Abstract

Diabetic macular edema is major cause of vision loss associated with diabetic retinopathy. Breakdown of blood-retinal barrier, especially inner BRB, is an early event in pathogenesis of DR. Apelin, an endogenous ligand of APJ, mediates angiogenesis and is involved in the development of DR. The present study aimed to investigate effects and mechanism of apelin-13 in vascular permeability during DME. We verified apelin-13 was upregulated in DME patients’ vitreous. High glucose incubation led to a progressive increase of apelin-13, APJ, cytoskeleton, and tight junction proteins, including VE-Cadherin, FAK, Src, ZO-1, and occludin. Apelin-13 promoted HRMEC proliferation and migration and phosphorylation of both cytoskeleton and tight junction under both normal and high glucose conditions. Besides, apelin-13 activated PI-3K/Akt and MAPK/Erk signaling pathways, including PLCγ1, p38, Akt, and Erk both in HRMEC and in C57BL/6 mice. Meanwhile, F13A performed opposite effects compared with apelin-13. In in vivo study, apelin-13 was also upregulated in retina of db/db mice. Taken together, apelin-13 increased biologic activity of HRMEC, as well as expression of both cytoskeleton and tight junction in DME via PI-3K/Akt and MAPK/Erk signaling pathways. Apelin-13 as an early promoter of vascular permeability may offer a new perspective strategy in early treatment of DR.

1. Introduction

Diabetes mellitus (DM) is one of the most important chronic diseases worldwide. By the year 2030, about 440 million people in the age group of 20–79 years are estimated to be suffering from diabetes mellitus worldwide (prevalence 7.7%), while in 2010 there were 285 million people with diabetes mellitus (prevalence 6.4%) [1]. DR is one of the most serious and common microvascular complications of DM and the predominant cause of new emerging blindness in adults between 20 and 74 years all over the world [2]. Clinically, DR is classified into nonproliferative DR (NPDR) and proliferative DR (PDR), with diabetic macular edema (DME) present in both types [3]. DME results from breakdown of the blood-retinal barrier (BRB), especially the inner BRB which constitutes of vascular endothelial cells, causing leakage of plasma constituents into the surrounding retina and inducing retinal edema. As macular edema develops, moderate to severe vision loss occurred [3]. The incidence of DME is 20% in patients with younger onset diabetes over a 10-year pathogenesis versus approximately 40% in older onset diabetes. Nowadays, the treatment of DME and DR is mainly limited to laser photocoagulation, vitrectomy, and intravitreal injection of drugs, such as anti-VEGF drugs and corticosteroids; the first two therapies cannot avoid the possibility of recurrence; and the incidence of complications of the last remains to be further clear.

Apelin is a recently isolated bioactive growth factor from bovine gastric extract regarding as an endogenous ligand for APJ and able to mediate angiogenesis and vascular formation [4]. Current investigations demonstrated that apelin played critical roles in the occurrence and development of DM and its complications [4]. Apelin encodes a preproprotein of 77 amino acids, which generate several active polypeptides [4]. The relative potency of the apelin peptides varies between experimental systems, (Pyr 1)-apelin-13 and apelin-13 being the most potent activators of apelin receptor APJ [4]. Our previous study demonstrated that apelin-13 promoted proliferation and migration of retinal Müller cells [5], human RPE cells [6], pericytes [7], and retinal microglial BV2 cells [8], offering new perspectives in the prevention and treatment of DR [9] and PDR [10]. Apelin has been reported to promote the phosphorylation of extracellular signal-regulated kinases (Erks), protein kinase B (Akt), and p70S6 kinase in umbilical endothelial cells [11]. However, there is no information about the role of apelin in the expression of cytoskeleton and tight junction proteins in human retinal microvascular endothelial cells (HRMECs) in DME.

In the present study, we investigated the effect and signaling pathways of apelin-13 in cytoskeleton and tight junction in vitro and in vivo. According to our results, we identified that apelin-13 was upregulated in DME patients. Exogenous apelin-13 increased the biologic activity of HRMECs, as well as the phosphorylation of both cytoskeleton and tight junction proteins via upregulating PI-3K/Akt and MAPK/Erk signaling pathways. Apelin-13 as an early promoter of vascular permeability probably offers a new perspective strategy in the early prevention and treatment of DME and DR.

