Activation of the Small GTPase Rap1 Inhibits Choroidal Neovascularization by Regulating Cell Junctions and ROS Generation in Rats

ABSTRACT

Purpose: Choroidal neovascularization (CNV) is a common vision-threatening complication associated with many  fundus diseases. The retinal pigment epithelial (RPE) cell junction barrier has critical functions in preventing CNV, and oxidative stress can cause compromise of barrier integrity and induce angiogenesis. Rap1, a small guanosine triphosphatase (GTPase), is involved in regulating endothelial and epithelial cell junctions. In this work, we explored the function and mechanism of Rap1 in CNV in vivo.

Methods: A laser-induced rat CNV model was developed. Rap1 was activated through intravitreal injection of the Rap1 activator 8CPT-2ʹ-O-Me-cAMP (8CPT). At 14 days after laser treatment, CNV size in RPE/choroid flat mounts was measured by fluorescein isothiocyanate-dextran staining. Expression of vascular endothelial growth factor (VEGF) and cell junction proteins in RPE/choroid tissues were analyzed by western blots and quantitative real-time PCR assays. Reactive oxygen species (ROS) in RPE cells were detectedbydichloro-dihydro-fluorescein diacetate assays. The antioxidant apocynin was intraperitoneally injected into rats.

Results: Activating Rap1 by 8CPT significantly reduced CNV size and VEGF expression in the rat CNV model. Rap1 activation enhanced protein and mRNA levels of ZO-1 and occludin, two tight junction proteins in the RPE barrier. In addition, reducing ROS generation by injection of apocynin, a NADPH oxidase inhibitor, inhibited CNV formation. Rap1 activation reduced ROS generation and expression of NADPH oxidase 4.

Conclusions: Rap1 activation inhibits CNV through regulating barrier integrity and ROS generation of RPE in vivo, and selectively activating Rap1 may be a way to reduce vision loss from CNV.

KEYWORDS:

• Rap1
• choroidal neovascularization
• retinal pigmentepithelial
• cell junction
• reactive oxygen species

Introduction

Choroidal neovascularization (CNV) is a major vision-threatening complication associated with several common retinal degenerative and inflammatory diseases, such as age-related macular degeneration (AMD), central exudative chorioretinopathy, pathological myopia, angioid streaks, ocular histoplasmosis syndrome, and chorioretinitis.¹ CNV often damages the macula, leading to exudation or hemorrhage and scar forming that causes visual loss.² Activated choroidal endothelial cells (CECs) originating from the choriocapillaris migrate through the Bruch membrane toward the retinal pigment epithelium (RPE), and severe vision loss occurs most often when they transmigrate the RPE into the neurosensory retina to proliferate into CNV.³ The RPE is thought to be critical for preventing CNV development. Normally, the RPE forms a monolayer connected by tight junctions that create the outer blood-retinal barrier and breakdown is a step in CNV pathogenesis.⁴ However, in aging eyes, oxidative stress, metabolic stress, hypoxia, and inflammation compromise the RPE barrier and induce angiogenesis.⁵ Specifically, excessive reactive oxygen species (ROS) induce dysregulation of functional networks and deleterious changes in the RPE that result in CNV.⁶

Rap1 is a member of the Ras guanosine triphosphatase (GTPase) superfamily. It has two isoforms, Rap1a and Rap1b, that share 95% amino acid identity.⁷ Like all GTPases, Rap1 acts as a molecular switch, cycling between an active (GTP-bound) and an inactive (GDP-bound) form through guanine nucleotide exchange factors and GTPase-activating proteins (GAPs). Cumulative evidence shows that activation of Rap1 strengthens cell junction integrity and reduces cell migration by affecting actin cytoskeleton remodeling.^8–10 Both Rap1a and Rap1b are expressed in RPE. Previous studies have showed that Rap1 expression and activity are involved in the formation and development of CNV.¹¹ In this work, we developed a rat laser-induced CNV model and activated Rap1 in vivo by a cAMP analog 8CPT-2ʹ-O-Me-cAMP (8CPT) to explore the mechanism of Rap1 regulating CNV. 8CPT specifically activates Rap1 via effects on upstream exchange factors (Epac1 and Epac2), but does not activate other cAMP-responsive pathways, e.g., Ras and ERK.¹² Our results show that Rap1 activation reduces CNV size and ROS generation of RPE, and enhances RPE cell junction in a rat laser-induced CNV model.

