Intravitreal Anti-VEGF Therapy Blocks Inflammatory Cell Infiltration and Re-Entry into the Circulation in Retinal Angiogenesis

Abstract

Purpose.: Anti–VEGF-A antibody (Ab) (e.g., bevacizumab, ranibizumab) is widely used as a treatment against retinal angiogenesis and edema. The purpose of this study was to evaluate whether intravitreal anti-VEGF Ab injection modulates inflammatory cells in retinal angiogenesis.

Methods.: To investigate whether intravitreal bevacizumab injections affect the number of inflammatory cells in proliferative diabetic retinopathy (PDR) membranes in patients, immunohistochemical staining with CD45 Ab (pan-leukocyte marker) was performed using the surgically obtained membranes in pars plana vitrectomy with or without pretreatment with bevacizumab. To check whether anti–VEGF-A Ab affects leukocytes going in and out of blood vessels during retinal angiogenesis, the authors performed their novel leukocyte transmigration assay and CD45 immunostaining in a mouse model of oxygen-induced retinopathy (OIR).

Results.: The authors' new imaging approach revealed that intravitreal injection of anti–VEGF-A Ab blocks leukocyte infiltration as well as angiogenesis. The Ab injection inhibited leukocyte transmigration before affecting the angiogenenic area. CD45 staining showed no significant difference in the leukocyte number in the angiogenic retina or the human PDR membranes between the anti–VEGF-A Ab injected group and the control group. Furthermore, VEGF-A inhibition also affected leukocytes going out from the retina.

Conclusions.: Intravitreal injection of anti-VEGF-A Ab could inhibit leukocyte trafficking in the retina, suggesting that anti-VEGF-A therapy could serve as a treatment in retinal inflammation.

Introduction

Leukocyte infiltration is an important step that occurs during angiogenesis in inflammatory diseases. ^1,2 Various studies have reported that infiltrated leukocytes are a prerequisite for inflammatory angiogenesis, including diabetic retinopathy. ^3,4  

VEGF-A promotes retinal neovascularization in mice as well as in patients with diabetic retinopathy. ^5,6 VEGF-A is also a chemoattractant for leukocytes, especially macrophages, via the VEGF receptor(R)-1 (flt-1) in vitro. ⁷ In vivo VEGF-A–dependent angiogenesis is also accompanied by leukocyte infiltration. ²  

The use of intravitreal anti–VEGF-A treatment (e.g., bevacizumab, ranibizumab) has increased in recent years in the management of various complications in retinal diseases. A histologic study of fibrovascular membranes in patients with proliferative diabetic retinopathy (PDR) showed that intravitreal anti–VEGF-A induces apoptosis and lessens fenestration in vascular endothelial cells, indicating that the treatment affects vascular endothelial cells. ^8,9 Leukocytes are a key component of fibrovascular membranes in PDR. ^10,11 However, the impact of anti–VEGF-A agents on leukocytes have not been examined in neovascular tissues from PDR patients. Intravitreal injection of anti–VEGF-A agents are associated with ocular complications, including endophthalmitis or uveitis ^12,13 ; however, whether intravitreal injection of anti–VEGF-A agents affects leukocyte infiltration from retinal angiogenesic vessels is unknown. Furthermore, whether the angiostatic effect of anti–VEGF-A Ab is related to reduced inflammation is unclear. 

Materials and Methods

Mouse Model of Oxygen-Induced Retinopathy (OIR)

All experimental procedures on the animals were performed according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. OIR was induced in C57BL/6J mice as described in detail by Connor et al. ¹⁴ Briefly, litters of postnatal day seven (P7) C57BL/6J pups along with their mothers were placed in a 75% ± 2% oxygen atmosphere (hyperoxia) for 5 days and then returned to room air at age P12. After the pups were returned to room air, and the avascular areas of retina became hypoxic, intraretinal physiological revascularization of the avascular areas and preretinal pathological neovascularization (NV) developed simultaneously. Pathological NV reached its maximum at P17. Mice were killed by cervical dislocation, and the eyes were enucleated. 