2. Materials and Methods

2.1. Reagents

Human exogenous recombinant apelin-13 peptide (number 057-30, St. Louis, MO, USA) and the special antagonist of apelin receptor APJ, F13A (number 427197, St. Louis, MO, USA) were purchased from Sigma. Antibodies used in the present study were shown in Table 1. All experiments were performed in accordance with the Research Ethics Committees of People’s Hospital Peking University and with the approval of People’s Hospital Eye Institute.

2.2. Human Material

The Ethical Committee and Institutional Review Board of Peking University People’s Hospital (Beijing, China) approved the human patient study protocol (the ethical approving number is 2012-23), which was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from each study subject. Inclusion criteria [12] included (1) age between 40 and 70 years, (2) first-episode DME or epiretinal membrane (ERM) patients, untreated with laser photocoagulation, vitrectomy, or other surgery within half a year, (3) a central macular thickness (CMT) greater than 250 μm based on optical coherence tomography (OCT), and (4) the fewest possible chronic pathologies other than diabetes. The medical records of patients who were diagnosed with DME and received intravitreal injections of bevacizumab (Avastin; Genentech; off-label usage, 1.25 mg) (IVB) at the Ophthalmology Department of Peking University People’s Hospital (Beijing, China) were used. The vitreous humor (100 μl) of DME patients was collected during IVB treatment, while the vitreous humor (100 μl) from ERM patient was collected from vitrectomy. In total, 4 DME patients and 4 controls (ERM) were enrolled in this study.

2.3. Cell Culture and Treatments

Human retinal microvascular endothelial cells (HRMECs, Angio-Proteomie, USA), which are primary cells, were used in this study. HRMECs were cultured in special endothelial cell medium (endothelial cell medium [ECM], number 1001, ScienCell, USA) supplemented with 5% fetal bovine serum (FBS; ScienCell, USA) as recommended by the manufacturer. To evaluate the effect of apelin-13 on treatment-naïve HRMECs, the cells were cultured with apelin-13 (0, 1, 10, 100, and 1000 ng/ml) [5] for 24 h before assessments; to evaluate the effect of apelin-13 on glucose-treated HRMECs, median or high glucose (15.5 mM or 25 mM) [13] were incubated with cells for 48 h and then treated with F13A (0, 0.2, 2, 20, and 200 ng/ml) [7] for another 24 h before assessments. The cells were used between passages 3 and 6.

2.4. Animals and Treatments

The animal study was approved by Peking University People’s Hospital Ethics Committee. Animals were purchased from Beijing Key Laboratory of Innovative Drug Discovery of Traditional Chinese Medicine (Natural Medicine) and Translational Medicine. Eighty db/db mice and 40 C57BL/6J mice were kept in a humidity-controlled room on a 12 h light-dark cycle with food and water available ad libitum. The db/db mice were then divided randomly into four groups with 10 animals in each group. The C57BL/6J mice were fed with a standard rodent chow (Harlan Teklad Mouse/Rat Diet 7002) [14]. The db/db mice were fed with a high-fat diet that contained 60% fat, 14% protein, and 26% carbohydrate [15]. To evaluate the effect of apelin/APJ system on the retina of mice, F13A (100 ng/ml, 1.5 μL/eye) [9] was continuously intravitreal injected into the right eye every month until sacrifice, and injection was started from 1 month after the mice’s birth. Vehicle-treated animals received a single injection of phosphate-buffered saline of the left eye were treated as controls. The retinal vascular permeability was assessed at 1, 3, 6, 9 months in adult db/db mice with or without treatment with F13A using western blot and immunofluorescence.

