Erianin alleviates diabetic retinopathy by reducing retinal inflammation initiated by microglial cells via inhibiting hyperglycemia-mediated ERK1/2–NF-κB signaling pathway

Abstract

Blood-retinal barrier (BRB) breakdown is a typical event in the early stage of diabetic retinopathy (DR). This study aims to elucidate the protection of erianin, a natural compound isolated from Dendrobium chrysotoxum Lindl, against DR development. Erianin alleviated BRB breakdown and rescued the reduced claudin1 and occludin expression in retinas from streptozotocin-induced diabetic mice. Erianin reduced microglial activation, ERK1/2 phosphorylation, NF-κB transcriptional activation, and the elevated TNF-α expression both in vitro and in vivo. ERK1/2 inhibitor U0126 abrogated NF-κB activation in d-glucose–treated BV2 cells. Erianin reduced cellular glucose uptake, and molecular docking analysis indicated the potential interaction of erianin with glucose transporter (GLUT)1. GLUT1 inhibitor (STF31) reduced the activation of the ERK1/2–NF-κB signaling pathway. Coculture with d-glucose–stimulated microglial BV2 cells and with TNF-α stimulation both induced inner BRB and outer BRB damage in human retinal endothelial cells and APRE19 cells, but erianin improved all these damages. In summary, erianin attenuated BRB breakdown during DR development by inhibiting microglia-triggered retinal inflammation via reducing cellular glucose uptake and abrogating the subsequent activation of the downstream ERK1/2–NF-κB pathway. Moreover, erianin also alleviated BRB damage induced by TNF-α released from the activated microglia.—Zhang, T., Ouyang, H., Mei, X., Lu, B., Yu, Z., Chen, K., Wang, Z., Ji, L. Erianin alleviates diabetic retinopathy by reducing retinal inflammation initiated by microglial cells via inhibiting hyperglycemia-mediated ERK1/2–NF-κB signaling pathway.

Diabetes mellitus (DM) is a chronic and multisystem disease that requires continuing medical care (1). Medical advances over the past years have greatly increased the survival rate of patients with diabetes, and the life expectancy of patients has almost doubled since the 1940s (2–4). However, the increased survival rate also brings an enhanced risk of the high incidence of diabetes-related microvascular complications. Diabetic retinopathy (DR) is the most common microvascular disease among all these diabetes-related microvascular complications (5). Recent studies showed that the number of people with diabetes will reach ∼600 million by 2040, and one-third of them are expected to have DR (6, 7). DR has remained the leading cause of visual loss and new-onset blindness in working-age adults (8).

Nonproliferative DR (NPDR) is the early stage of DR. In this stage, because of the breakdown of the blood-retinal barrier (BRB), the loss of pericytes in the capillaries, and the thickening of basement membrane, the vascular permeability in retinas is increased (9, 10). Moreover, the death of retinal endothelial cells in the early stage of diabetes will contribute to the BRB breakdown (11). NPDR will further progress into proliferative DR (PDR) that is associated with an increased risk of visual loss owing to the formation of neovascular vessels, which can lead to vitreous hemorrhage and fibrous metaplasia of the endothelium and ultimately cause traction retinal detachment (9). Diabetic macular edema may occur at any stage of DR, and it is caused by the accumulation of exudative fluid due to the disruption of BRB (12).

A recent study showed that inflammation is a main pathologic feature of DR, and inhibiting inflammation has been found to obviously alleviate DR development (13, 14). Increased expression of proinflammatory cytokines and the adhesion molecules like intracellular adhesion molecule 1 will lead to the activation of immune cells and induce chronic inflammation in retinas and thus eventually cause the disruption of BRB function (15). Resident microglial cells can be regarded as the immunologic watchdogs in the retina and play critical roles in retinal degenerative diseases (16). The activation of microglial cells has already been found in patients with diabetes and in animal models of DR, and microglia have been considered key players during DR development (17–19). Erianin is the major bibenzyl compound found in the traditional Chinese medicine Dendrobium chrysotoxum Lindl. Erainin is generally used as the chemical marker for the quality control of D. chrysotoxum, which is a species of medicinal Dendrobium indexed in the Chinese Pharmacopoeia (2010 version) (20). According to the records of ancient medical books in China, Dendrobium has the function to improve eyesight in clinic. Our previous studies have already demonstrated that D. chrysotoxum has the capacity to reduce retinal inflammation and neoangiogenesis during DR development (21, 22). Moreover, erianin was also reported to attenuate PDR development by inhibiting retinal neoangiogenesis (23). However, whether erianin also inhibits retinal inflammation in the early-state NPDR and its engaged mechanism remain unknown. In this study, we elucidated the crucial role of microglia-triggered inflammatory injury in hyperglycemia-induced BRB breakdown during DR development and the alleviation of erianin.