Quantification of Avascular Areas and Neovascular Tuft

Quantification of the avascular areas and neovascular tuft areas in P17 OIR retinas was performed according to a described protocol. ¹⁴ After the mouse VEGF-A Ab (50 ng/1 μL; AF-493-NA; R&D Systems, Minneapolis, MN) or nonimmune IgG was injected into P16 OIR eyes, the retinas were flat-mounted at 8 or 24 hours after the Ab injection, and stained with fluorescein-labeled isolectin B4 (1:150 dilution; F-1201; Vector Laboratories, Burlingame, CA). The retinas were photographed under a fluorescent microscope (BZ-9000; KEYENCE, Osaka, Japan) and merged to obtain whole-mount retinal images using image-joint software (BZ-Analyzer software; KEYENCE). 

Ex Vivo Leukocyte Transmigration Assay

The experimental mice were anesthetized at P16 in an OIR model. One hundred μL acridine orange (AO) (1 mg/mL; A6014; Sigma, St. Louis, MO) was injected into the heart. Untreated mice (4 week old) were injected with 200 μL AO. The AO concentration in the intravascular leukocytes and the endothelial cells significantly diminished 30 minutes after dye injection due to a washout effect. In contrast, transmigrated leukocytes retained their staining, which allowed for their visualization. At 2, 4, or 8 hours after AO injection, the eyes were enucleated and the retinas were carefully removed and fixed with 4% paraformaldehyde for 30 minutes at 4°C, and flat mounts were prepared using a mounting medium (TA-030-FM, Mountant Permafluor; Lab Vision Corporation, Fremont, CA). To examine leukocyte extravasated from the retina, the authors injected AO at 2 hours before VEGF-A Ab or IgG intravitreal injection. After 6 hours, the retinas were harvested. The flat mounts were examined by fluorescence microscopy, and digital images were recorded using a fluorescent microscope (BZ-9000; KEYENCE) with standardized illumination and contrast. AO-positive transmigrated leukocytes (outside the vessels) in angiogenic areas or peripheral areas were counted manually. 

PDR Membrane Collection

This study was approved by the institutional ethics committees, and the surgical specimens were handled in accordance with the Declaration of Helsinki. Fifteen preretinal fibrovascular membranes from 14 patients with PDR were included if there were fibrovascular membranes that required pars plana vitrectomy at Kyushu University and Fukuoka University Chikushi Hospital. The clinical data of patients are summarized in Table 1. All patients had neovascularization of the disc and neovascularization elsewhere noted along the arcade vessels. In cases of uncontrolled hypertension, or in cases with a history of thromboembolic events, cerebral insult or renal diseases, no off-label use of intravitreal bevacizumab was offered. Intravitreal injection of bevacizumab (1.25 mg) was performed 5.3 ± 2.5 (2–20) days before surgery. Surgically removed membranes were fixed in 4% paraformaldehyde at 4°C for 24 hours and were subjected to histologic analysis. 

Table 1. 

 

Patient Characteristics before Intravitreal Bevacizumab Injection for Proliferative Diabetic Retinopathy

Characteristic	IVB(–)	IVB(+)	P
n	8 (7 patients)	7 (7 patients)
Age, y	57 (37–69)	52 (30–71)	0.565*
Duration of DM, y	20 (9–30)	14 (2–30)	0.281*
HbA1c, %	7.1 (4.5–9.9)	8.8 (4.5–13.1)	0.224*

 Values are expressed as median (range) or number (%).

*  Mann-Whitney U test.

