Establishment of a New Animal Model of Focal Subretinal Fibrosis That Resembles Disciform Lesion in Advanced Age-Related Macular Degeneration

Abstract

Purpose.: Subretinal fibrosis causes damage to visual acuity, especially if the lesion is in the macula, as is frequently observed in advanced age-related macular degeneration. Exudate leukocytes form abnormal vessels that initiate regional inflammation accompanied with local glial proliferation and matrix production. The purpose of this study was to establish an animal model of focal subretinal fibrosis.

Methods.: Macrophage-rich peritoneal exudate cells (PECs) were injected into the subretinal space of C57BL/6 or MCP-1 knockout (KO) mice. Seven days later, the size of the subretinal fibrotic tissue was evaluated by the adherent area of glial fibrillary acidic protein (GFAP)–positive retinal glial cells on choroidal flat mounts. Myofibroblastic changes and collagen synthesis were detected by α-smooth muscle actin (α-SMA) and Masson trichrome staining of the histologic section, respectively. α-SMA expression was also examined on retinal pigment epithelium (RPE) cells during co-culture with activated macrophages.

Results.: Subretinal fibrous tissue was observed by funduscopy in PEC-injected mice after 7 days. The tissue consisted of a monotonous, low-cell-density area that expressed α-SMA with collagen synthesis. Both steroid and antioxidant treatment can reduce residual glia. Because PEC-injected MCP-1 KO mice showed less residual glia, not only exogenous macrophages, but also intrinsic macrophages were activated. The macrophages directly induced myofibrotic changes in RPE cells in vitro.

Conclusions.: Activated macrophages form subretinal fibrosis when they are placed in the subretinal space and induce myofibrotic changes in RPE cells.

Tissue fibrosis is an important wound-healing process that is essential to the repair of damaged organs, such as burned skin, sutured tissues during surgery, and other such injuries. Fibrosis is initiated by conformational changes in the damaged cells that cause expansion of regional fibroblasts that synthesize certain collagens and extracellular matrixes to cover the defective tissues. ¹ This process is also mediated by the accumulation of bone marrow–derived macrophages and dendritic cells. ² Macrophage-derived soluble factors are critical in this process. ³  

In contrast to these physiological points of view, fibrosis sometimes causes organ dysfunction such as in the lung and liver. In line with pathogenic fibrosis, subretinal fibrosis has been recognized as a feature of advanced vitreoretinal diseases that lead to loss of vision. During progression of refractory vitreoretinal diseases, such as proliferative diabetic retinopathy (PDR), proliferative vitreoretinopathy (PVR), uveitis, and age-related macular degeneration (AMD), fibrosis often develops in the subretinal space, with proliferation of glial cells. ^{4 –6} In PDR and PVR, shrinking subretinal fibrous tissues causes irreversible retinal detachment. In AMD, focal subretinal fibrosis forms due to long-lasting, active choroidal neovascularization (CNV), which then damages macular function. ⁷  

There are animal models of AMD that focus on CNV formation. ^8,9 However, there is no model of focal retinal fibrosis, which is frequently observed in advanced AMD. Several reports have demonstrated that activated macrophages play an essential role in tissue fibrosis. ^1,3,10 In the eye, macrophage infiltration is also thought to be one of the important pathologic processes in subretinal proliferation in retinal diseases, because macrophages can produce fibrogenic factors, proliferative factors, and angiogenic factors. ¹¹ They can also promote extracellular matrix synthesis by regional fibrogenic cells. ³ We thus attempted to establish a mouse model of subretinal fibrosis by introducing activated macrophages into the subretinal space and then analyzing the pathogenic process. 