Materials and methods

Animals

Brown Norway (BN) rats (male, aged 8–10 weeks, 180–200 g) were from Vital River Laboratory Animal Technology Co., Ltd (Beijing, China). Prior to experiments, slit-lamp and ophthalmoscopic examinations were performed to exclude abnormalities in eyes. The study protocols were approved by the Ethical Committee of Affiliated Second Hospital of Hebei Medical University.

Laser-induced CNV model and injections

BN rats were anesthetized with intraperitoneal injections of 10% chloral hydrate (0.35 ml/kg), and pupils were dilated with 1% tropicamide and 1% mydriacyl. Using a 532 nm laser (Coherent Novus Omni, Palo Alto, CA, USA), a slit-lamp delivery system, and a coverslip as a contact lens, six spots (300 mW, 200 μm, and 50 ms) were placed in each eye. Lesions were 2–3 disk diameters from the optic nerve head. Development of a bubble under laser radiation confirmed rupture of Bruch’s membrane. Lesions with hemorrhage were excluded. Right after laser injury, a 1 μl intravitreal injection of 8CPT (20.5 μM; Calbiochem, Darmstadt, Germany) was delivered to activate Rap1 in vivo. In separate rats, phosphate buffer saline (PBS) intravitreal injection was the control. Ten rats were used in each group, and both eyes of each rat were lasered and injected with the same reagent. To inhibit ROS generation in vivo, an intraperitoneal injection of apocynin (10 mg/kg/d, Sigma–Aldrich, St. Louis, MO, USA) was delivered daily for 14 days. PBS injection was the control.

Measurement of CNV and CNV lesions

At 14 days after laser treatment, animals were perfused with 1 ml PBS containing 50 mg/ml fluorescein isothiocyanate-labeled dextran (FITC-dextran, average MW 2 × 10⁶; Sigma–Aldrich) before sacrifice. Eyes were harvested and fixed for 1 h in 10% phosphate-buffered formalin. Cornea, lens, and vitreous were removed. Posterior eyecups of the RPE, choroid, and sclera were dissected and flat-mounted on microscope slides by cutting radial incisions. CNV images were captured using a laser scanning confocal microscope (Bio-Rad, Hercules, CA, USA), and images at different focal planes were stacked and flattened. Lesions with obvious hemorrhage or bridging were excluded. The fluorescent area of lesion spot was quantified with ImageJ software (National Institutes of Health, Bethesda, MD, USA). Ten lesions from at least three rats per group were analyzed.

Western blots

RPE/choroid tissues were triturated, lysed on ice in RIPA buffer, and boiled in SDS loading buffer. Protein extracts were subjected to 6–12% SDS-PAGE and processed with a standard protocol. Primary antibodies for GTP-Rap1 (1:1000, NewEast Biosciences, King of Prussia, PA, USA), vascular endothelial growth factor (VEGF, 1:1000, Santa Cruz Biotechnology, Santa Cruz, CA, USA), zonula occludens-1 (ZO-1, 1:2000, Santa Cruz Biotechnology), occludin (1:2000, Santa Cruz Biotechnology), NADPH oxidase 4 (NOX4, 1:1000, Abcam, Cambridge, UK), and β-actin (1:4000, Abclone, Wuhan, China) were used.

Quantitative real-time PCR assays

Total RNA was extracted from RPE/choroid tissues and reverse transcribed to cDNA. Quantitative real-time PCR (qPCR) analysis was performed with SYBR Premix Ex Taq™ kits (TAKARA Bio, Dalian, China) on an ABI7500 Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). Relative gene expression was normalized to GAPDH with the 2^−ΔΔCt method. The qPCR primer sequences are in Table 1.

Table 1. Primers for qPCR.

