Interleukin-18 Induces Retinal Pigment Epithelium Degeneration in Mice

Abstract

Purpose.: To examine the effectiveness of interleukin-18 (IL-18) on choroidal neovascularization (CNV) and retinal pigment epithelium (RPE) in humans and mice.

Methods.: Serum IL-18 levels in patients with wet and dry AMD who were older than 50 years were measured and compared with those of age-matched controls. In mice, laser photocoagulation was performed in the retina to induce experimental CNV, and CNV volume was measured in eyes injected with recombinant IL-18 (rIL-18) and IL-18 neutralizing antibody (nIL-18Ab) compared with those injected with control. Tube formation assay was performed on human retinal endothelial cells (HREC) with rIL-18 administration in vitro. After subretinal injection of rIL-18, fundus change in the injected eyes was evaluated; active caspase-3 level was measured in the RPE/choroid complex, and tight junction integrity in RPE was visualized by zonula occludens-1 (ZO-1) staining.

Results.: Serum IL-18 levels in dry AMD patients were higher than those in control. Mouse rIL-18 or nIL-18Ab did not induce significant change in CNV volume compared with controls or change tube formation in HREC. Subretinal injection of rIL-18 induced retinal degeneration in the mice fundus; ZO-1 staining showed considerably disturbed RPE structure, and active caspase-3 expression was significantly higher after rIL-18 induction.

Conclusions.: Interleukin-18 did not show a pro- or antiangiogenic effect on mouse laser-induced CNVs (laser-CNVs), whereas it directly induced RPE cell apoptosis in the mouse eye. Our results suggested that IL-18 is associated with dry AMD, but not with wet AMD.

Introduction

Age-related macular degeneration (AMD) is one of the major causes of blindness in most industrialized nations. ^1,2 Age-related macular degeneration has two different forms, wet and dry AMD. Geographic atrophy in dry AMD is characterized by atrophy of the retinal pigment epithelium (RPE), ³ and wet AMD is characterized by the invasion of choroidal neovascularization (CNV) into the sensory retina. Enhanced expression of the proangiogenic cytokine vascular endothelial growth factor (VEGF) has been strongly involved in patients with wet AMD, and anti-VEGF antibody is the current standard treatment for wet AMD. ^4–6 On the other hand, there is no standard treatment to prevent RPE atrophy in geographic atrophy. Recent studies have revealed that multiple types of inflammation have a strong role in the pathogenesis of both wet and dry AMD. ^7,8  

Interleukin-18 (IL-18) is an 18-kDa cytokine that was identified as an important factor in the production of interferon-γ in response to toxic shock. It has been reported that IL-18 is expressed in Kupffer cells and activated macrophages, ⁹ keratinocytes, ¹⁰ intestinal epithelial cells, ¹¹ osteoblasts, ¹² adrenal cortex cells, ¹³ and human testis. ¹⁴ Interleukin-18 acts on T helper type 1 (Th1) T cells and, in combination with IL-12, strongly induces production of interferon-γ. ¹⁵  

Recently, IL-18 was reported to play a critical role in the pathogenesis of AMD. ^16,17 Tarallo et al. ¹⁷ found that IL-18 has a promotive function toward dry AMD; that is, it induced RPE cell death in the eye. In contrast, Doyle et al. ¹⁶ found that IL-18 showed a protective role with regard to wet AMD, and its decrease resulted in the pathogenesis of CNV. Following those major reports, several additional reports showed the roles of inflammasomes including IL-18 in the pathogenesis of AMD. ^18–21 However, the genuine role of IL-18 in the pathogenesis of AMD is not yet clear. Therefore, we here examined the role of IL-18 in patients with AMD by various methods using human and rodent samples. 

Methods

Human Serum IL-18 Measurement

Interleukin-18 levels in serum from human AMD patients were measured by enzyme-linked immunosorbent assay (ELISA) as previously described. ^22,23 Briefly, serum was prepared from 43 patients with wet AMD, 17 patients with dry AMD, and 40 controls. Patients with polypoidal choroidal vasculopathy (PCV), retinal angiomatoid proliferation (RAP), maculopathy with myopic CNV, and CNV based on angioid streaks were excluded. The diagnosis of wet and dry AMD was established on the basis of age (>50 years old), clinical examination, fundus photography, optical coherence tomography, and fluorescein fundus angiography. Patients who had wet AMD in one eye and dry AMD in the other eye were excluded from both the wet AMD patient and dry AMD patient groups. Control serum was obtained from patients with other ocular diseases, for example, cataract, glaucoma, retinal detachment, macular hole, and epiretinal membrane. In both patients and control groups, those who had inflammatory diseases such as asthma and rheumatoid arthritis were excluded. The levels of human IL-18 were measured with a human IL-18 ELISA kit (MBL, Nagoya, Japan) according to the manufacturer's protocol. All procedures were conducted at 4°C until the final washing step was completed. Plates were analyzed by measuring absorbance at 450 nm (reference 620 nm) using a plate reader (Bio-Rad, Richmond, CA, USA). This study protocol complied with the ethical guidelines of the Declaration of Helsinki and was approved by the institutional review board of Nagoya University Hospital, and informed consent was obtained from each patient. 