Methods

Animals

Female C57BL/6 (B6) mice between 6 and 10 weeks old were purchased from Japan SLC (Shizuoka, Japan), and monocyte chemoattractant protein (MCP)-1 knockout (KO) mice with a B6 background ¹² were kindly provided by Barrett J. Rollins (Harvard Medical School, Boston, MA). All animal experiments were approved by the Committee on the Ethics of Animal Experiments, Kyushu University Graduate School of Medical Sciences, Japan. Animals were treated according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Peritoneal Exudate Cell Preparation

For the isolation of macrophages, thioglycolate-induced peritoneal exudates were used as a source of recently extravasated macrophages. Mice received an intraperitoneal injection of 2.5 mL of thioglycolate. Three days later, the mice were euthanatized and the peritoneal exudate cells (PECs) were isolated. The cells were washed with PBS, and cell counting was performed to calculate consistent amounts of PECs for subretinal injection. A concentration of PECs of 4 × 10⁷/mL was prepared for subretinal injection. 

Induction and Evaluation of Subretinal Fibrosis in Mice

We injected PECs into the subretinal space of the right eye in each mouse. To achieve the most accurate subretinal injection volume, two types of needle were used. The first type was a sharp, beveled 32-gauge needle with the tip cut off, for puncturing both the sclera and retina, and the other was a blunt, nonbeveled 32-gauge needle. To make the subretinal bubble and to rupture Bruch's membrane, laser photocoagulation (0.05 second, 200 mW, 532 nm diode laser, Iridex, Mountain View, CA) was performed in the temporal and posterior regions of the retina of the right eye, and 0.5 μL of prepared PECs (4 × 10⁷/mL) were injected into the subretinal space with a 10-μL syringe (Hamilton Company, Reno, NV) with a blunt-tipped needle attached. After injection, the retinal break was sealed to prevent leakage of the injected PECs into the vitreous cavity (Fig. 1A). Eyes in which there was bleeding or no focal retinal detachment were excluded. 

Figure 1.

 

Establishment of a mouse model of focal subretinal fibrosis. (A) Induction of focal subretinal fibrosis. (Aa) Bruch's membrane was ruptured by laser, (Bb) creating a subretinal air bubble. (Ac) A sharp needle was inserted to make a small retinal break. (Ad) PECs that mostly consist of activated macrophages were injected into the subretinal space with a blunt, nonbeveled needle. (B) Fundus photograph of a PEC-injected right eye 7 days later. (C) Histologic examination of the PEC-induced lesion. Representative HE staining of subretinal fibrosis is shown. Left: PBS injection site; right: PEC injection site. Scale bar, 50 μm. (D) Representative α-SMA staining of focal subretinal fibrosis. Myofibroblastic change was confirmed by α-SMA staining (red) in the PEC-injected group. (E) Representative collagen type I staining (green) of subretinal fibrosis.

Seven days after injection, the mice were euthanatized, and the PEC-injected right eyes were removed. After dissecting the cornea and lens, we made radial relaxing incisions in the eye cups. GFAP staining was performed to visualize the area of residual glia on a choroidal flat mount. The tissues were fixed for 20 minutes with 4% paraformaldehyde, after which the retina and flattened choroid were easily separated. The flattened choroid was rinsed with PBS and soaked in 100% methanol at room temperature for 10 minutes. The samples were blocked with 5% skim milk in PBS for 1 hour at room temperature and then incubated with a GFAP primary antibody (dilution 1:400, rabbit IgG; Dako, Glostrup, Denmark) at 4°C for 24 hours. The sections were then incubated with fluorescein isothiocyanate (FITC)–conjugated anti-rabbit IgG. All samples were mounted (Crystal/Mount; Biomedia, Foster City, CA). Some samples also were stained with anti-collagen type I mAb (ab292; Abcam, Cambridge, UK), followed by APC-conjugated anti-rabbit Ig (Abcam). 

Photomicrographs were captured with a digital camera (DC350F with allied software QFluoro; Leica Microsystems, Cambridge, UK) mounted on a fluorescence microscope (DMIRE2; Leica). The area of visualized residual glia on the choroidal flat mount was measured with Image J software (National Institutes of Health, Bethesda, MD). To evaluate the effect of PEC injection for CNV formation followed by the laser, we stained some samples with fluorescein-isothiocyanate–conjugated Griffonia simplicifolia isolectin B4 (Vector Laboratories, Burlingame, CA), and visualized the vessels by excitation with a blue argon laser at 488 nm and emission between 515 and 545 nm. 