Gene	Forward primers (5ʹ-3ʹ)	Reverse primers (5ʹ-3ʹ)
VEGF	CAGGACCATGCACCATAGCA	GCAGTGCTCACCTCTAACGA
ZO-1	AGTCTCGGAAAAGTGCCAGG	GGGCACCATACCAACCATCA
NOX4	CAGCAGAGAAGCACCTAGCA	CACAGCACAGCCACTTGTTC
Occludin	ACAGGTTTGCAGGACTCGAC	AACTCCTGAGAAGCGCTGTG
GAPDH	AAGAGGGATGCTGCCCTTAC	TACGGCCAAATCCGTTCACA

Measurement of ROS productions

Rat RPE cells were isolated according to the method of Heller et al.¹³ Isolated cells were immediately incubated in 5 μM 2,7-dichlorodihydro-fluorescein diacetate (DCFH-DA, Invitrogen, Carlsbad, CA, USA) in serum-free medium for 30 min at 37°C and washed twice with PBS. DCFH-DA is cell-permeable and hydrolyzed intracellularly to the DCFH carboxylate anion; this is retained in cells, and induced by ROS to form a fluorescent product, dichlorofluorescein (DCF). ROS-induced DCF fluorescence was measured with a fluorescent plate reader with excitation 485 nm and emission 520 nm. Simultaneously, fluorescence emitted by cells on slides was observed with a confocal microscope. Cells incubated with 10 μM H₂O₂ were the positive control.

Statistical analysis

All experiments were performed in triplicate. Data are means ± SEM, and differences between groups were analyzed with independent-sample t-tests (two-tailed, GraphPad Prism 5, GraphPad Software, Inc., La Jolla, CA, USA). P < 0.05 was considered statistically significant. One representative set of data from western blots and fluorescent images were shown.

Results

Rap1 activation inhibits laser-induced CNV

To explore the effect of Rap1 on CNV, 8CPT, a known Rap1 activator, was used in a rat CNV model. We confirmed that 8CPT treatment activated Rap1 in rats. Western blots showed that active Rap1 increased after 8CPT intravitreal injection (Figure 1(A,B)). We observed the response of rats to laser injury after 8CPT treatment. CNV was induced by photocoagulation using an argon laser, and all spots were positive for CNV at 14 days after laser treatment. Confocal analyses of RPE-choroid-scleral flat mounts by FITC-dextran revealed significantly smaller CNV lesion areas in 8CPT-injected groups than in PBS control groups (Figure 1(C,D)). Western blots showed VEGF expression in RPE/choroid tissue was significantly decreased in 8CPT-injected rats compared with control rats (Figure 1(E,F)).

Figure 1. Rap1 activation inhibits laser-induced CNV in rats.

A. Western blot of GTP-Rap1 in RPE/choroid tissues from CNV rats with PBS and 8CPT injection. B. Quantification of results in A. Protein levels normalized against Rap1, n = 3. **P < 0.01. C. Representative confocal images of RPE/choroid flat mounts stained with FITC-dextran in PBS and 8CPT groups. Scale bar = 100 μm. D. Quantification of CNV area. n = 10 lesions. ***P < 0.001. E. Western blot of VEGF in RPE/choroid tissues from CNV rats with PBS and 8CPT injection. F. Quantification of results in E. Protein levels normalized against β-actin, n = 3. **P < 0.01.

Rap1 activation strengthens RPE cell junctions

Previous reports implicated Rap1 in cell junction regulation. We detected expression of ZO-1 and occludin, components of the tight junction complex, after 8CPT treatment. Western blots showed that ZO-1 and occludin protein in RPE/choroid tissue was significantly increased in 8CPT-injected rats compared with control rats (Figure 2(A,B)). qPCR indicated that ZO-1 and occludin mRNA also increased with 8CPT (Figure 2(C)).

Figure 2. Rap1 activation enhances expression of RPE cell junction proteins.

A. Western blot of ZO-1 and occludin in RPE/choroid tissues from CNV rats with PBS and 8CPT injection. B. Quantification of results in A. Protein normalized against β-actin. C. Quantitative real-time PCR for ZO-1 and Occludin mRNA in two groups, n = 3. *P < 0.05. **P < 0.01. ***P < 0.001.