CD45 Immunohistochemical Analysis

Surgically removed membranes were fixed in 4% paraformaldehyde, embedded in paraffin wax, cut into 5-μm sections, and deparaffinized according to standard procedures. Heat-induced epitope retrieval was performed by immersing sections of tissue in citrate buffer (pH 6.0). The sections were blocked with a 3% nonfat dried milk bovine working solution for 30 minutes. Tissues were incubated overnight at 4°C with anti-human CD45 mAb (1/400, M0701; Dako, Carpinteria, CA) in PBS. Tissues were washed three times for 5 minutes in PBS followed by incubation with Alexa Fluor647 goat anti-mouse IgG (10 μg/mL, A21235; Invitrogen, Carlsbad, CA) for 30 minutes at room temperature. The number of CD45(+) leukocytes in the PDR membrane was counted manually and PDR membrane areas were measured using Photoshop (Adobe, Mountain View, CA). 

Whole Mount Immunofluorescence

The mice eyes were enucleated and fixed with 4% paraformaldehyde for 30 minutes at 4°C. For whole mount preparation, the retinas were microsurgically exposed by removing other portions of the eye. Tissues were washed with PBS, three times for 5 minutes and then placed in methanol for 20 minutes. Tissues were incubated overnight at 4°C with fluorescein-labeled isolectin B4 (1:150 dilution; Vector Laboratories, Burlingame, CA), anti-mouse CD45 mAb (6.25 μg/mL; 550539; BD Pharmingen, San Diego, CA) and anti–VEGF-A Ab (10 μg/mL; sc-152; Santa Cruz Biotechnology, Santa Cruz, CA) in PBS containing 10% goat serum and 1% Triton X-100. Tissues were washed four times for 20 minutes in PBS followed by incubation with Alexa Fluor488 goat anti-rat IgG (20 μg/mL; A11006; Invitrogen) overnight at 4°C. Radial cuts were then made in the retina. Retinal flat mounts were prepared on glass slides using a mounting medium (TA-030-FM, Mountant Permafluor; Lab Vision Corporation). The flat mounts were examined by fluorescence microscopy and digital images were recorded using a fluorescent microscope (BZ-9000; KEYENCE) with standardized illumination and contrast. 

Statistical Analysis

Values are expressed as mean ± SEM. Data were analyzed by Student's t-test. Baseline characteristics of each group in diabetic retinopathy (DR) patients were compared with the Mann-Whitney U test or the χ ² test. Differences between the experimental groups were considered statistically significant or highly significant, when P was less than 0.05 or less than 0.01, respectively. 

Results

Effect of Intravitreal Injection of Anti–VEGF-A Ab on Retinal Angiogenesis in Mice

The authors first examined whether the intravitreal injection of anti–VEGF-A Ab inhibits retinal angiogenesis in a mouse model of OIR. In line with their previous report, ¹⁵ the neovascular area in mice with the intravitreal injection of anti–VEGF-A Ab was significantly smaller than in mice with control Ab 24 hours after Ab injection (Figs. 1A, 1B). The intravitreal injection did not increase the size of the avascular area significantly (Fig. 1C). These data indicate that after 24 hours, intravitreal anti–VEGF-A Ab treatment inhibits retinal neovascularization. 

Figure 1. 

 

Angiostatic effect of intravitreal anti-VEGF Ab injection. (A) Representative images of a flat-mounted OIR retina with IgG or αVEGF Ab treatment (P17). IgG or αVEGF Ab was administrated at P16. Arrows indicate retinal neovascularization. (B, C) Quantitation of neovascular area (B) or avascular area (C) with 50 ng IgG or αVEGF Ab treatment (P17) (n = 5). *P < 0.05.