2.5. Cell Viability/Cell Proliferation

The cell proliferation capacity was assessed using MTS (3-4, 5-dimethylthiazol-2-yl]-53-carboxymethoxyphenyl]-2-4-sulfopheny]-2H-tetrazolium inner salt, one kind of new synthetic tetrazole compounds) assay according to manufacturer’s instructions (CellTiter96 Aqueous One Solution Assay; Promega, Madison, WI, USA). HRMECs were starved in ECM with 1% FBS for 12 h before being cultured into 96-well plates. Cells were pretreated with apelin (0, 1, 10, 100, and 1000 ng/ml) for 24 h or were incubated with median or high glucose for 48 h and then treated with F13A (0, 0.2, 2, 20, and 200 ng/ml) for another 24 h with 5% fetal bovine serum (FBS, Gibco, USA) before assessments. MTS reagent (10 μl/well) was added to the culture medium, and the cells were incubated in 96-well plates for an additional 4 h. The absorbance was measured at 490 nm. Each experiment was repeated at least six times, and each sample was measured in triple wells.

2.6. Cell Migration/Transwell Assay

The HRMEC migration study was performed using transwell (Cat. number 3422; Corning, Tewksbury, MA) as described [16]. Briefly, HRMECs in 200 μL serum-free ECM were added to the upper chamber and 600 μL ECM (containing 5% FBS) with apelin (0, 1, 10, 100, and 1000 ng/ml) or high glucose with F13A (0, 0.2, 2, 20, and 200 ng/ml) to the lower chamber, respectively. The chambers were incubated at 37°C for 6 h. The filters were fixed and stained with 4,6-diamidino-2-phenylindole (DAPI, 1 : 5000; number D9542, Sigma, USA). The remaining cells on the upper surface of the filter were removed with a cotton swab gently. The number of migrated cells were quantified by counting in five random fields (10x magnification), using a Nikon 50i (Nikon, Tokyo, Japan) fluorescence microscopy. The data are shown as the mean ± standard deviation (SD). Each experiment was repeated at least three times.

2.7. Immunofluorescence Staining

HRMECs, pretreated with apelin (0, 1, 10, 100, and 1000 ng/ml) in normal (4.5 mM) or high glucose ECM, were cultured on cover glasses (Fisher, USA) and then fixed with acetone for 10 min [17]. Primary antibodies were incubated overnight at 4°C. The primary antibodies used for immunofluorescence (IF) staining were rabbit anti-apelin, anti-VE-Cadherin, anti-ZO-1, and anti-occludin (Table 1). After blocking, the sections were washed and then incubated with relevant fluorescence-conjugated secondary antibodies (1 : 1000) (Table 1). The samples were counterstained with DAPI. The images were acquired using a Nikon 50i microscope with NIS-Elements Imaging Software under 200x magnifications.

For mice retina detection, IF frozen sections (8 mm thick) were blocked with 1% BSA (30 min–1 h, RT), and zonula occludens-1 (ZO-1) and occludin (Table 1) were used. For double-labeling IF study, they were then incubated with the relevant fluorescence-conjugated secondary antibodies (Table 1). The samples were counterstained with DAPI. IF was acquired using a Nikon 50i microscope.

2.8. Quantitative Real-Time PCR

HRMECs were lysed with TRIzol (Invitrogen, Carlsbad, CA, USA), and total RNA was extracted according to the manufacturer’s protocol. The concentration and integrity of total RNA were detected with UV spectrophotometry (NANODROP 2000C, Thermo, USA). Reverse transcriptase reactions were performed using a Thermo Revert Aid TM First Strand cDNA Synthesis Kit (K1622, Thermo, USA). Real-time PCR reactions were performed with Select Master Mix (Cat. number 4472908, Invitrogen, USA) using a real-time PCR system (Piko Real 96 PCR system, Thermo Scientific, USA). Each sample was measured in triplicate wells. The target gene primers are shown in Table 2. Data were normalized to the housekeeping genes human glyceraldehyde 3-phosphate dehydrogenase (GAPDH). We calculated the changes in mRNA expression according to the method, with ΔCT = and ΔΔCT = . Each experiment was repeated at least three times.

2.9. Western Blot Analysis

Briefly, HRMECs and retinal tissue were prepared with RIPA lysis buffer (number P0013B, Beyotime, Shenzhen, Guangdong, China). After centrifugation, the supernatant was collected, and the protein lysate was measured with a bicinchoninic acid (BCA) protein assay kit (Pierce) according to the manufacturer’s protocol. Equal amounts of protein were separated by 8% or 10% NuPAGE Bis-Tris Gel (Cat. number NW00080BOX, NP0315BOX, Invitrogen, USA) and transferred electrophoretically to polyvinylidene fluoride (PVDF) membranes (Millipore, Billerica, MA, USA). The proteins were visualized using an enhanced chemiluminescence (ECL) substrate (PerkinElmer, Inc., MA, USA). Primary antibodies concentration was used as shown in Table 1, followed by incubation with secondary antibody. Each experiment was repeated at least three times.