Some mice received dexamethasone (2 mg/kg, 100 μL; Sigma-Aldrich, St. Louis, MO) intraperitoneally (IP) on days 0, 1, 2, 3, and 4. 

Histologic Analysis of Murine Eyes

Seven days after PEC injection, the eyes were removed, fixed in 4% paraformaldehyde, embedded in paraffin, and stained with hematoxylin and eosin (HE). In addition, Masson trichrome staining was used to evaluate local collagen synthesis. Some freshly enucleated eyes were embedded in freezing compound (Tissue-Tek OCT; Sakura, Tokyo, Japan) and α-smooth muscle actin (α-SMA, dilution 1:100, mouse IgG; Dako, Glostrup, Denmark) staining was performed to see myofibroblastic changes on 10-μm-thick frozen sections. Horse anti-mouse IgG (1:500; Vector Laboratories) was used as the secondary antibody, and HRP-streptavidin (Vector Laboratories) and DAB were used as the substrates. Photomicrographs were then captured with a microscope (model BX50; Olympus, Tokyo, Japan). 

Qualitative Determination of Glutathione in Sectioned Specimens

For histologic evaluation, three eyes from either PEC- or PBS-treated mice were randomly selected and enucleated, frozen, embedded, and sectioned. They were stained with 100 mM monochlorobimane (MCB, M-1381; Molecular Probes, Eugene, OR) and propidium iodide. Fluorescence intensity, reflecting the amount of intracellular glutathione, was inspected under a confocal laser scanning microscope (TCS SP5; Leica). Glutathione levels were detected with a fluorescent MCB probe with excitation and emission wavelengths of 405 nm and 410 to 480 nm, respectively. The cell-permeating MCB probe is nonfluorescent but forms a fluorescent adduct with glutathione in a reaction catalyzed by glutathione-S-transferase. 

Treatment of Antioxidant Reagents

We suspended PECs (4 × 10⁷/mL) into either N-acetyl-l-cysteine (NAC, 5 mM in PBS; Senju, Osaka, Japan,) solution or glutathione monoethyl ester (GSH-OEt, 100 mM in PBS; CalBiochem, San Diego, CA) solution, as described before. ¹³ Prepared PECs (0.5 μL) were injected into the subretinal space. 

RPE Cell Preparation and Culture

Mice RPE cells were cultured for approximately 2 weeks until becoming confluent in 24-well plates in Dulbecco's modified Eagle's medium containing 20% heat-inactivated fetal calf serum supplemented with 100 U/mL penicillin, 100 mg/mL streptomycin, 1% l-glutamine, and 0.1 mM nonessential amino acids. ¹⁴ Designated concentrations of PECs were added to the primary culture and stained for α-SMA 48 hours later. Fixed cultured cells were rinsed with PBS and permeated in 100% methanol at room temperature for 10 minutes. The samples were blocked using 5% skim milk in PBS for 1 hour at room temperature, then incubated in FITC-conjugated anti-α-SMA antibody (dilution 1:100, mouse IgG; Dako, Glostrup, Denmark) at 4°C for 24 hours. All samples were counterstained with DAPI, mounted (Crystal/Mount; Biomedia), and subjected to fluorescence microscopy (BZ-9000; Keyence, Osaka, Japan). TUNEL analysis and quantification of TUNEL-positive cells were performed with a cell viability assay (ApopTag Fluorescein In Situ Apoptosis Detection Kit; Chemicon International-Millipore, Billerica, MA). Some samples were co-stained with FITC-conjugated anti-F4/80 mAb (clone A3–1; Caltag Laboratory, Inc., San Francisco, CA) to visualize the macrophages. 

Flow Cytometry

RPE cells were prepared from C57BL/6 mice and cultured for approximately 2 weeks until becoming confluent in 24-well plates. Designated concentrations of PECs were added to the primary culture. RPE cells were removed from the dish by a vigorous wash, stained with α-SMA, and analyzed by flow cytometry. 