Animals

Male wild-type C57BL/6J mice (CLEA, Tokyo, Japan) between 6 and 8 weeks of age were used. For all procedures, the animals were anesthetized with intraperitoneal injection of 400 mg/kg Avertin (2.5% 2,2,2-tribromoethyl and tertiary amyl alcohol; Sigma-Aldrich Corp., St. Louis, MO, USA), and pupils were dilated with a combination of tropicamide 0.5% and phenylephrine 0.5% (Mydrin-P; Santen, Osaka, Japan). The use of animals in the experimental protocol was approved by the Nagoya University Animal Care Committee. All animal experiments were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Laser-CNV Volume Analysis

Four spots of laser photocoagulation (532 nm, 180 mW, 100 ms, 75 μm; Novus Verdi; Coherent, Inc., Santa Clara, CA, USA) were performed in each fundus of eyes on day 0 by one individual masked to the group assignment to induce laser-induced choroidal neovascularization (laser-CNV) as described previously. ^23,24 The laser spots were created around the optic nerve using a slit-lamp delivery system, and a coverslip was used as a contact lens. Laser-CNV volume was measured with a method similar to that previously described. Briefly, 1 week after the laser injury, eyes were enucleated and fixed with 4% paraformaldehyde (PFA). Eye cups obtained by removing the anterior segments were incubated with 0.5% fluorescein-isothiocyanate (FITC)-isolectin B4 (Sigma-Aldrich Corp.). Choroidal neovascularization was visualized using a blue argon laser wavelength (488 nm) and a scanning laser confocal microscope (Eclips C1 confocal; Nikon, Tokyo, Japan). Horizontal optical sections were obtained at 1-μm intervals from the top of the CNV to the surface of the RPE. The images of each layer were digitally stored, and the area sizes were measured. The area of CNV-related fluorescence was measured using ImageJ software (NIH, Bethesda, MD, USA). The summation of the whole fluorescent area in each horizontal section was used as an index for the volume of CNV. Imaging was performed by an operator masked to the group assignments. The average volume obtained from all three or four laser spots per eye was generated (n = number of eyes). 

Intravitreous and Subretinal Injections of Mouse Recombinant IL-18 and Anti-IL-18 Neutralizing Antibody

To evaluate the effect of IL-18 on CNV, 1 μg mouse recombinant IL-18 (rIL-18, B004-5; MBL) dissolved in ultrapure PBS or the same volume of PBS was administered intravitreously at day 0 immediately after inducing laser-CNV in wild-type mice eyes. Furthermore, 0.3 μg anti-mouse IL-18 neutralizing antibody (nIL-18Ab, D048-3; MBL) or isotype IgG antibody (Vector Labs, Burlingame, CA, USA) was intravitreously administered after induction of laser-CNV. Recombinant IL-18 (18 kDa) is small enough to be dissolved in PBS and reaches RPE as if other cytokines/proteins, as well as IgGs (~150 kDa). ^25,26 Phosphate-buffered saline is not toxic, whereas normal saline is reported to be toxic to the mouse retina. ²⁷ Intravitreous injection was performed with a 33-gauge needle (Ito Corporation, Shizuoka, Japan) under a surgical microscope. To evaluate the direct effect of IL-18 on mouse RPE, 1 μg mouse rIL-18 dissolved in ultrapure PBS or the same volume of PBS was administered subretinally with a blunt-tip 33-gauge needle (Ito Corporation) under a surgical microscope with a coverslip used as a contact lens. Successful subretinal injection was confirmed by the bullous retinal detachment on the hemisphere of the fundus. 