Inhibition of ROS generation reduces CNV

To confirm the effect of ROS in CNV, we used apocynin, an NADPH oxidase inhibitor, to reduce ROS generation in CNV. After apocynin injection, NOX4 expression was reduced in rat RPE/choroid by western blot (Figure 3(A,B)). ROS generation was measured as DCF fluorescence in the RPE cells of rats treated with intraperitoneal apocynin or PBS. Fluorescence was reduced in the apocynin-injected group (Figure 3(C,D)). Administration of apocynin reduced CNV lesion size in rats (Figure 3(E,F)).

Figure 3. Inhibiting ROS generation with apocynin reduces CNV.

A. Western blot of NOX4 in RPE/choroid tissues from CNV rats with PBS and apocynin injection. B. Quantification of results in A. Protein normalized against β-actin, n = 3. **P < 0.01. C. Representative image of ROS generation in RPE cells from CNV rats with PBS and apocynin injection. RPE cells were isolated and cultured in DCFH-DA, and DCF fluorescence from cells on slides observed by confocal microscope. Scale bar = 20 μm. D. Quantification of ROS generation. ROS-induced DCF fluorescence measured with a fluorescence plate reader with excitation 485 nm and emission 520 nm, n = 3. *P < 0.05. E. Representative confocal images of RPE/choroid flat mounts stained with FITC-dextran in PBS and apocynin groups. Scale bar = 100 μm. F. Quantification of CNV area, n = 10 lesions. ***P < 0.001.

Rap1 activation reduces ROS generation in the RPE at CNV lesions

To determine if activation of Rap1 reduced CNV area by inhibiting ROS generation, ROS in the RPE cells was measured in rats treated with intravitreal 8CPT or control. The strength of DCF fluorescence in the RPE cells was significantly decreased in rats that received 8CPT injections compared to PBS (Figure 4(A,B)). Moreover, 8CPT injection caused a decrease in NOX4 protein and mRNA by western blot and qPCR (Figure 4(C–E)). These results suggested that activation of Rap1 could reduce CNV formation by inhibiting ROS generation in the RPE.

Figure 4. Rap1 activation reduces ROS generation in RPE at CNV lesions.

A. Representative image of ROS generation in RPE cells from CNV rats with PBS and 8CPT injection. RPE cells were isolated and cultured in DCFH-DA, and DCF fluorescence emitted by cells on slides was observed by confocal microscope. Scale bar = 20 μm. B. Quantification of ROS generation. ROS-induced DCF fluorescence was measured with a fluorescence plate reader with excitation 485 nm and emission 520 nm, n = 3. *P < 0.05. C. Western blot of NOX4 in RPE/choroid tissues from CNV rats with PBS and 8CPT injection. D. Quantification of results in C. Protein normalized against β-actin, n = 3. *P < 0.05. E. Quantitative real-time PCR of NOX4 mRNA in two groups, n = 3. *P < 0.05.

Discussion

In this work, we showed that activating Rap1 by 8CPT, a known Rap1 activator, reduced CNV size and VEGF expression in a rat laser-induced CNV model. Rap1 activation enhanced expression of ZO-1 and occludin, two tight junction proteins in the RPE barrier. Furthermore, Rap1 activation reduced NOX4 expression and ROS generation. These results demonstrated that Rap1 activation inhibited CNV through regulating barrier integrity and ROS generation in the RPE (Figure 5).

Figure 5. Schematic of Rap1 in regulating CNV. Rap1 inhibits CNV through enhancing cell junctions and decreasing ROS generation in RPE cells. An alternative mechanism may exist in which Rap1 inhibits CNV by regulating these function of CECs.