Statistics

Data were analyzed by ANOVA and Scheffé's test. Differences between experimental groups with P ≤ 0.05 were considered to be significant. 

Results

Subretinal Injection of Activated Macrophages Can Induce Focal Subretinal Fibrosis That Resembles Disciform Lesion in Advanced AMD

To examine the mechanism of focal subretinal fibrosis in advanced AMD, we created a mouse model of disciform lesions. Since macrophages can initiate fibrotic changes in multiple organs ^2,3 and can actually exist in the subretinal scar in AMD patients, ¹⁵ we decided to inject thioglycolate-elicited peritoneal macrophages into the subretinal space as activated macrophages (Fig. 1A). Immediately after injection, focal retinal detachment was confirmed in all cases. After 3 days, the primary induced retinal detachment gradually flattened. Seven days later, subretinally injected PECs successfully caused subretinal fibrosis detected by fundus scope (Fig. 1B, yellow dotted circle). HE staining showed that a low-cell-density fibrotic area (Fig. 1C, right panel, yellow dotted circle) had formed beneath the retina that was not obvious in the PBS-injected control. There was no retinal fibrotic scar in the noninjected area (data not shown). 

It was notable that myofibroblastic changes were also induced in the subretinal space and were confirmed by α-SMA staining. The PEC-injected group had much more myofibroblastic changes than the PBS-injected group (Fig. 1D). α-SMA was predominantly stained in the subretinal space. It was also confirmed that subretinal fibrosis contained green-stained collagen fibrils (Fig. 1E; Masson trichrome), and thus the PEC-injection–induced experimental subretinal fibrosis in mice was complete. 

To quantify the subretinal fibrosis, the area of glial fibrillary acidic protein (GFAP)–positive retinal glial residual was measured on the choroidal flat mount 7 days later. Since retinal glia tended to adhere to the choroid in the scarred portion, the retina and choroid were separated (Fig. 2A). We included not only the strongly adherent fibrous area (located at the center of the fibrotic tissue), but also the peripheral scattered adherent area (Fig. 2B, yellow dotted circle). The residual glia were not always fibrous, and some portion of them (especially in the peripheral area) looked like scattered small dots on the flat mount. As shown in Figure 2B, the peripheral area was well differentiated from the other dark area, and we easily discriminated and measured the area accurately by using the contrast-based selection function in the ImageJ software. In fact, compared with the PBS-injected control (sham procedure), PEC-injected mice showed a marked increase in residual glial area on the choroidal flat mounts (Figs. 2B, 2C). 

Figure 2.

 

The area measurement of residual glia on the choroidal flat mounts associated with subretinal fibrosis. (A) Residual glia on the choroidal flat mount. (B) Representative GFAP-stained residual glia. Left: PBS-injection site; right: PEC-injection site. Scale bar, 500 μm. (C) Comparison of the area of GFAP staining (n = 9 of each group, P < 0.01). The experiments were repeated twice with similar results. (D) Representative double staining of lectin (green) and GFAP (red). (Top) PBS-injected and (bottom) PEC-injected flat mounts. (E) Representative double staining of GFAP (green) and collagen type I (purple) in the PEC-injected flat mount.

Fibrosis is a wound-healing event, and high-power laser treatment is known to induce CNV. The question may arise of whether the fibrosis process is affected by CNV. It is possible that CNV plus additional macrophages induce fibrosis, and we thus performed double staining of isolectin (for CNV)/GFAP (for fibrosis) in our model. We found that CNV and fibrosis were co-existent (Fig. 2D). Importantly, compared with the PBS-injected control, the PEC-injected eyes showed not only GFAP-positive fibrosis, but also a markedly expanded isolectin-positive vascularized area (Fig. 2D). 