Fundus Imaging

Mouse ocular fundus images were obtained using a high-resolution digital fundus camera (TRC-50DX; Topcon, Tokyo, Japan). To adjust focus on the mouse fundus, a 20-diopter lens was placed in contact with the fundus camera lens. ¹⁷  

Tube Formation Assay

To evaluate the effectiveness of IL-18 on ocular angiogenesis, a tube formation assay was performed as previously described. ²⁸ Briefly, human retinal endothelial cells (HREC) from Cell Systems (Kirkland, WA, USA) were cultured with EGM-2 medium (Lonza, Tokyo, Japan) in an incubator with 5% CO₂-enriched air. Extracellular matrix gels were prepared with a Chemicon in vitro angiogenesis assay kit (EMD Millipore, Billerica, MA, USA). Gels were solidified over a 96-well microplate. Using this kit, 1.25 × 10⁴ HREC were added onto the surface of gels with administration of human rIL-18 (B003-5; MBL) at 1 to 100 nM in the medium. After 4 hours of incubation, tubes were labeled by Calcein-AM solution (Dojindo Laboratories, Kumamoto, Japan) and photographed. The area sizes of tube formation per were measured and calculated by ImageJ. 

Zonula Occludens-1 Staining of Retinal Pigment Epithelium

To visualize the integrity of the RPE structure, zonula occludens-1 (ZO-1) staining was performed as previously described. ^3,17 Briefly, mouse RPE/choroid flat mounts were fixed with 4% PFA or 100% methanol, stained with rabbit antibodies against ZO-1 (1:100; Invitrogen, Carlsbad, CA, USA), and visualized with Alexa 594 (Invitrogen). All images were obtained using the Nikon confocal microscope (Eclips C1). 

Caspase-3 Activity After IL-18 Injection

On days 3 and 5 after subretinal injections of mouse rIL-18 (1 μg, B004-5; MBL), the eyes were enucleated and RPE/choroid flat mounts were created after radial incisions and retina removal. The RPE/choroid complexes were carefully isolated with a scraper and were lysed in RIPA buffer (R0278; Sigma-Aldrich Corp.) with a protease inhibitor cocktail (Roche Diagnostics, Indianapolis, IN, USA). The lysate was centrifuged at 15,000g for 15 minutes at 4°C, and supernatant was collected. Protein concentrations were determined using a Bradford assay kit (Bio-Rad) with bovine serum albumin as a standard. Caspase-3 activities were measured as previously described. ²⁹ Briefly, 100 μg total protein was assayed for active caspase-3 by a commercial kit (BF-1100; R&D Systems, Minneapolis, MN, USA) and analyzed on a fluorescent plate reader according to the manufacturer's recommendations (PowerScan 4; DS Pharma Biomedical, Osaka, Japan). Duplicate evaluations were performed for each sample. 

Statistical Analysis

Results are expressed as mean ± SEM (n = number of samples). Serum IL-18 levels and tube formation assay of HREC with rIL-18 were analyzed with the Kruskal-Wallis test; and if significance was detected (P < 0.05), this was followed by Steel's test for comparison with the control. The other results were analyzed statistically using the Mann-Whitney U test (unpaired samples). Differences were considered to be statistically significant at P < 0.05. 

Results

Higher Serum IL-18 Level in Dry AMD Patients

First we examined the serum IL-18 levels in patients with wet and dry AMD, excluding PCV, RAP, myopic CNV, and CNV secondary to angioid streaks, and compared them with levels in controls. The wet AMD patients comprised 43 patients aged from 50 to 92 years (average age, 75.3; 31 males and 12 females), and the dry AMD patient group comprised 17 patients aged from 50 to 94 years (average age, 68.8 years; 11 males and 6 females). Control patients were composed of 40 patients aged from 53 to 85 years (average age, 69.5; 18 males and 22 females) who did not have any retinal diseases related to angiogenesis or retinal degeneration. Serum IL-18 levels were higher in dry AMD patients (71.5 ± 10.6 pg/mL) by 48.9% compared with the controls (48.0 ± 2.6 pg/mL, Fig. 1, P = 0.033, Steel's test). By contrast, the serum IL-18 in wet AMD patients (57.4 ± 9.9 pg/mL) did not show a significant difference compared to control (P = 0.097, Steel's test). These results suggested a possible association of serum IL-18 with dry AMD, but not with wet AMD. 

Figure 1

 

Serum interleukin-18 (IL-18) levels were higher in patients with dry age-related macular degeneration (AMD). Serum IL-18 levels from patients with wet AMD and dry AMD and controls older than 50 years were measured. Serum IL-18 levels were higher in dry AMD patients by 48.9% compared with control (P = 0.033), but IL-18 levels in wet AMD patients were not significantly different (P = 0.097). *P < 0.05, N.S., no significant difference.