Effect of Intravitreal Injection of Anti–VEGF-A Ab on Leukocyte Transmigration in Mice

To next investigate the effect of intravitreal injection of anti–VEGF-A Ab on leukocyte transmigration during retinal neovascularization, the authors performed a novel assay that they developed for in vivo quantification of leukocyte infiltration into the retina. They observed very few AO(+) transmigrated leukocytes in nonischemic retinas without angiogenesis (0.1 ± 0.1 leukocytes/mm², n = 27 from three mice) (Fig. 2A). Furthermore, they also confirmed most AO(+) leukocytes could exist outside of the vessels in OIR model (see Supplementary Material and Supplementary Fig. 1). At 24 hours after the intravitreal injection of anti–VEGF-A Ab or control Ab, the authors examined AO(+) transmigrated leukocytes in the angiogenic areas of the retina (Fig. 2B). Intravitreal anti–VEGF-A Ab treatment significantly inhibited AO(+) leukocyte transmigration into the angiogenic retina (Figs. 2D, 2E). However, the injection could not affect AO(+) leukocyte transmigration in the nonangiogenic retina significantly (see Supplementary Material and Supplementary Figs. 2A and 2B). These data suggest that intravitreal anti–VEGF-A Ab treatment blocks leukocyte transmigration as well as angiogenesis. 

Figure 2. 

 

Impact of anti-VEGF Ab on leukocyte infiltration in retinal angiogenesis. (A) Double staining of retinal flat mounts for transmigrated leukocytes (AO, green) and perfused blood vessels (concanavalin A [ConA], red) in 4-week-old mice (untreated). Bar equals 100 μm. (B) Time courses of the experimental protocol. IgG or αVEGF Ab was treated at P16. After 24 hours of Ab injection, AO was injected, and 2 hours after, AO(+) leukocyte was examined. (C) Dotted square indicates the examined angiogenic area in the OIR model (P17). (D) Representative images of AO(+) cells (arrows) in a flat-mounted OIR retina with 50 ng IgG or αVEGF Ab treatment (P17). Bar equals 200 μm. (E) Quantitation of the number of AO(+) cells in retinal angiogenesis (n = 4–6). **P < 0.01.

Effect of Intravitreal Injection of Anti–VEGF-A Ab on Leukocyte Transmigration Prior to Affecting Angiogenesis

To examine whether the anti-inflammatory effect of anti–VEGF-A Ab is due to less angiogenesis, the authors first checked the neovascular area at an earlier time after the Ab injection. Eight hours after the Ab injection, there was no significant difference in the neovascular area, or in the avascular area between anti–VEGF-A Ab- and IgG-treated eyes (Figs. 3A–C). They checked the AO(+) leukocyte number, before angiogenesis was affected significantly by the treatment (Fig. 3D). The number of AO(+) infiltrated cells in angiogenic area in the anti–VEGF-A Ab-treated retinas was less than in the IgG-treated retinas (Figs. 3E, 3F) However, the authors could not observe a significant difference of AO(+) cell numbers in the peripheral nonangiogenic area between anti–VEGF-A Ab-treated retinas. These results suggest that the angiostatic effect of anti–VEGF-A Ab is partially due to early blockage of leukocyte infiltration in angiogenesis. 

Figure 3. 

 

Direct effect of anti-VEGF Ab on leukocyte infiltration. (A) Representative images of a flat-mounted OIR retina with IgG or αVEGF Ab treatment (P16 + 8h). IgG or αVEGF Ab (50 ng) was administrated at P16. Arrows indicate retinal neovascularization. (B, C) Quantitation of neovascular area (B) or avascular (C) area with IgG or αVEGF Ab treatment (P16 + 8h) (n = 6). (D) Time courses of the experimental protocol. IgG or αVEGF Ab was treated at P16. After 8 hours of Ab injection, AO was injected, and 2 hours after AO(+) leukocyte was examined. (E) Representative images of AO(+) cells (arrows) in a flat-mounted OIR retina with IgG or αVEGF Ab treatment (P16 + 10h). Bar equals 200 μm. (F) Quantitation of the number of AO(+) cells in retinal angiogenesis (n = 4–6). *P < 0.05.