GFAP is not a direct marker, but is a good indirect marker, of fibrosis. We tried direct staining for fibrosis by using an anti-collagen type I antibody. Double-staining results (collagen and GFAP) showed that the fibrotic area visualized by these two antibodies corresponded well, suggesting that both markers detect fibrosis effectively (Fig. 2E). However, the direct staining did not constantly detect the fibrotic area, because of the irregular surface of the extracellular matrix. We chose indirect staining because the anti-GFAP antibody is a strongly staining antibody that consistently detects retinal glial residues on choroidal flat mounts. Subretinal fibrotic tissue strongly adheres to the retina and always had residual glia when we detached the retina from the choroid. 

Inflammatory Process Is Important to Form Subretinal Fibrosis

Fibrosis is mediated by the accumulation of bone marrow–derived inflammatory cells and their expressed inflammatory cytokines. ² Moreover, tissue stromal reaction is frequently accompanied by the generation of reactive oxygen species, which in turn regulate the cytokine milieu at inflamed tissue sites. ¹⁶ To determine the contribution of the inflammatory process, we examined the effects of a systemic corticosteroid on subretinal fibrosis. Dexamethasone (2 mg/kg, 100 μL) was injected intraperitoneally on days 0, 1, 2, 3, and 4 after PEC injection. The steroid-injected group had a smaller area of subretinal fibrosis than did the control group treated with PBS (Fig. 3). The inflammatory process therefore contributes to subretinal fibrosis. 

Figure 3.

 

The area measurement of residual glia in steroid-treated mice. (A) Steroid treatment (2 mg/kg, 100 μL, intraperitoneal) began immediately after PEC injection, on days 1, 2, 3, and 4 after treatment). Representative immunohistochemical staining for GFAP in a choroidal flat mount from a PBS-treated control mouse (left) or steroid-treated mouse (right). Scale bar, 500 μm. (B) The mean ± SEM areas of fibrosis stained with GFAP, measured 1 week after PEC injection. Steroid-treated mice (n = 9) or PBS-treated control mice (n = 9; *P < 0.05). The experiment was repeated three times (18 mice each) with similar results.

Endogenous Macrophage Recruitment Is Critical to the Formation of Focal Fibrosis

Injected exogenous macrophages initiated local fibrogenic responses and began to recruit endogenous bone marrow–derived cells, especially macrophages. Therefore, not only injected macrophages, but also subsequently recruited endogenous macrophages can be critical in focal inflammation and can form subretinal fibrotic tissues. To clarify this point, we injected PECs into either control B6 or MCP-1 KO mice (B6 background). MCP-1 mediates the recruitment of monocytes/macrophages in several inflammation models and diseases. As a result, mice lacking MCP-1 cannot recruit endogenous bone-marrow macrophages to the inflammatory site. ¹²  

Although we used activated macrophages from syngeneic mice for the injections, MCP-1 KO mice had a smaller area of GFAP-positive staining than that in the control group (Fig. 4). We concluded that not only exogenous macrophages, but also endogenous macrophage infiltration must be an essential process in subretinal fibrosis. 

Figure 4.

 

The area measurement of residual glia in MCP-1 KO mice. (A) Representative immunohistochemical staining for GFAP in a choroidal flat mount of a PEC-injected B6 mouse (left) or an MCP-1 KO mouse (right). Scale bar, 500 μm. (B) The mean ± SEM areas of fibrosis stained with GFAP, measured 1 week after PEC injection. PEC-injected B6 mice (n = 9), MCP-1 KO mice (n = 9; *P < 0.05). The experiment was repeated three times (18 mice each) with similar results.

Oxidative Injury May Be Related to Fibrosis

Oxidative stress can trigger various eye diseases, including AMD. ^17,18 It is also known that macrophages can vary and/or change their function according to the regional circumstances. Glutathione constitutes the first line of cellular defense against oxidative injury ¹⁹ and its concentration indicates the functional status of the macrophages. ²⁰ Macrophages with decreased intracellular glutathione are known as oxidative macrophages and those with increased amounts as reductive macrophages. ²¹ We thus examined the glutathione concentration within the PEC-injected site in our model. Figure 5A (left) shows the glutathione concentration by color gradation. The reductive colors were predominant in the injected subretinal space at 24 hours after injection, but the oxidative colors came to the fore up to 72 hours after the procedure. The right panels also show the reduction in glutathione concentration (blue) on the same slices. The data indicate that oxidative changes are somehow related to the fibrotic process. 