Effect of IL-18 on Laser-CNV in Mice

To evaluate the role of IL-18 in CNV as the culprit in wet AMD, laser-CNVs were performed in wild-type mice. The CNV volume of the wild-type mice injected with mouse rIL-18 was not statistically different from that in animals injected with control (Figs. 2a–c, 1.00 ± 0.07 vs. 0.91 ± 0.07, P = 0.35, n = 9). Also, the CNV volume of the wild-type mice injected with mouse nIL-18Ab was not statistically different from that in mice injected with isotype control IgG (Figs. 2d–f, 1.00 ± 0.09 vs. 0.98 ± 0.06, P = 0.87, n = 11). 

Figure 2

 

Laser-induced choroidal neovascularizations (laser-CNVs) were not significantly affected by IL-18. (a–c) The volume of laser-CNVs in wild-type mice injected with mouse recombinant IL-18 (rIL-18, 1 μg) did not show significant change compared with control (P = 0.35). (a, b) Representative images of laser-CNV in wild-type mice injected with control solution (a) and rIL-18 (b). (d–f) The volume of laser-CNVs in wild-type mice injected with anti-mouse IL-18 neutralizing antibody (IL-18Ab, 0.3 μg) did not show significant change compared with volume in those injected with isotype IgG (P = 0.87). (e, f) Representative images of laser-CNV in wild-type mice injected with isotype IgG (d) and IL-18Ab (e). Scale bars: 50 μm. N.S., no significant difference.

Effect of IL-18 on Human Retinal Endothelial Cells

We also examined the effect of IL-18 on retinal endothelial cells in vitro (Fig. 3). The area size of HREC vessels in control medium was 1.0 ± 0.04 (n = 15); values in medium with 1, 10, and 100 nM rIL-18 were 0.98 ± 0.05, 1.04 ± 0.05, and 1.00 ± 0.04, respectively (n = 15). The administration of human rIL-18 did not affect tube formation in HREC (P = 0.821, Kruskal-Wallis test). 

Figure 3

 

Human recombinant IL-18 did not affect tube formation in retinal endothelial cells. (a) Various concentration of IL-18 (0 [control], 1, 10, and 100 nM) did not induce significant changes in vessel formation. (b–e) Representative images of tube formation in retinal endothelial cells with 1 nM (c), 10 nM (d), and 100 nM (e) human rIL-18 and control (b). Scale bars: 100 μm. N.S., no significant difference.

Direct Effect of IL-18 on RPE

To evaluate the direct effect of IL-18 on the RPE cells, mouse rIL-18 was delivered directly into the subretinal space. Subretinal injection of rIL-18 showed that IL-18 induced RPE degeneration on the hemisphere of the fundus at day 7 after the injection (Fig. 4a). Additionally, ZO-1 staining revealed strongly disturbed structures of RPE in eyes injected with rIL-18 (Fig. 4c), whereas subretinal injection of PBS into the wild-type mice did not lead to RPE degeneration as shown in fundus images (Fig. 4b) or ZO-1 disruption (Fig. 4d). Caspase-3 activities were measured from the RPE/choroid complex in eyes subretinally injected with rIL-18 at days 3 and 5 after the injection. Compared to what was seen in the PBS-injected eyes, the caspase-3 activities were upregulated and significantly higher both at day 3 (1.0 ± 0.24 vs. 1.88 ± 0.19, P = 0.012, n = 30 and 34) and at day 5 (1.0 ± 0.17 vs. 2.15 ± 0.21, P = 0.0003, n = 18) in the RPE/choroid complex after subretinal injection of mouse rIL-18 (Fig. 4e). These results indicated that IL-18 induced cell apoptosis of RPE in vivo. 

Figure 4

 

Subretinal injection of mouse recombinant IL-18 induced retinal degeneration in wild-type mice. (a, b) Color fundus image from wild-type mouse eye at 7 days after subretinal injection of 1 μg mouse recombinant IL-18 (rIL-18) showed retinal/RPE degeneration ([a] surrounded by white arrowheads), whereas eyes treated with PBS did not (b). (c, d) Zonula occludens-1 (ZO-1) image from wild-type mouse eye subretinally injected with 1 μg mouse recombinant IL-18 (rIL-18) showed strongly damaged RPE structure (c), whereas eyes treated with PBS did not (d). (e) Relative expression of caspase-3 activities in RPE/choroid complex from eyes injected with mouse rIL-18 was significantly increased compared with those injected with control solution. Scale bars: 100 μm. *P < 0.05. Arrow: Injection site.