2.10. Statistical Analysis

The results were expressed as mean ± SD. Difference between two groups was compared with Student’s -test (SPSS17.0 software, Chicago, IL). Differences among groups were assessed using one-way analysis of variance (ANOVA, Prism version 5.0 [GraphPad Software, Inc., San Diego, CA, USA]), followed by Dunnett’s test. The clinical data was analyzed using repeated measurement analysis of variance. A value of less than 0.05 was considered to be statistically significant.

3. Results

3.1. Apelin Is Increased in the Vitreous Humor of Patients Suffering from Diabetic Macular Edema

Table 3 presented the demographics of patients selected for vitreous biopsy. 4 patients diagnosed with DME and 4 ERM controls were involved in the detection of apelin in the vitreous humor with western blot. Figure 1(a) showed significant retinal swelling, especially in the macular and perimacular zones compared to normal control, with ocular fundus color image and spectral-domain optical coherence tomography (SD-OCT). Western blot detection and analysis of vitreous from patients revealed that apelin was robustly increased in patients with DME (Figure 1(b)).

3.2. Apelin Induces Cell Proliferation, Migration, and Phosphorylation of Cytoskeleton and Tight Junction Proteins in HRMECs

Experiments were performed to evaluate whether apelin-13 produced an effect on HRMECs proliferation and migration in normal glucose. MTS and transwell assay were used to detect the proliferation and migration in HRMECs. According to the results, HRMECs proliferation and migration capability were increased in a dose-dependent manner, significantly higher in the apelin group, compared with the control one, and the optimum concentration was 100 ng/ml (Figures 2(a)2(c)). Cytoskeleton (VE-Cadherin) in HRMECs was visualized with fluorescence microscopy after immunofluorescence staining with rabbit polyclone anti-VE-Cadherin antibody (red). Cells were stained with DAPI for visualization of nuclei (blue). Our IF staining showed that the phosphorylation of VE-Cadherin was upregulated after being treated with apelin in HRMECs (Figure 2(d)). Two optimum concentrations (10 ng/ml and 100 ng/ml) were used to verify the phosphorylation changes of cytoskeleton and tight junction proteins with western blot; the results showed that the phosphorylation of cytoskeleton and tight junction were also upregulated in a dose-dependent way after being stimulated with apelin (Figures 2(e)2(g)).

3.3. High Glucose Increases the Expression of Apelin and the Phosphorylation of Cytoskeleton and Tight Junction Proteins in HRMECs

In order to mimic the in vivo environment, HRMECs were incubated with different glucose concentrations of ECM medium. The mRNA expression of apelin (a), APJ (b), vascular endothelial (VE)-Cadherin (c), focal adhesion kinase (FAK) (d), and Src (e) were determined using QRT-PCR. The Quantitative Real-Time PCR results showed that the expression of apelin, its receptor APJ, VE-Cadherin, FAK, and Src increased with median and high glucose ECM medium, compared with normal glucose one (Figures 3(a)3(e)). Likewise, staining intensities of apelin (f) in HRMECs under normal, median, and high glucose conditions were performed using IF and the results revealed that the expression of apelin increased glucose concentration independently (Figure 3(f)). APJ, the phosphorylation of cytoskeleton proteins, such as VE-Cadherin, FAK, and Src also increased under median and high glucose conditions, compared with normal glucose one using western blot (Figures 3(g)3(j)).