The RPE maintains the outer blood-retinal barrier. Intact barrier integrity is important for the health and function of the retinal photoreceptors and choriocapillaris.¹⁴ The barrier comprises a single layer of epithelium cells that have tight junctions. Impairment of junctional regulation interferes with compartmentalization of angiogenic agonists normally secreted from the RPE basal aspect, and inhibitors secreted from the apical aspect, by permitting movement of factors to other compartments.¹⁵ This change in concentrations of angiogenic factors such as VEGF activates motile CECs. Tight junctions consist of multiprotein complexes including transmembrane protein components, such as occludin and cadherins, and cytoplasmic protein components, such as ZO-1 and β-, α-, and p120-catenin. Some of these bind to the actin cytoskeleton.¹⁶ Of the signaling proteins involved in junctional regulation, small GTPases are focused because of their ability to cycle between active and inactive states. Small GTPases of the Rho family regulate cell junctions by affecting actin cytoskeleton remodeling.^17,18

For the small GTPase Rap1, Hartnett ME and colleagues reported that in the RPE line ARPE-19, knockdown or inhibition of Rap1 activity impairs barrier functions. This result was shown with reduced transepithelial electrical resistance and electrical impedance, mislocalization of cadherins, formation of gaps within the monolayer, and higher transmigration of CECs.¹⁵ These results suggest Rap1 regulates RPE junctional barrier function and might potentiate CNV. For the specific functions of Rap1a and Rap1b isoforms, in vitro, Rap1a short hairpin RNA (shRNA) educes steady-state barrier integrity, whereas Rap1b shRNA affects dynamic junctional responses. In vivo, Rap1b^−/− mice have larger CNV volumes compared to wild-type or Rap1a^−/−.¹¹ In addition, increasing Rap1a activity specifically in RPE cells by self-complementary adeno-associated virus 2 vectors reduces CNV volume in vivo and inhibits CEC transmigration across RPE monolayers in vitro.¹⁹ Our data showed activating Rap1 reduced CNV size and VEGF expression, and thus inhibited CNV development in a rat laser-induced CNV model. Activating Rap1 also enhanced the RPE barrier by increasing protein and mRNA for ZO-1 and occludin, two tight junction proteins. Our conclusion was consistent with previous reports from the Hartnett laboratory. We used a rat model and different experiments to further confirm the critical function of Rap1 in CNV formation.

The barrier function of vascular endothelial cells maintained by adherens and tight junctions is central for regulating angiogenesis and inflammation.^20,21 In microvascular endothelial cells, Rap1 regulates the accumulation of adherens and tight junction proteins to the cell–cell contact sites.^17,21,22 In CNV, besides the RPE barrier, Rap1 may also regulate the barrier function of CECs to affect angiogenesis.^15,23 Therefore, an alternative interpretation to increased ZO-1 and occludin expression by 8CPT is that CECs have more stable tight junctions, and thus they are less likely to remodel and migrate (Figure 5). This could yield an inhibition to CNV lesion. We need further data to identify the role of Rap1 in RPE and CECs.

Multiple risk factors, including obesity, hypertension, smoking, and light exposure, are reported to contribute to AMD pathogenesis through increasing oxidative stress generated by several mechanisms, including constant exposure to light and ROS.²⁴ ROS cause cell damage directly but can also trigger intracellular signal transduction that leads to various outcomes in different cell types involved in AMD.²⁵ Hartnett et al. demonstrated that in CECs, overexpressing active Rap1a by 8CPT inhibits TNF-α-induced CEC migration by inhibiting NADPH oxidase-dependent NF-kB and Rac1 activation. This result suggests that Rap1a de-escalates CNV development by interfering with ROS-dependent signaling.²³ Similarly, in RPE inhibition of Rap1 by Rap1GAP increases ROS generation, whereas activation of Rap1 by 8CPT reduces ROS by interfering with NADPH oxidase assembly.²⁶ Our results in a rat laser-induced CNV model showed that reducing ROS generation by antioxidant apocynin treatment inhibited CNV formation in rats. Rap1 activation decreased NOX4 expression and ROS generation, suggesting that Rap1 affected CNV by regulating oxidative stress. Our results confirmed and extended the conclusion of Wang et al.²⁶

In conclusion, our findings indicated that Rap1 GTPase is important in CNV through regulating RPE junctional barrier function and ROS generation in vivo. Our results suggested that selectively activating Rap1 may be a way to reduce vision loss from CNV and improve clinical outcomes.

Declaration of interest

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this article.

Additional information

Funding

Key Basic Research Program of Applied Basic Research Program of Hebei Province (15967741D).