Discussion

In this study, we showed that IL-18 did not have a direct pro- or antiangiogenic effect on the pathogenesis of CNV, whereas it directly induced RPE degeneration. We found that serum IL-18 levels were higher in dry AMD patients than in controls. Interestingly, as far as we know, there have been no reports showing an association between IL-18 levels in serum and AMD. However, the serum levels in wet AMD patients did not show a statistical difference from controls. The serum IL-18 examinations that we performed were subjected to cross-sectional analysis; thus, it is not possible to completely exclude factors that could possibly increase individual variances. For instance, in most patients, blood was collected when they visited our clinic for the first time, and some of the wet AMD patients had already received anti-VEGF treatment by the time the blood was extracted. Furthermore, some were taking antioxidant supplements that could possibly prevent AMD deterioration. It is interesting, however, that dry AMD patients had higher IL-18 levels than the controls considering such a wide variation. It is also interesting that in patients with ocular disease that was very spatially limited compared with body size, serum levels of interleukin were different, especially in such a limited number of samples. Although we excluded all AMD and control patients who potentially had an inflammatory disease not only in the eye but also in the whole body based on medical records, there might be some other factors that affect IL-18 levels. Unfortunately, we could not obtain anterior chamber fluid from the patients with dry AMD and control groups in our study, and it would probably be meaningful to measure IL-18 levels in aqueous fluid from those patients in the future. Only a small number of reports show cytokine/chemokine changes in serum of AMD patients. ³⁰ Nevertheless, these findings, including ours, imply that AMD is not a simple ocular disease but has a relation to whole-body changes. 

Our primary hypothesis was that IL-18 has an important function in the pathogenesis of AMD. To confirm this hypothesis, first we performed laser-CNV and injections of rIL-18 and nIL-18Ab and compared with controls. However, neither rIL-18 nor mouse nIL-18Ab had either a pro- or an antiangiogenic effect. Hence we speculated that IL-18 has relevance to RPE degeneration, which is the main pathology of geographic atrophy in AMD. To confirm this speculation, we directly delivered rIL-18 into the subretinal space. We found that IL-18 directly induced RPE cell apoptosis. As to the effectiveness of IL-18 on the retina, there have been several reports in the last few years. Tarallo et al. ¹⁷ found that DICER1 decrease caused Alu accumulation in RPE, followed by NLRP3 activation, resulting in RPE cell death. In these studies, IL-18 was reported to exist downstream of NLRP3 and was activated to induce RPE cell death. ¹⁷ On the other hand, Doyle et al. ¹⁶ found that IL-18 has a protective role to suppress CNV and that its decrease led to the pathogenesis of CNV. ¹⁶ Subsequently Hirano et al. ²⁰ found that a high concentration of glycerol had a proangiogenic effect on mouse laser-CNV, and IL-18 antibody without glycerol did not affect laser-CNV. Interleukin-18s seem to have a complicated multifunction in many diseases. For instance, in dextran sulfate sodium–induced colitis (an animal model of inflammatory bowel disease), IL-18 was primarily reported to have a promotive role to worsen colitis, ³¹ whereas it was later reported to have a protective role since IL-18^−/− mice showed more severe colitis. ³² Interleukin-18 may have a different function in different tissues and different states. In our study, IL-18 did not promote or prevent ocular angiogenesis either in vivo or in vitro. On the other hand, IL-18 induced RPE apoptosis in vivo. As seen in Figure 4, the injected areas were less than a hemisphere. We prepared RPE/choroid lysates from whole eyes including another half (noninjected area) of the RPE. The results for caspase-3 activities were affected by the large size of intact RPE/choroid cells and thus may be underestimated. 

Previous reports have shown that IL-18s are expressed in macrophages, ³³ and macrophages have a strong relevance to the pathogenesis of AMD. ^34,35 It is possible that induction of RPE cell death by IL-18 occurred under the effects of other cells, for example, macrophages. Unlike VEGF in wet AMD, none of the factors that strongly influence and destine the severity of dry AMD have been found. Moreover, in clinical situations, there are many cases in which wet and dry AMD coexist in the same patients. Properly understanding and controlling the function of IL-18 will be the key to taming AMD. To evaluate the genuine function of IL-18, further investigation will be necessary. 

Acknowledgments

The authors thank Reona Kimoto and Seina Ito for technical assistance. 

Supported by a Grant-in-Aid for Young Scientists (A) and a Grant-in-Aid for Challenging Exploratory Research from the Japan Society for the Promotion of Science (HK), Chukyo Longevity Medical and Promotion Foundation (HK), Takeda Science Foundation (HK), a Takayanagi Memorial Grant (RI), and Hori Science and Arts Foundation (FY). 

Disclosure: R. Ijima, None; H. Kaneko, None; F. Ye, None; Y. Nagasaka, None; K. Takayama, None; K. Kataoka, None; S. Kachi, None; T. Iwase, None; H. Terasaki, None