3.4. Inhibition of Apelin Receptor APJ Using F13A Decreased Cell Proliferation, Migration, and the Phosphorylation of Cytoskeleton and Tight Junction Proteins in HRMECs under High Glucose Condition

HRMECs with F13A (0.2, 2, 20, and 200 ng/mL) proliferation was determined with MTS (a). HRMECs migration in response to F13A was measured using a transwell assay (b and c, 200x magnification). Tight junction (occludin and ZO-1) in HRMECs was visualized with fluorescence microscopy after immunofluorescence staining with mouse monoclone anti-occludin antibody (d) (red) and mouse monoclone anti-ZO-1 antibody (e) (green). According to the MTS and transwell assay results, the HRMECs proliferation and migration capability were decreased in a dose-dependent manner, significantly lower in the F13A group, compared with the control one, and the optimum concentration was 20 ng/ml (Figures 4(a)4(c)). Our IF staining showed that the phosphorylation of occludin and ZO-1 was downregulated after being treated with F13A in HRMECs under high glucose condition (Figures 4(d) and 4(e)). Two optimum concentrations (2 and 20 ng/ml) were used to verify the phosphorylation changes of cytoskeleton and tight junction proteins with western blot; the results showed that the phosphorylation of cytoskeleton and tight junction were also downregulated in a dose-dependent way after stimulated with F13A (Figures 4(f) and 4(g)).

3.5. Apelin Plays Roles in HRMECs via PI-3K/Akt and MAPK/Erk Signaling Pathways

The PI-3K/Akt and MAPK/Erk are the most significant signaling factors involved in stimulated endothelial cells and also appear to play critical roles in angiogenesis [11]. Phosphorylation of Akt and Erk is involved in proliferation, migration, vascular remolding, and angiogenesis [18]. In order to illuminate the mechanism of apelin to HRMECs and retinal vascular permeability in the genesis and development of DME, we detected the protein phosphorylation of PI-3K/Akt and MEK/Erk signaling pathway such as PLC1, p38, Akt, and Erk using western blot. The results suggested that elevated expression of phosphorylated PLCγ1, p38, Akt, and Erk in apelin-treated (10 and 100 ng/ml) HRMECs, compared with untreated control ones (Figures 5(a)5(d)). On the contrary, F13A played the opposite role under high glucose condition (Figures 5(i)5(l)). High glucose also increased the phosphorylation of PI-3K/Akt and MEK/Erk signaling pathway proteins (Figures 5(e)5(h)).

3.6. Apelin Induces the Expression of Cytoskeleton and Tight Junction Proteins in C57/BL6 Mice via PI-3K/Akt and MAPK/Erk Signaling Pathways

In order to determine the effectiveness of apelin on BRB and its direct signaling pathways, we detected APJ expression (a), phosphorylation of cytoskeleton (VE-Cadherin and FAK) (b and e), tight junction (occluding and ZO-1) (c and d), and PI-3K/Akt and MAPK/Erk signaling pathways (p38, Akt and Erk) (f–h) proteins with western blot in C57BL/6 mice. Apelin (10 and 100 ng/ml) was injected in vitreous of C57BL/6 mice for 10 days according to our previous study [10]. Our results found that the expression of APJ and the phosphorylation of cytoskeleton (VE-Cadherin and FAK) and tight junction proteins (occludin and ZO-1) increased significantly in apelin-treated (10 and 100 ng/ml) C57BL/6J mice (Figures 6(a)6(e)). Besides, the phosphorylation of PI-3K/Akt and MEK/Erk signaling pathways p38, Akt, and Erk also increased significantly (Figures 6(f)6(h)).

3.7. Inhibition of Apelin Receptor APJ with F13A Dramatically Reduces Pathological Vascular Permeability in db/db Mice via PI-3K/Akt and MAPK/Erk Signaling Pathways

F13A, the inhibitor of APJ, was used to investigate the effect of apelin in early stage DR animal model. We detected the changes of F13A treatment using IF and western blot. The phosphorylated and total protein of PLCγ1, p38, Akt, and Erk were detected to verify the role of PI3K/Akt and MAPK/Erk signaling pathways. According to our IF results, the phosphorylation of tight junction proteins (occluding and ZO-1) decreased in F13A treated groups (Figures 7(a) and 7(b)). The western blot results revealed that the expression of apelin receptor APJ was higher in diabetic retinopathy mice, compared with the F13A treated groups (Figure 8(a)). The results also showed that treatment with F13A produced a remarkable reduction in the phosphorylation of cytoskeleton (VE-Cadherin and FAK) and tight junction protein (occludin) (Figures 8(b)8(d)). Furthermore, the phosphorylation of PI-3K/Akt and MEK/Erk signaling pathways also decreased significantly in F13A-treated db/db mice (Figures 8(e)8(h)).

