and that in WT mice (28.2 ± 10.1 [n = 10 eyes] vs. 27.1 ± 7.0 [n = 10 eyes] pg/mg, P = 0.90, respectively; Student's t-test; Fig. 5B), suggesting that DP2 is not involved in the concentration of MCP-1 in the eyes. 

Figure 5.

 

The concentrations of VEGF (A) and MCP-1 (B) in whole eye at 3 days after laser photocoagulation in WT and DP2KO mice.

To explore the molecular mechanism of DP2 in vitro, we investigated DP2 expression in ARPE-19 and HUVECs. RT-PCR analysis indicated that DP2 was expressed in both ARPE-19 cells and HUVECs (Fig. 6A). 

Figure 6.

 

Expression of DP2 and GAPDH in ARPE-19 cells and HUVECs detected by RT-PCR (A). The concentrations of VEGF (B) and MCP-1 (C) in the cultured medium of ARPE cells after incubation with 15-methyl PGD₂ without (group A) or with (group B) CAY10471. Size markers of cDNA fragments are applied in a middle lane of A.

We then measured VEGF and MCP-1 production in ARPE-19 cells after stimulation with 15-methyl PGD₂, a stable DP2-selective agonist, without (group A) or with (group B) CAY10471. The VEGF level was significantly lower in group B than in group A (553.2 ± 24.5 [n = 12], vs. 443.0 ± 19.2 pg/mL [n = 12], **P < 0.01; Student’s t-test; Fig. 6B). There was no significant difference in MCP-1 concentration between the two groups (3946.1 ± 135.3 [n = 12], vs. 4158.0 ± 282.6 pg/mL [n = 12], p = 0.56; Student's t-test; Fig. 6C). These results indicated that stimulation of DP2 on ARPE-19 cells increased the release of VEGF and unchanged that of MCP-1 from these cells. 

To investigate the effect of DP2 antagonists on the differentiation and proliferation of vascular endothelial cells, we performed a tube formation assay using HUVECs. No significant difference in the number of branches was observed at 1 µM of CAY10471 (1-way ANOVA followed by Tukey's post hoc test; P = 0.44; Fig. 8a); however, the number of branches significantly decreased in a dose-dependent manner at 10 and 100 µM of CAY10471 (n = 8 wells in each group, **P < 0.01 and ***P < 0.001; 1-way ANOVA followed by Tukey's post hoc test; Fig. 7A). These results suggest that DP2 on HUVECs is involved in branching of capillary vessels in an autocrine manner, because HUVECs are known to constitutively produce PGD₂ by lipocalin-type PGD₂ synthase.²³ 

Figure 7.

 

Effect of DP2 antagonist on branching of tube formation of HUVECs. HUVECs were incubated with 1, 10, and 100 µM of CAY10471 for 18 hours and the branching of microvessels are counted under a microscope (A). The representative micrographs of branching of tube formation of HUVECS in vehicle (B) and 100 µM CAY10471 (C) were shown.

Discussion

In the current study, we investigated the involvement of the PGD₂ receptor, DP2, in the development of CNV in a mouse model of AMD. Our results showed that DP2 antagonists significantly reduced CNV size (see Fig 2), and DP2KO mice presented a reduction in CNV size (see Fig. 3), macrophage accumulation (see Fig. 4), and VEGF secretion (see Fig. 5A), suggesting that DP2 plays an important role in CNV formation. In vitro assays also indicated that PGD₂ signaling through DP2 is associated with secretion of VEGF, but not of MCP-1, from ARPE cells (see Fig. 6) and also involved in tube formation of human endothelial cells (see Fig. 7). 

Kim et al. demonstrated that lipocalin-type prostaglandin D synthase and PGD₂ metabolites produced by normal prostate stromal cells inhibit tumor cell growth through a peroxisome proliferator-activated receptor γ-dependent mechanism.²⁴ Murata et al. showed that endothelial cells, which promote angiogenesis, highly express DP1 receptors in tumors, and that loss of DP1 receptors promotes vascular leakage and angiogenesis.²⁵ Although there have been reports that angiogenesis is promoted through DP1, there have been no reports that angiogenesis is inhibited by DP2 inhibition in HUVECs, and this is a new finding. 