Effect of Intravitreal Injections of Bevacizumab on the Leukocyte Number in the PDR Membrane

A total number of eight membranes from seven patients with PDR with bevacizumab pretreatment, and seven membranes from seven patients with PDR without the pretreatment were included (Table 1). To check the effect of intravitreal bevacizumab on leukocytes in the PDR membrane, the authors performed immunohistochemistry with CD45 Ab (Fig. 4A). Unexpectedly, the immunostaining showed that the PDR membranes with bevacizumab pretreatment contained more leukocytes than membranes without the pretreatment; however, the difference did not reach statistical significance (Fig. 4B). Furthermore, to check whether the results in the human tissue and the mouse model are congruent, they investigated the number of CD45(+) cells in the mouse retina with VEGF-A Ab and the control IgG treatment. Surprisingly, the CD45(+) cell number around neovascular tufts in VEGF-A Ab-administered retinas was similar to the control retina (Figs. 4C, 4D). 

Figure 4. 

 

Impact of bevacizumab on the leukocyte number in the preretinal membrane of diabetic retinopathy patients and mouse retinal angiogenesis. (A) Representative images of CD45 (red) and 4′,6-diamidino-2-phenylindole (DAPI) (blue) immunostaining in PDR membrane with or without the intravitreal bevacizumab treatment. Bar equals 200 μm. (B) Quantitation of the number of CD45(+) leukocytes per millimeters squared in the PDR membrane with or without the intravitreal bevacizumab treatment (n = 7). (C) Representative images of lectin (green) and CD45 (red) immunostaining in a flat-mounted OIR retina with IgG or αVEGF Ab treatment (P17). The dotted circle indicates neovascularization. The dotted line indicates border between avascular and vascular area. Bar equals 200 μm. (D) Quantitation of the number of CD45-positive leukocytes per millimeters squared in OIR retinas with IgG or αVEGF Ab treatment (n = 12).

Effect of the Intravitreal Injection of Anti–VEGF-A Ab on Leukocyte Re-Entry into the Circulation

The authors' observation led to the hypothesis that intravitreal injection of anti–VEGF-A Ab stops infiltrated leukocytes from leaving the retina. To examine the hypothesis, they first injected AO and followed the time course of AO(+) cells outside the vessels during retinal angiogenesis. The number of AO(+) cells outside the vessels significantly increased 4 hours after AO injection, then the number significantly decreased at 8 hours (Figs. 5A, 5B). These data indicate that the AO assay could detect leukocyte emigration as well as extravasation from the retina during angiogenesis. To check that anti–VEGF-A Ab could block the leukocyte extravasation from the angiogenic retina, the authors administrated VEGF-A Ab and control IgG after AO-stained leukocytes infiltrated the retina (Fig. 5C). Interestingly, the number of AO(+) cells outside the vessels in anti–VEGF-A Ab-treated eyes was more than in IgG-treated eyes (Figs. 5D, 5E). This suggests that leukocyte re-entry into the circulation from the angiogenic retina is partially inhibited through VEGF-A inhibition. 

Figure 5. 

 

Impact of bevacizumab on leukocytes in the preretinal membrane of diabetic retinopathy patients. (A) Double staining of retinal flat mounts for transmigrated leukocytes (AO, green) and perfused blood vessels (ConA, red) in the OIR model (P17) 2, 4, or 8 hours after AO injection. Representative images of AO(+) cells (arrows) in the flat-mounted retina. Bar equals 100 μm. (B) Quantitation of the number of AO(+) cells (outside the vessels) 2, 4, or 8 hours after AO injection (n = 11–25). (C) Time courses of the experimental protocol. AO was injected at P16. Two hours after the AO injection, IgG or αVEGF Ab (50 ng) was treated, and at 6 hours AO(+) leukocytes were counted. (D) Representative images of AO(+) cells (arrows) in a flat-mounted OIR retina with IgG or αVEGF Ab treatment (P17 + 8h). Bar equals 200 μm. (E) Quantitation of the number of AO(+) cells in retinal angiogenesis (n = 4). **P < 0.01, *P < 0.05.