4. Discussion

There are three main findings in the present study. First, we reported that apelin is upregulated in the vitreous of DME patients; second, we detected that apelin induced the proliferation and migration of HRMECs and meanwhile promoted the protein phosphorylation of cytoskeleton and tight junction under both normal and high glucose condition; finally, we demonstrated that apelin activated the PI-3K/Akt and MEK/Erk signaling pathways both in vitro and in vivo. Above all, apelin/APJ system is involved in the pathology of DR, and apelin-13 is an early promoter of cytoskeleton and tight junction in DME via PI-3K/Akt and MAPK/Erk signaling pathways.

In 1998, Tatemoto et al. purified apelin from bovine stomach extracts a novel adipokine that binds to the oligo G-protein coupled receptor APJ [4]. Studies proved that apelin/APJ mainly localized in vascular endothelial cells (ECs), adipose tissue, and epithelial cells [19], which is involved in cardiovascular regulation, fluid homeostasis, modulation of the adipoinsular axis, and HIV coreceptor function [20]. In ocular diseases, Kasai et al. first reported that apelin is an angiogenic factor in retinal endothelial cells [21] and suggested that spatiotemporally regulated apelin/APJ signaling participates in retinal vascularization in a cooperative manner with vascular endothelial growth factor (VEGF) or fibroblast growth factor 2 (FGF2) and contributes to normal ocular development [22]. Yonem et al. demonstrated that apelin is upregulated at the leading edge of retinal vessel formation [23]. Our previous studies verified that apelin/APJ system is involved in retinal neovascularization development in PDR patients [10]. However, the function of apelin in DME, which is the most common complication and early event of DR, is not fully understood. Our data on human subjects provide the rationality that apelin plays roles in the development of DME, which is the expression of apelin in vitreous fluid of the DME patients being higher (Figure 1). This finding correlated with previous studies which suggested that expression of apelin was significantly higher in the plasma of diabetes mellitus patients [2426].

As stated before, DME is derived from breakdown of blood-retinal barrier (BRB). The BRB has two components, the retinal vascular endothelium (inner BRB) and the retinal pigment epithelium (RPE) (outer BRB). This physiological barrier has tight junctions which can prevent the circulation component from leaking into the eye. It is strongly suggested that the inner BRB is the primary site of the vascular leak that results in DME in both human and animal studies [27, 28]. Thus, the relation between apelin/APJ in DME is further studied on HRMECs and diabetic animal model in the present study.

Our in vitro studies focused on the cytoskeleton and tight junction of microvascular endothelial cells, which are key factors to maintain the normal function of BRB. Any abnormity of BRB will induce retina edema. We first demonstrated that apelin/APJ was positively expressed in HRMECs. Then, we proved that apelin induced the proliferation and migration and upregulated the phosphorylation of cytoskeleton (VE-Cadherin, FAK, and Src) and tight junction proteins (ZO-1 and Occludin) in HRMECs (Figure 3), which indicated that apelin induced loosening of the BRB. As demonstrated before, proliferation and migration are key factors for angiogenesis and cell attachment [29]. VE-cadherin is an endothelial-specific cell-cell adhesion protein of the adherents junction complex, plays a critical role on endothelial barrier function and angiogenesis, and also sustains the stability of adherents junction of the endothelial cells [30]. FAK, a nonreceptor protein tyrosine kinase, is a pivotal mediator of angiogenesis, deletion of FAK proved by the lethality of the early embryonic mice [31]. Src kinase is a protooncogene encoding a protein tyrosine kinase that regulates cellular morphology, motility, metabolism, survival, migration, and proliferation [32]. In accordance with other reports, several research also showed that apelin was involved in vessel endothelial proliferation and migration, such as human umbilical microvascular endothelial cell (HUMEC), vascular endothelial cell (VEC), and retinal endothelial cell (REC) [11, 21, 33]. Considering that hyperglycemia is the major risk factor of DME, the expression of apelin/APJ in HRMECs under glucose culture condition was evaluated. Apelin/APJ and phosphorylation of cytoskeleton and tight junction proteins were all increased under high glucose condition, compared with the control group (Figure 2). After adding F13A, the phosphorylation of FAK and occludin decreased significantly under high glucose condition, which suggested the negative effect of apelin/APJ on vascular permeability, which can promote cell disruption under hyperglycemia condition and loosen the cell-to-cell connections.