MATERIALS AND METHODS

Reagents and antibodies

Erianin was purchased from Shanghai Tauto Biotech (Shanghai, China) (purity, ≥98.0%). Antibodies for p65 (p65 subunit of NF-κB), phosphorylated (p)65 (Ser536), phosphorylated NFκB inhibitor (p-IкB)-α (Ser32/36), phosphorylated inhibitor of nuclear factor κ-B (p-IKK)-α or -β (Ser176/180), phosphorylated RAF proto-oncogene (p-cRaf) (Ser338), phosphorylated MAPkinse-ERK kinase1/2 (p-MEK1/2) (Ser221), total-extracellular regulated protein kinase 1/2 (t-ERK1/2), p-ERK1/2 (Thr202/Tyr204), Lamin B1, and β-actin were all purchased from Cell Signaling Technology (Danvers, MA, USA). Antibody for ionized calcium-binding adapter molecule 1 (Iba1) was purchased from GeneTex (Alton Parkway, Irvine, CA, USA). Antibodies for claudin1 and occludin were purchased from Santa Cruz Biotechnology (Dallas, TX, USA). Peroxidase-conjugated goat anti-rabbit IgG (H+L) and anti-mouse IgG (H+L) were purchased from Jackson ImmunoResearch Laboratories (West Grove, PA, USA). Alexa Fluor 488 goat anti-rabbit IgG were purchased from BD Biosciences (Franklin Lakes, NJ, USA). Antibody of cluster of dfferentiation 11b/c (OX42) used for the immunofluorescence staining assay was purchased from ABclonal Technology (Woburn, MA, USA). Nuclear and cytoplasmic extraction reagents, and BCA Protein Assay Kits were purchased from Thermo Fisher Scientific (Waltham, MA, USA). The whole-cell protein extraction kits and the ECL kits were obtained from MilliporeSigma (Burlington, MA, USA). ELISA kits were obtained from R&D Systems (Minneapolis, MN, USA). Kits for detecting glucose content were obtained from Yeli Biotech (Shanghai, China). Trizol and DAPI were purchased from Thermo Fisher Scientific. PrimeScript RT Master Mix and Sybr Premix Ex Taq were purchased from Takara Bio (Kusatsu, Japan). U0126 was purchased from Enzo Biochem (Farmingdale, NY, USA). GLUT1 inhibitor STF31 and NF-κB inhibitor QNZ were purchased from Selleckchem (Houston, TX, USA). FITC-conjugated dextran, d-glucose, and other reagents unless noted were purchased from MilliporeSigma.

Experimental animals

C57BL/6 male mice (18–22 g; 6 wk age) were purchased from the Shanghai Laboratory Animal Center of the Chinese Academy of Sciences. Animals were maintained under controlled temperature (23 ± 2°C), humidity (50%), and lighting (12-h light/dark cycle). The mice were fed with a standard laboratory diet and given free access to tap water. All animals received humane care according to the institutional animal care guidelines approved by the Experimental Animal Ethical Committee of Shanghai University of Traditional Chinese Medicine.

Treatment of animals

Mice were unfed for 12 h before streptozotocin (STZ) injection. Mice were intraperitoneally injected with STZ (55 mg/kg) for 5 consecutive days, whereas the other 50 mice were intraperitoneally injected with physiologic saline and served as control animals. Blood glucose concentration was measured at 7 d after the last injection, and the mice with high-glucose concentration (>16.5 mM) were considered in the further experiments. For the 2-mo experiment, 45 mice were randomly divided into 3 groups: DR model (n = 15), DR + erianin (1 mg/kg) (n = 15), and DR + erianin (10 mg/kg) (n = 15), respectively. At 1 mo after STZ injection, mice were intraperitoneally injected with erianin (1, 10 mg/kg/d) consecutively for 1 mo. For the 3-mo experiment, the other 45 mice were also randomly divided into 3 groups: DR model (n = 15), DR + erianin (1 mg/kg) (n = 15), and DR + erianin (10 mg/kg) (n = 15), respectively. At 2 mo after STZ injection, mice were intraperitoneally injected with erianin (1, 10 mg/kg/d) consecutively for 1 mo. To observe the effects of erianin alone on normal nondiabetic mice, 21 mice were randomly divided into 3 groups: control (n = 7), erianin (1 mg/kg) (n = 7), and erianin (10 mg/kg) (n = 7), respectively. Mice were intraperitoneally injected with erianin (1, 10 mg/kg/d) consecutively for 1 mo.

The body weight was monitored, and blood glucose concentration was determined by glucometer (Accu-Check Performa Nano; Roche, Basel, Switzerland) during the whole experimental process (the result of blood glucose concentration was shown in Supplemental Fig. S1A). After treatment, the mice were anesthetized by sodium pentobarbital (30 mg/kg, i.p.), the blood samples were taken from the abdominal aorta, and the eyes were removed immediately.

Evan’s blue leakage assay

Retinal vascular permeability in mice with STZ-induced diabetes was detected by using Evan’s blue leakage assay as described in our previously published study (22).

Cell culture

Primary human retinal endothelial cells (HRECs) were purchased from ScienCell (Carlsbad, CA, USA) and cultured in extracellular matrix medium (The concentration of glucose in extracellular matrix medium is 5.5 mM) supplemented with 5% (v/v) fetal bovine serum (FBS), 1% endothelial cell growth supplement, 100 U/ml penicillin, and 100 mg/ml streptomycin. ARPE19, a cell line of retinal pigment epithelia cells, was gifted from Prof. Yu Chen (Shanghai University of Traditional Chinese Medicine, Shanghai, China) and cultured in DMEM-F12 medium supplemented with 10% (v/v) FBS, 100 U/ml penicillin, and 100 mg/ml streptomycin. Cells were incubated in a humidified atmosphere of 5% CO₂ at 37°C. Experiments were performed on HRECs and ARPE19 cells from passages 4–10.