Figure 5.

 

The effects of antioxidant reagents on the formation of subretinal fibrosis. (A) Glutathione concentration at the PEC-injected site was visualized with MCB. Frozen sections were stained with MCB to monitor intracellular glutathione. The serial sections were co-stained with MCB (blue) and propidium iodide (red), to confirm the location. (B) Representative immunohistochemical staining for GFAP in a choroidal flat mount from a PBS-injected control mouse (top) or GSH-OEt-treated mouse (bottom). Scale bar, 500 μm. (C) The mean ± SEM area of fibrosis stained with GFAP, measured 1 week after injection of PECs. PBS-suspended control mice (n = 9), GSH-OEt-treated mice (n = 9; *P < 0.05). The experiment was repeated twice (18 mice each) with similar results.

To further confirm the effect of oxidative changes on fibrosis, we suspended PECs into two well-known antioxidant reagents, N-Acetyl-l-cysteine (NAC) ¹³ and glutathione monoethyl ester (GSH-OEt), ²² and injected them into the subretinal space. Both antioxidant-treated mice showed a marked reduction in subretinal fibrosis (especially with GSH-OEt) compared with the PBS-treated control (Figs. 5B, 5C), which suggests that oxidative changes in the subretinal space are a crucial process in this fibrosis model. 

Activated Macrophages Induce Conformational Changes in RPE Cells

All in vivo data indicated that activated macrophages could initiate subretinal fibrosis. To confirm that the cellular mechanism in the subretinal space in vivo could be found in vitro, we examined the morphologic changes in primary-cultured RPE cells after PECs were added. RPE cells were chosen, because they are closely located in the chorioretinal interface that is damaged in AMD and work as the first line of defense against external pathogens. RPE cells have antigen presenting properties ²³ and co-exist in human fibrotic tissues. ¹⁵  

RPE cells are known to cause myofibrotic changes themselves and to express intracellular α-SMA. ^24,25 After 24 hours of co-culture with PECs, the RPE cells became fibrous, and showed a dense concentration of α-SMA (red) in the presence of PECs (Fig. 6A). On flow cytometry, we confirmed the upregulation of α-SMA expression according to the dose of PECs (Fig. 6B). We also performed TUNEL staining, which clearly showed a few apoptotic (green) cells, even after co-culture with 1 × 106 PECs (Fig. 6A, arrows). Moreover, to visualize macrophages on the RPE cells, we double stained with α-SMA and F4/80 (Fig. 6C). F4/80+ macrophages showed completely different features from RPE cells and were not stained with α-SMA. Collectively, we conclude that activated macrophages can actually induce conformational myofibroblastic changes in RPE cells (but not in themselves) without cell death in vitro. 

Figure 6.

 

In vitro expression of α-SMA in the RPE cells cultured with PECs. RPE cells were prepared from C57BL/6 mice and cultured approximately 2 weeks until becoming confluent in a 24-well plate. Designated concentrations of PECs were added to the primary culture and stained by α-SMA 48 hours later. (A) Representative images of α-SMA-stained RPE cells (red) on the dish. Left: isotype control; middle: RPE cells, right: RPE cells + 1 × 10⁶ PECs. All three cultures received TUNEL staining (green, arrows). (B) RPE cells were detached from the dish by vigorous pipetting, stained with α-SMA, and analyzed by flow cytometry. The experiments were repeated twice and with similar result. (C) A representative image of α-SMA-stained RPE cells (red) and F4/80-stained macrophages (green) on the dish.