To further investigate the protective effects of apelin on retinal vascular, more in vivo studies were performed. We found that the phosphorylation of cytoskeleton and tight junction proteins increased in apelin-treated C57BL/6 mice (Figure 6), which is consistent with in vitro HRMECs results and further confirmed our hypothesis of the negative effect of apelin on BRB. Besides, db/db mice were used as an early stage of DR animal model, which have been demonstrated that the breakdown of BRB is performed as early as at eight weeks [34]. Apelin receptor APJ and cytoskeleton and tight junction were higher in diabetic retinopathy mice, which correlated with lots of studies [19, 35, 36]. After being treated with F13A, the phosphorylation of these proteins was decreased (Figure 4), the same as our previous STZ-induced diabetic C57/BL/6J mice study [9], which suggested that apelin weakened the vascular permeability directly.

Furthermore, we investigated the underlying mechanisms of vascular permeability by apelin/APJ in HRMECs and retina of DR. The activation of Akt has been shown to be a downstream effector of apelin via a PTX-sensitive G-protein and PKC [11]. Apelin-13- mediated Akt phosphorylation in HUVECs [37], osteoblasts [38], rat hippocampal neurons [39], mouse cortical neurons [40], rat and human VSMCs respectively [41, 42]. ERK1/2 is another reported downstream of apelin also. In pluripotent embryonic stem cells [43] and mouse embryonic endothelial cells [44]. Our previous study showed that apelin promoted proliferation, migration, and collagen I expression through the PI3K/Akt and MEK/Erk signaling pathways in human RPE cells [6]. Wang et al. demonstrated that endothelial progenitor cells (EPCs) and neural progenitor cells (NPCs) synergistically protects cerebral ECs against H/R injury in barrier function (tube formation and permeability) via PI3K/Akt signaling pathways [45]. Our results showed that the phosphorylation of PI-3K/Akt and MEK/Erk was significantly higher with apelin-treated group (Figures 3 and 4) both in vitro and in vivo. In particular, in db/db mice, inhibition of APJ could downregulate the phosphorylation of cytoskeleton (VE-Cadherin and FAK), tight junction (Occludin), and PI-3K/Akt and MEK/Erk signaling pathways.

In summary, our data indicate that apelin plays a negative role in cytoskeleton and tight junction during the pathogenesis of DME through the PI-3K/Akt and MEK/Erk signaling pathways. Although our study has some limitations, including apelin/APJ knockout mice should be used to verify the protection effects of apelin in DME, the present results provided the rationality that apelin-13 is an early promoter of vascular permeability which offers a new perspective strategy in the early prevention and treatment of DME.

Ethical Approval

The Ethical Committee and Institutional Review Board of Peking University People’s Hospital (Beijing, China) approved the human patient study protocol (the ethical approval number is 2012-23), which was conducted in accordance with the Declaration of Helsinki.

Conflicts of Interest

The authors declare that there are no conflicts of interest regarding the publication of this paper.

Authors’ Contributions

Yu-jing Bai and Yan-rong Jiang contributed to research design. Yang Li, Yu-jing Bai, Li Chen, Wenzhen Yu, Xuan Shi, and Gui-bo Sun conducted the experiments. Yang Li and Yu-jing Bai assisted in the data analysis. Yang Li and Yu-jing Bai wrote or contributed to writing the manuscript. Yu-jing Bai, Yan-rong Jiang, and Gui-bo Sun reviewed the manuscript. Yan-rong Jiang obtained the grant. Yang Li and Yu-jing Bai contributed equally to this paper and are co-first author.

Acknowledgments

This work was supported by the National Natural Science Foundation of China (81271027) and the EFSD/CDS/Lilly Grant (2127000043).