Microglial cell culture and d-glucose treatment

Microglial BV2 cells were gifted from Prof. Xinhua Liu (Fudan University, Shanghai, China) and cultured in DMEM high-glucose medium supplemented with 10% FBS, 100 U/ml penicillin, and 100 μg/ml streptomycin. Cells were incubated in a humidified atmosphere of 5% CO₂ at 37°C. BV2 cells were treated with d-glucose (25 mM), and 25 mM mannitol was used as an isotonic contrast.

Primary rat retinal microglia isolation and culture

Retinas were obtained from the eyeballs of newborn (3 d) Sprague-Dawley rats as described in the previously published study by Jiang et al. (24). Briefly, retinas were washed in cold Hank’s solution and separated into 1–3-mm pieces. The mixture was transferred into sterile centrifuge tubes and centrifuged (1000 g, 3 min), and the supernatant was discarded. Trypsin (0.25%) was added to digest, and then the mixture was incubated at 37°C for 5 min. The digested tissues were mechanically dissociated into a single cell suspension and collected by centrifugation (1000 g, 3 min). The cells were then resuspended in DMEM-F12 containing 20% FBS and plated in T75 cell culture dishes at a density of 10⁶ cells per dish. After 10 d, when the cells were almost confluent (∼80–90%), the flasks were agitated at 37°C and 110 g for 1 h. The suspension was collected and centrifuged at 1000 g for 5 min. The supernatant was discarded, and the cells were then resuspended at the appropriate density depending on the experiment. Rat retinal microglial cells were cultured in DMEM-F12 containing 10% FBS and 1% penicillin-streptomycin and were identified using microglia-specific antibody OX42. The rat retinal microglial cells were preincubated with or without erianin (10, 50 nM) for 6 h and then incubated with or without mannitol (25 mM), l-glucose (25 mM), and d-glucose (25 mM) for additional 24 h.

Immunofluorescence staining assay

Paraffin-embedded sections of retinas (5 μm) were deparaffinized in xylene and rehydrated in an ethanol gradient with distilled water. Retinas were incubated with 5% bovine serum albumin to minimize nonspecific binding. After rinsing 3 times, retinas were incubated with Iba1 antibody (1:100 dilution) at 4°C overnight and further incubated with Alexa Fluor 488 goat anti-rabbit IgG (H+L) antibody at room temperature for 1 h. After rinsing 3 times again, retinas were further incubated with DAPI for 10 min. Images were captured under an inverted microscope (IX81; Olympus, Tokyo, Japan) (the statistical results were calculated by using cellSens Dimension; Olympus).

BV2 cells were preincubated with or without erianin (10, 50 nM) for 6 h and then incubated with or without d-glucose (25 mM) for an additional 24 h. After treatment, BV2 cells were fixed by 4% paraformaldehyde solution in PBS for 15 min and then incubated with 0.3% Triton X-100 for 10 min and further blocked with 1% bovine serum albumin for 1 h. Cells were probed with an appropriate combination of primary antibodies (1:100 dilution overnight at 4°C) and Alexa 488–labeled goat anti-rabbit antibody (1 h avoiding light at room temperature). DAPI was added into cells to stain the nuclei for 10 min. Fluorescence photographs were pictured using an Olympus IX81 inverted fluorescence microscope (the statistical results were calculated by using the program cellSens Dimension).

Real-time PCR assay

Cellular and retinal total RNA was isolated by using Trizol reagents according to the manufacturer’s instructions. The RNA content was determined by measuring the optical density at 260 nm, and cDNA was synthesized by using the PrimeScript RT Master Mix Kit. Real-time PCR was performed by using Sybr Premix Ex Taq Kit, and the relative expression of target genes was normalized to actin. The results were analyzed by the 2^−ΔΔCt method and given as a ratio compared with the control. The primer sequences used in this study are shown in Supplemental Table S1.

ELISA assay

The whole blood and vitreous cavity suspended in PBS were centrifuged at 850 g at 4°C for 15 min, and the serum and supernatant of the vitreous cavity were collected for further ELISA analysis as described in the kits.

BV2 cells were preincubated with or without erianin for 6 h and then incubated with or without d-glucose (25 mM) for 24 h. The supernatants from BV2 cells were collected by centrifugation at 850 g at 4°C for 15 min, and TNF-α content in the isolated cell supernatants was further detected by ELISA analysis as described in the kits.

Western blot analysis

Cells were pretreated with erainin (10, 50 nM) or STF31 (2.5 μM) for 6 h or with U0126 (20 μM) for 15 min and then incubated with d-glucose (25 mM) for the additional times. After treatment, nuclear and cytoplasmic proteins in cells were isolated by using nuclear and cytoplasmic extraction reagents as described in the kits. Cells were pretreated with erainin (10, 50 nM) for 6 h and then incubated with TNF-α (20 ng/ml) for an additional 18 h. After treatment, cellular proteins were extracted by using the whole-cell protein extraction kits for detecting the expression of occludin and claudin1. Cells were pretreated with erainin (10, 50 nM) or STF31 (2.5 μM) for 6 h or with U0126 (20 μM) for 15 min and then incubated with d-glucose (25 mM) for the additional times. After treatment, to detect the phosphorylated and total proteins, including p-cRaf, p-MEK1/2, ERK1/2, p-ERK1/2, etc., cell supernatants were quickly discarded, and cells were lysed and scraped by using 2-times protein loading buffer for Western blotting.