Discussion

In the current study, we injected peritoneal macrophages into the subretinal space and successfully established an animal model of focal subretinal fibrosis, which mimics the fibrotic subretinal scarring observed in late-stage AMD. Not only injected exogenous macrophages, but also infiltrating intrinsic macrophages are critical in the development of fibrosis. Both steroid and antioxidant treatments can reduce fibrosis. Macrophages have a potential to induce α-SMA in RPE cells, suggesting conformational myofibroblastic changes. 

To establish this mouse model, we performed several trial experiments. Using bone marrow–derived cells other than macrophages, we injected red blood cells or bead-enriched lymphocytes by the same technique immediately after photocoagulation. These cells did not cause any advanced pathogenic subretinal changes (data not shown). In contrast, if we injected primary cultured RPE cells that were prepared in the same way as the experiment in Figure 6, then the RPE cells frequently caused extensive retinal detachment with proliferative changes in the vitreous cavity—so-called proliferative vitreoretinopathy. Only thioglycolate-elicited macrophages could cause focal fibrosis that is observed in advanced AMD. 

An important point of our model was performing high-power laser photocoagulation, which induced both retinal coagulation and rupture of Bruch's membrane ahead of PEC injection. The retinal coagulation prevents further enlargement of the retinal break that was necessary to inject PECs into the subretinal space. The rupture of Bruch's membrane is known to induce CNV, which enabled easy accumulation of endogenous macrophages to the inflammatory site. Moreover, an air bubble eventually formed in the subretinal space when Bruch's membrane was ruptured. After injection, the air bubble sealed the retinal break, which prevented initial leakage of injected macrophages into the vitreous cavity. 

Another point of discussion is the purity of macrophages in our PECs. Flow cytometry analysis using F4/80-specific antibody (a marker of macrophages) showed that the ratio of macrophages in our PEC preparation was constantly higher than 90%. However, neutrophils and lymphocytes are potential contaminants in the PECs, and these cells may contribute to the subretinal fibrotic process. However, as shown in Figure 4, subretinal fibrosis did not form in MCP-1 KO mice, which lacked secondary macrophage infiltration into the injected site. We can at least say that reactively accumulated intrinsic macrophages are essential in the formation of fibrosis. 

The detailed mechanisms of how injected macrophages mediate subretinal fibrosis still has to be elucidated. However, as shown in Figure 6, macrophages have the direct ability to induce α-SMA in co-cultured RPE cells. Macrophages can produce proinflammatory cytokines including IL-1β, TNF-α, and TGF-β, ²⁶ which are known to promote myofibrosis. ²⁷ In addition, macrophages with decreased intracellular glutathione, oxidative macrophages, strongly express IL-6 and PGE2. ¹³ These soluble factors mediate focal inflammation consisting of a mixture of both resident cells (e.g., RPE cells and microglia cells) and bone marrow–derived inflammatory cells. 

Many researchers have investigated the mechanism of CNV formation and have focused on preventing CNV and/or reducing already-established CNV. However, leukocyte exudation or bleeding from abnormal vessels can cause focal inflammation and tissue fibrosis that causes a decrease in macular function. Therefore, management for CNV, but also for subsequent inflammation may be important to achieve better visual prognosis in AMD. Our present study highlighted subsequent subretinal inflammation and subretinal fibrotic scar formation in an animal model for the first time. We identified the critical role of exudate macrophages and modified the function by steroid or antioxidant reagents which minimized fibrotic changes with glial proliferation. Our study will hopefully allow better management in the future of advanced-phase AMD and other CNV-related diseases. 

Footnotes

 Supported by Japanese Ministry of Education, Science, Sports and Culture Grants B2 14770962 (K-HS) and B2 13470369 (TI) and by a contribution from Hiroki Sanui.

Footnotes

 Disclosure: Y.-J. Jo, None; K.-H. Sonoda, None; Y. Oshima, None; A. Takeda, None; R. Kohno, None; J. Yamada, None; J. Hamuro, None; Y. Yang, None; S. Notomi, None; T. Hisatomi, None; T. Ishibashi, None

The authors thank Michiyo Takahara for excellent technical support with all the experiments and Mari Imamura for help in preparing the histologic sections.