The number of macrophages migrating to the laser spot was significantly lower in DP2KO mice than in WT mice (see Fig. 4). Localization of DP2 on many, but not all, macrophages within the laser spot was immunohistochemically detected in WT mice (see Fig. 1). The PGD₂-DP2 pathway may regulate the migration of macrophages in the mouse CNV model. However, in our current study, there was no significant difference in MCP-1 levels between DP2KO and WT mice (see Fig. 5B), although the VEGF level in DP2KO mice was significantly lower than that in WT mice (see Fig. 5A). Tajima et al. reported that RAW264.7 macrophages initiate migration through the PGD₂/DP2 and PGE₂/EP4 signaling pathways.¹⁴ According to Katharina et al., both DP1 and DP2 are expressed in human macrophages, and stimulation of DP1 or DP2 resulted in intracellular Ca (²⁺) flux, cytokine release, and macrophage migration.²⁶ This was consistent with the fact that DP2 inhibition inhibited macrophage migration to the laser spot. 

Previous reports have indicated that COX2 inhibitors reduce CNV size via the attenuation of macrophage infiltration and downregulation of VEGF¹⁰ and another report suggested that topical COX inhibitors reduce the CNV size, VEGF, and PGE₂ levels in the retina.²⁷ Similar results were obtained in our current experiments. Inhibiting the PGD₂-DP2 pathway led to a reduction in CNV size through decreased macrophage infiltration and downregulation of VEGF expression. The PGD₂-DP2 pathway is possibly a COX2 downstream signaling pathway involved in CNV formation. Furthermore, it would be interesting to investigate whether PGD₂ stimulation accelerates CNV formation and increases CNV size in COX2 mutant mice. 

In this study, we found that VEGF production was suppressed in the human RPE cell line, ARPE-19, by a DP2 antagonist after stimulation with 15-methyl PGD₂ (see Fig. 6B). Zhang et al. found that COX2 inhibitors decreased the secretion of VEGF and TGFβ2 in mouse RPE cells.¹⁰ COX2 inhibition is predicted to decrease PGD₂ production in RPE cells, suggesting that the decreased production of VEGF is mediated by the DP2-PGD₂ pathway. 

In this study, there was no significant difference in MCP-1 levels between DP2KO and WT mice in the laser induced CNV model and DP2 blockage did not reduce MCP-1 secretion in ARPE-19 cells. These results indicated that DP2 is not involved in MCP-1 production and that the DP2-mediated macrophage migration in the CNV model may be mediated by other systems but not by MCP-1.²⁸ 

Female mice tend to have more inflammatory response than male animals.¹⁵ COX-inhibitors are well known as anti-inflammatory agents. DP2-mediated pathway is involved in CNV-related inflammation, as shown in this report. It is therefore interesting to examine the effect of DP2 blockade on the CNV model in female mice. 

This study included several limitations. First, the CNV size was manually measured in this study, however, it should be measured automatically and objectively. In the present work, a masked grader measured the CNV size, however, microscope brightness should be standardized between retinae in an automated fashion. Second, in the laser-induced CNV model, the laser irradiation was not completely constant, resulting in a variation in the number of irradiations. In order to investigate the impact of gene ablation on the CNV size, we basically performed laser photocoagulation at the 3, 6, 9, and 12 o'clock positions around the optic nerve (4 laser spots per eye). On the other hand, additional laser shots were applied when analyzing the macrophage infiltration or the concentrations of VEGF and MCP-1. It is unknown whether the uneven number of irradiations per eye might have affected the results, further investigation should be required to address this issue. Third, the laser-induced CNV model is artificially generated by the wound healing response following injury at the Bruch membrane and is highly dependent on inflammation,^29,30 therefore it is possible that the effect of DP2 inhibition might be different in human AMD. Further studies are needed to clarify the molecular mechanism of DP2 in AMD. 

In conclusion, our results suggest that DP2 is involved in the pathogenesis of AMD through macrophage accumulation and VEGF upregulation and that DP2 inhibition may be a new therapeutic target for AMD. 

Acknowledgments

Disclosure: H. Soga, None; T. Inoue, None; Y. Urade, None; T. Ueta, None; H. Kawashima, None; T. Kaburaki, None; M. Aihara, None