Nuclear and cytoplasmic proteins in retinas were isolated by using nuclear and cytoplasmic extraction reagents as described in the kits. Retinas were homogenized in ice-cold lysis buffer containing 50 mM Tris (pH 7.5), 150 mM NaCl, 1 mM EDTA, 20 mM NaF, 0.5% NP-40, 10% glycerol, 1 mM PMSF, 10 μg/ml aprotinin, 10 μg/ml leupeptin, and 10 μg/ml pepstatin A.

The protein concentration was determined by using kits and normalized to the equal amount of protein of each sample. Samples containing 20 μg proteins were separated by SDS-PAGE (12% SDS-PAGE gels for detecting the expression of claudin1 and Iba1 and 10% SDS-PAGE gels for detecting all the other proteins) and transferred onto PVDF membranes. Membranes were blocked with 5% bovine serum albumin at room temperature for 1 h and were incubated with primary antibodies (1:200–1:1000) at 4°C overnight and then incubated with secondary antibodies for 2 h at room temperature. The antibody-reactive bands were identified by ECL kits. The gray densities of the protein bands were normalized by using β-actin or Lamin B1 density as an internal control, respectively.

Molecular docking analysis

The crystal structure of glucose transporter (GLUT)-1 that is opening inward has been resolved (Protein Data Bank; http://www.rcsb.org/; code 4PYP). The 3-dimensional structure of GLUT1 that is opening outward is constructed by using the method of homologous modeling Molecular Operating Environment (MOE) program v.2016.08; Chemical Computing Group; Montreal, QC, Canada). The generated model is evaluated by the GBVI/WSA dG scoring function. The confirmation of erianin is generated by using Comformational Search (MMFF94X force field) in MOE v.2016.08. The Amber10 extended Huckel theory (EHT) force field was set in MOE first, classic triangle matching was chosen as the placement method, and the force field was employed as a refinement method for pose refinement. The best model was generated and evaluated by GBVI/WSA dG score, and the reasonable binding model is analyzed in MOE.

Glucose content detection

BV2 cells were preincubated with or without erianin (10, 50 nM) or STF31 (2.5 μM) for 6 h and then incubated with d-glucose (25 mM) for the additional 2.5 min. Cell supernatants were collected for further assay, and cellular glucose content was calculated according to the manufacturer’s instruction.

Measurement of transendothelial electrical resistance and transepithelial electrical resistance

HRECs or ARPE19 cells (2 × 10⁴) were plated into fibronectin-coated transwell inserts (0.4-μm pore size, Millicell Hanging Cell Culture Inserts; MilliporeSigma). The culture medium was replaced every other day. The cells were cultured for an additional 6–7 d until the transendothelial electrical resistance (TEER) and transepithelial electrical resistance (TER) stable values were obtained. BV2 cells were cultured under transwells in 24-well plates for 24 h and then preincubated with or without erianin (10, 50 nM) for 6 h. BV2 cells were further stimulated with or without d-glucose (25 mM) for an additional 6 or 24 h, and then TEER and TER were measured by using a Millicell ERS-2 Voltohmmeter (MilliporeSigma). Otherwise, the fresh medium containing TNF (20 ng/ml) and erianin (10, 50 nM) was added into the 24-well plates under transwells. TEER and TER were measured at 6 and 18 h thereafter. Blank wells were transwell inserts without endothelial or epithelial cells. The following formula was used: TEER/TER of monolayer (ΔΩcm²) = (sample well resistance – blank well resistance) × area of cell monolayer.

FITC-conjugated dextran cell permeability assay

HRECs or ARPE19 cells (2 × 10⁴) were plated into fibronectin-coated transwell inserts just like for detecting TEER and TER. The culture medium was replaced every other day. The cells were cultured for an additional 6–7 d. BV2 cells were cultured under transwells in 24-well plates for 24 h and then preincubated with or without erianin (10, 50 nM) for 6 h. BV2 cells were further stimulated with or without d-glucose (25 mM) for an additional 24 h. Otherwise, the fresh medium containing TNF (20 ng/ml) and erianin (10, 50 nM) was added into the 24-well plates under transwells and incubated for an additional 18 h. After the culture medium in inserts and bottom chambers was replaced with fresh medium, 70 kDa FITC-dextran (100 μg/ml) was added to transwell inserts and incubated for 4 h in a CO₂ incubator. Medium was collected from transwell inserts and bottom chambers, and the fluorescence intensity of FITC-dextran was measured at an excitation of 485 nm and an emission of 538 nm. The leakage of FITC-dextran was calculated using the following formula: permeability (%) = [FITC-dextran content in bottom chamber/(FITC-dextran content in transwell insert + bottom chamber)] × 100.