Discussion

Leukocyte recruitment is an essential part of angiogenic disorders including PDR. ^2,3,16,17 The authors' recent study showed anti–VEGF-A Ab could block leukocyte transmigration in corneal inflammation. ¹⁸ Therefore, they expected that bevacizumab-treated PDR membranes would show less leukocyte infiltration than untreated membranes. However, intravitreal injections of anti–VEGF-A Ab did not reduce the leukocyte number in PDR patients, or during mouse angiogenesis. These results are in line with a previous report regarding the effect of bevacizumab in human choroidal neovascularization. ¹⁹ That article showed a higher leukocyte density in the bevacizumab-treated group compared to the control; however, the results were not statistically significant. Tatar et al. ¹⁹ suggested leukocyte infiltration by several mechanisms, including endothelial apoptosis or immune response to the fragment crystallizable (Fc) portion of bevacizumab. ¹⁹ The current study's data indicate that another explanation to why VEGF-A inhibition increases leukocyte number could be due to the blockage of leukocyte re-entry into the circulation. 

The authors' observation with AO staining showed leukocyte trafficking within 8 hours during retinal angiogenesis. The leukocyte turnover rate in the retina with angiogenesis is higher than in the normal retina ²⁰ and VEGF-A inhibition makes the rate slower. Because VEGF-A inhibition induced less fenestration and apoptosis of retinal vessels, ⁸ anti–VEGF-A Ab treatment might affect various molecules to modulate leukocyte adhesion and the blood–ocular barrier in the retinal blood vessels. ²¹  

In this study, the human PDR membrane and mouse OIR model were examined. VEGF-A inhibition could block leukocyte trafficking in the mouse model. Furthermore, the authors could confirm transmigrated leukocytes expressed VEGF-A (see Supplementary Material and Supplementary Fig. 2), suggesting extravasated leukocytes contribution to retinal angiogenesis in the OIR model. However, their observation in the OIR model was not that diabetic status and diabetic retinal endothelium has been reported to be damaged by leukostasis. ²² Many other compounding factors might also affect the mechanism of leukocyte trafficking in the diabetic retina. 

Recent clinical reports showed endophthalmitis in patients occurred after anti–VEGF-A Ab administration. ^12,13 The authors' data showed that VEGF-A inhibition reduces leukocyte infiltration as well as re-entry into the circulation. Leukocyte trafficking, especially macrophages and dendritic cells, have been implicated in the maintenance of immunological homeostasis in the retina. ^20,23 The side effects of the anti–VEGF-A agent might be due to a break down of the immune homeostasis in the retina, which might increase the rate of infection. 

Furthermore, previous articles showed macrophage is important for retinal angiogenesis. ^3,24 Another article indicated that T lymphocyte-mediated immune responses could inhibit retinal angiogenesis. ¹⁶ Anti–VEGF-A treatment might affect certain leukocyte populations by inhibition of the infiltration as well as re-entry into the circulation. 

Diabetic retinopathy is accepted as a low-grade inflammatory disease. ^4,22 This study provides a detailed analysis of leukocyte trafficking in anti–VEGF-A treated eyes in mice and human diabetic patients. The authors' results show that VEGF-A inhibition blocks leukocyte infiltration into the retina and, thus, could benefit both low-grade and acute inflammation in diabetic patients. 

Supplementary Materials

pdf S1 

Acknowledgments

The authors thank Mari Imarura and Michiyo Takahara for their technical assistance. 

Footnotes

 Supported by grants from the Ministry of Education, Culture, Sports, Science, and Technology, Japan, and Okinaka Memorial Institute for Medical Research, Japan.

Footnotes

 Disclosure: S. Nakao, None; M. Arima, None; K. Ishikawa, None; R. Kohno, None; S. Kawahara, None; M. Miyazaki, None; S. Yoshida, None; H. Enaida, None; A. Hafezi-Moghadam, None; T. Kono, None; T. Ishibashi, None