Statistical analysis

Data are expressed as means ± sem. The significance of differences between groups was evaluated by 1-way ANOVA with least significant difference post hoc test, and a value of P < 0.05 was considered as indicating statistically significant differences.

RESULTS

Erianin alleviated BRB disruption in STZ-induced diabetic mice

The results of blood glucose concentration detection during the whole experimental process and serum glucose concentration measurement at the end of the experiment both showed that erianin had no obvious effect on the elevated glucose content in mice with STZ-induced diabetes for 2 or 3 mo (Supplemental Fig. S1). Data in Fig. 1A showed that erianin (1, 10 mg/kg) significantly reduced the elevated Evan’s blue dye leakage in retinas of mice with STZ-induced diabetes for 2 mo. Data in Fig. 1B showed that the protein expression of occludin and claudin1 was decreased in mice with STZ-induced diabetes as compared with control normal mice. Erianin (1, 10 mg/kg) rescued the decreased expression of claudin1, and erianin (10 mg/kg) also rescued the decreased expression of occludin in mice with STZ-induced diabetes (Fig. 1B). As shown in Fig. 1C, in mice with STZ-induced diabetes for 3 mo, the leakage of Evan’s blue in retinas was remarkably elevated, but erianin (1, 10 mg/kg) reduced this elevation. Additionally, erianin (1, 10 mg/kg) also obviously rescued the decreased protein expression of occludin and claudin1 in retinas of mice with STZ-induced diabetes for 3 mo (Fig. 1D). The retinal expression of claudin1 and occludin was also increased in normal nondiabetic mice treated with erianin (10 mg/kg) (Supplemental Fig. S2A).

Figure 1.

Erianin alleviated BRB breakdown and rescued the decreased expression of occludin and claudin1 in mice with STZ-induced diabetes. At 1 mo after STZ injection, mice were injected with erianin (1, 10 mg/kg/d) consecutively for 1 mo. A) BRB breakdown is detected by Evan’s blue dye leakage assay (n = 6–7). B) Retinal protein expression of occludin and claudin1 is detected. The quantitative densitometric analysis of occludin and claudin1 is shown below (n = 3–4). At 2 mo after STZ injection, mice were injected with erianin (1, 10 mg/kg/d) consecutively for 1 mo. C) BRB breakdown is detected by Evan’s blue dye leakage assay (n = 5). D) Retinal protein expression of occludin and claudin1 is detected. The quantitative densitometric analysis of occludin and claudin1 is shown below (n = 4–5). Data represent means ± sem. *P < 0.05, **P < 0.01, ***P < 0.001 compared with control; ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 compared with DM.

Erianin inhibited the activation of microglial cells both in vivo and in vitro

Iba1 is the most reliable and widely used marker for reflecting the activation of microglia (25). Erianin (1, 10 mg/kg) had no effect on the expression of Iba1 in retinas from normal nondiabetic mice (Supplemental Fig. S2B). As shown in Fig. 2A, in mice with STZ-induced diabetes for 2 mo, the number of Iba1-positive microglial cells was remarkably increased in the inner plexiform layer and ganglion cell layer. Erianin (1, 10 mg/kg) reduced this increased number of activated microglial cells in retinas of mice with STZ-induced diabetes for 2 mo (Fig. 2A). The results of counting the number of Iba1-positive microglial cells and the fluorescence intensity of Iba1 were shown in Supplemental Fig. S3. Next, results showed that erianin (10 mg/kg) reduced the elevated Iba1 protein expression in retinas in mice with STZ-induced diabetes for 2 mo (Fig. 2B). Erianin also reduced the increased number of activated microglial cells in retinas of mice with STZ-induced diabetes for 3 mo (unpublished results).

Figure 2.

Erianin inhibited the activation of microglial cells both in vivo and in vitro. At 1 mo after STZ injection, mice were injected with erianin (1, 10 mg/kg/d) consecutively for 1 mo. A) Retinal expression of Iba1 in mice is detected. The representative pictures of retinal immunofluorescence staining of Iba1 and DAPI and also the merge of Iba1- and DAPI-stained images are shown at the left (scale bar, 20 μm). The enlarged representative pictures of retinal Iba1-stained images are shown at the right (scale bar, 5 μm). Red arrows indicate microglial cells. B) Erianin reduced the elevated protein expression of Iba1 in retinas. The quantitative densitometric analysis of Iba1 is shown below (n = 4). C) BV2 cells were preincubated with erianin for 6 h and then incubated with d-glucose (25 mM) for 24 h. BV2 cells were stained by using Iba1 antibody and pictured under an inverted fluorescence microscope (original magnification, ×200). Results represent 3 independent experiments. D) The quantitative analysis of immunofluorescence staining of Iba1 in BV2 cells. Data represent means ± sem. *P < 0.05, **P < 0.01 compared with control; ^#P < 0.05, ^##P < 0.01 compared with DM or d-glucose.

Microglial BV2 cells have already been commonly used in the studies about various ocular diseases, including DR, retinal degeneration, and ocular inflammation (26–29). Next, whether erianin also inhibits the activation of microglial cells induced by hyperglycemia in vitro was observed in BV2 cells stimulated by d-glucose. Data in Fig. 2C, D showed that erianin (50 nM) reduced the elevated expression of Iba1 in BV2 cells stimulated by d-glucose (25 mM) in vitro. However, erianin alone had no obvious cytotoxicity in BV2 cells (unpublished results).

Erianin inhibited the activation of the NF-κB signal pathway both in vivo and in vitro

Data in Fig. 3A, B showed that erianin (1, 10 mg/kg) reduced the elevated phosphorylation of IKK in retinas, and erianin (10 mg/kg) decreased the elevated phosphorylation of IκB in retinas in mice with diabetes induced by STZ for 2 mo. Erianin (1, 10 mg/kg) reduced the elevated nuclear accumulation of p65 and p-p65 in retinas in mice with diabetes induced by STZ for 2 mo, and erianin (1, 10 mg/kg) also reduced the elevated phosphorylation of cytosolic p65 in retinas (Fig. 3A, C, D). Results in Supplemental Fig. S4A showed that erianin (1, 10 mg/kg) alone had no effect on the phosphorylation of IκB, IKK, and p65 in retinas from normal nondiabetic mice. Data in Fig. 3E–G showed that erianin (1, 10 mg/kg) reduced the increased TNF-α mRNA expression in retinas and further decreased the elevated TNF-α content both in serum or vitreous body in mice with diabetes induced by STZ for 2 mo. Results in Supplemental Fig. S4B showed that erianin (1, 10 mg/kg) alone had no effect on the retinal mRNA expression of TNF-α in normal nondiabetic mice.

Figure 3.

Erianin inhibited the activation of the NF-κB signaling pathway both in vivo and in vitro. At 1 mo after STZ injection, mice were injected with erianin (1, 10 mg/kg/d) consecutively for 1 mo. A) Erianin (1, 10 mg/kg) reduced the increased phosphorylated IKK, IκB, p65, and the subsequent enhanced nuclear accumulation of p65 and p-p65 in retinas from mice with STZ-induced diabetes. B) The quantitative densitometric analysis of phosphorylated IKK and IκB (n = 3). C) The quantitative densitometric analysis of cytosolic and nuclear p65 (n = 3–4). D) The quantitative densitometric analysis of cytosolic and nuclear p-p65 (n = 4–5). E) Retinal mRNA expression of TNF-α (n = 5). F) Serum contents of TNF-α (n = 6). G) The contents of TNF-α in vitreous body (n = 4). H) BV2 cells were preincubated with erianin for 6 h and then incubated with d-glucose (25 mM) for an additional 6 h. Erianin reduced the increased nuclear accumulation of p65 induced by d-glucose (25 mM) in BV2 cells. The quantitative densitometric analysis of p65 is shown down (n = 4). I) BV2 cells were preincubated with erianin for 6 h and then incubated with d-glucose (25 mM) for an additional 3 or 6 h. Erianin reduced the increased expression of phosphorylated IKK, IκB, and p65 induced by d-glucose in BV2 cells. The quantitative densitometric analysis of phosphorylated IKK, IκB, and p65 is shown down (n = 3–5). J) BV2 cells were preincubated with erianin for 6 h and then incubated with d-glucose (25 mM) for 24 h. TNF-α content in cell supernatant is detected (n = 3–4). Data represent means ± sem. *P < 0.05, **P < 0.01, ***P < 0.001 compared with control; ^#P < 0.05, ^##P < 0.01, ^###P < 0.001 compared with DM or d-glucose or d-glucose (3 h) in Fig. 3I; ^$P < 0.05, ^$$P < 0.01, ^$$$P < 0.001 compared with d-glucose (6 h) in Fig. 3I.

Next, results showed that 25 mM d-glucose induced the nuclear accumulation of p65 in microglial BV2 cells (Fig. 3H), but 25 mM mannitol and 25 mM l-glucose both did not have this effect (Fig. 3H and Supplemental Fig. S5A, B). Erianin (10, 50 nM) decreased the elevated nuclear accumulation of p65 in microglial BV2 cells stimulated by 25 mM d-glucose (Fig. 3H). d-glucose (25 mM) increased the phosphorylation of p65 in microglial BV2 cells (Fig. 3I), but 25 mM mannitol and 25 mM l-glucose both did not have this effect (Fig. 3I and Supplemental Fig. S5A, C). Erianin (10, 50 nM) reduced the enhanced phosphorylation of p65, IκB, and IKK in BV2 cells when cells were incubated with 25 mM d-glucose for both 3 and 6 h (Fig. 3I). Additionally, erianin (10, 50 nM) also decreased the elevated TNF-α content in cell supernatants from d-glucose–stimulated BV2 cells (Fig. 3J). However, erianin (10, 50 nM) alone had no obvious effect on the phosphorylation of p65, IκB, and IKK in BV2 cells (Supplemental Fig. S5D).

Next, we isolated the retinal microglial cells from newborn Sprague-Dawley rats. The isolated microglial cells were identified by immunofluorescence staining with microglia-specific antibody OX42 (Supplemental Fig. S6A). The further results showed that erianin (10, 50 nM) also inhibited the elevated phosphorylation of p65, IκB, and IKK and the nuclear accumulation of p65 and p-p65 in d-glucose–treated primary rat retinal microglial cells (Supplemental Fig. S6B–E).

Erianin decreased ERK1/2 phosphorylation both in vivo and in vitro

As shown in Fig. 4A, B, d-glucose (25 mM) enhanced the phosphorylation of ERK1/2 in BV2 cells, but erianin (50 nM) and U0126 (20 μM) both reduced this increased phosphorylation. Erianin (10 nM) decreased the enhanced phosphorylation of cRaf and MEK1/2, which are both upstream signals of ERK1/2, induced by d-glucose (25 mM) in BV2 cells (Fig. 4C, D). Erianin (50 nM) also reduced the increased phosphorylation of cRaf in BV2 cells treated with d-glucose (25 mM) (Fig. 4C, D). The osmotic control d-mannitol (25 mM) and the metabolic control l-glucose (25 mM) both had no effect on the phosphorylation of ERK1/2, MEK1/2, and cRaf in BV2 cells (Fig. 4A–D and Supplemental Fig. S7A). Additionally, erianin (10 mg/kg) reduced the elevated phosphorylation of cRaf and ERK1/2 in retinas of mice with diabetes induced by STZ for 2 mo (Fig. 4E, F). Erianin (1 mg/kg) also decreased the elevated phosphorylation of ERK1/2 in retinas from mice with STZ-induced diabetes (Fig. 4E, F). Erianin (10, 50 nM) alone had no effect on the phosphorylation of ERK1/2, cRaf, and MEK1/2 in BV2 cells in vitro, and erianin (1, 10 mg/kg) alone also had no obvious effect on the phosphorylation of ERK1/2 in normal nondiabetic mice (Supplemental Fig. S7B, C).

Figure 4.

Erianin decreased ERK1/2 phosphorylation both in vivo and in vitro. A) BV2 cells were preincubated with erianin or U0126 for 6 h or 15 min and then incubated with d-glucose (25 mM) for the additional 2.5 min. Erianin and U0126 reduced the increased phosphorylation of ERK1/2 induced by d-glucose in BV2 cells. B) The quantitative densitometric analysis of phosphorylated ERK1/2 (n = 5). C) BV2 cells were preincubated with erianin for 6 h and then incubated with d-glucose (25 mM) for the additional 2.5 min. Erianin decreased the enhanced phosphorylation of cRaf and MEK1/2 induced by d-glucose in BV2 cells. D) The quantitative densitometric analysis of phosphorylated MEK1/2 and cRaf (n = 3–4). E) Erianin reduced the elevated phosphorylation of cRaf and ERK1/2 in retinas from mice with STZ-induced diabetes. F) The quantitative densitometric analysis of phosphorylated cRaf and ERK1/2 in retinas (n = 3–4). G) BV2 cells were preincubated with U0126 for 15 min and then incubated with d-glucose (25 mM) for the additional 6 h. U0126 reduced the increased nuclear accumulation of p65, p-p65, and the increased expression of cytosolic p-p65 in BV2 cells stimulated by d-glucose (25 mM). The quantitative densitometric analysis is shown at right (n = 4–6) (a, b). H) BV2 cells were preincubated with U0126 for 15 min and then incubated with d-glucose (25 mM) for the additional 6 h. U0126 reduced the increased expression of phosphorylated IKK and IκB induced by d-glucose in BV2 cells. The quantitative densitometric analysis of phosphorylated IKK and IκB is shown at right (n = 5–6). Data represent means ± sem. *P < 0.05, **P < 0.01 compared with control; ^#P < 0.05, ^##P < 0.01 compared with DM or d-glucose.

Next, results showed that U0126, an ERK1/2 inhibitor, reduced the nuclear accumulation of p65 and p-p65 and decreased the elevated phosphorylation of cytosolic p65 in BV2 cells treated with 25 mM d-glucose (Fig. 4G). Moreover, U0126 also reduced the increased phosphorylation of IκB and IKK in BV2 cells treated with 25 mM d-glucose (Fig. 4H).

Erianin reduced GLUT1-mediated cellular d-glucose uptake in BV2 cells

The 3-dimensional (Fig. 5A) and 2-dimensional (Fig. 5B) interaction map of the molecular docking assay showed that erianin formed 2 H-bond interactions between the hydroxyl group with Glu380 and Asn288 at the endofacial sugar binding site of GLUT1, and the binding energy of erianin is −6.965 kcal/mol. However, as shown in Fig. 5C, there is no interaction between erianin and the exofacial sugar binding site of GLUT1. To observe whether GLUT1-mediated cellular glucose uptake played an important role in regulating the activation of ERK1/2 and NF-κB signaling pathways, the GLUT1 inhibitor STF31 was used in the next experiment. Data in Fig. 5D showed that erianin (10, 50 nM) and STF31 (2.5 μM) both decreased the uptake of glucose into BV2 cells when cells were treated with d-glucose (25 mM). Further results showed that STF31 (2.5 μM) decreased the enhanced phosphorylation of cRaf, MEK1/2, and ERK1/2 induced by 25 mM d-glucose in BV2 cells (Fig. 5E). Meanwhile, STF31 (2.5 μM) also decreased the increased nuclear accumulation of p65 and p-p65 in BV2 cells induced by 25 mM d-glucose (Fig. 5F).

Figure 5.

Erianin reduced GLUT1-mediated d-glucose uptake in BV2 cells. A) Front view of the docking mode of erianin (green) in the binding site of the inward-facing structure of GLUT1 (shown in ribbon representation and colored by structure). B) Two-dimensional interaction map of erianin and inward-facing structure of GLUT1. C) Front view of the docking mode of erianin (green) in the binding site of the outward-facing structure of GLUT1 (shown in ribbon representation and colored by structure). D) BV2 cells were preincubated with erianin (10, 50 nM) or STF31 (2.5 μM) for 6 h and then incubated with d-glucose (25 mM) for the additional 2.5 min. Erianin and STF31 decreased the uptake of d-glucose into BV2 cells (n = 6). E) BV2 cells were preincubated with STF31 (2.5 μM) for 6 h and then incubated with d-glucose (25 mM) for the additional 2.5 min. STF31 decreased the increased phosphorylation of cRaf, MEK1/2, and ERK1/2 induced by d-glucose in BV2 cells. The quantitative densitometric analysis of phosphorylated cRaf, MEK1/2, and ERK1/2 is shown below (n = 5). F) BV2 cells were preincubated with STF31 (2.5 μM) for 6 h and then incubated with d-glucose (25 mM) for the additional 6 h. STF31 reduced the increased nuclear accumulation of p65 and p-p65 induced by d-glucose (25 mM) in BV2 cells. The quantitative densitometric analysis is shown below (n = 3–5). G) BV2 cells were preincubated with U0126 and QNZ for 15 min or 3 h and then incubated with d-glucose (25 mM) for the additional 12 h. Cellular mRNA expression of GLUT1 is detected (n = 4). t-ERK1/2, total ERK1/2. Data represent means ± sem. *P < 0.05, **P < 0.01, ***P < 0.001 compared with control; ^#P < 0.05, ^##P < 0.01 compared with d-glucose.

Next, we observed whether GLUT1 expression was in turn affected by ERK1/2 and NF-κB signaling pathways. Data in Fig. 5G showed that U0126 (20 μM) and QNZ (100 nM), an NF-κB inhibitor, both decreased the increased mRNA expression of GLUT1 in BV2 cells induced by 25 mM d-glucose.

Erianin weakened BRB damage in vitro triggered by d-glucose–treated BV2 cells

Data in Fig. 6A showed that HREC monolayers cocultured with d-glucose–stimulated BV2 cells for 6 or 24 h had an obviously lower TEER value, indicating the serious damage on the barrier structure. However, erianin (10, 50 nM) attenuated this decrease of TEER value (Fig. 6A). When HRECs were incubated with d-glucose alone or BV2 cells stimulated by mannitol (25 mM) or l-glucose (25 mM) for 6 or 24 h, TEER value was not altered (Fig. 6A and Supplemental Fig. S8A). Next, as compared with the control group, the leakage of FITC-dextran through the monolayer of HRECs that were cocultured with d-glucose–stimulated BV2 cells for 24 h was significantly increased (Fig. 6B). However, erianin (10, 50 nM) reduced this increased FITC-dextran leakage (Fig. 6B). Also, the leakage of FITC-dextran through the monolayer of HRECs that were cocultured with l-glucose–stimulated BV2 cells or mannitol-stimulated BV2 cells for 24 h had no alteration (Fig. 6B and Supplemental Fig. S8B).

Figure 6.

Erianin weakened BRB damage in vitro triggered by d-glucose–treated BV2 cells. A) TEER value is detected when HRECs were incubated with d-glucose–stimulated BV2 cells for 6 or 24 h, respectively (n = 6). B) FITC-dextran leakage assay is conducted when HRECs were incubated with d-glucose–stimulated BV2 cells for 24 h (n = 4). C) TER value is detected when ARPE19 cells were incubated with d-glucose–stimulated BV2 cells for 6 or 24 h, respectively (n = 6). D) FITC-dextran leakage assay is conducted when ARPE19 cells were incubated with d-glucose–stimulated BV2 cells for 24 h (n = 4). Data represent means ± sem. *P < 0.05, ***P < 0.001 compared with HRECs or ARPE19 cells alone; ^#P < 0.05, ^###P < 0.001 compared with HRECs or ARPE19 cells incubated with d-glucose–stimulated BV2 cells.

Data in Fig. 6C showed that ARPE19 cell monolayers cocultured with d-glucose–stimulated BV2 cells for 6 or 24 h had an obviously lower TER value. However, erianin (10, 50 nM) attenuated this decrease of TER value (Fig. 6C). When ARPE19 cells were incubated with d-glucose alone or BV2 cells stimulated by mannitol (25 mM) or l-glucose (25 mM) for 6 h or 24 h, TER value was not altered (Fig. 6C and Supplemental Fig. S8C). Further results showed that ARPE19 cells cocultured with d-glucose–stimulated BV2 cells for 24 h obviously increased the leakage of FITC-dextran through the confluent monolayer of ARPE19 cells (Fig. 6D). However, erianin (10, 50 nM) also reduced this increased FITC-dextran leakage (Fig. 6D). Also, the leakage of FITC-dextran through the monolayer of ARPE19 cells that were cocultured with l-glucose–stimulated BV2 cells or mannitol-stimulated BV2 cells for 24 h had no alteration (Fig. 6D and Supplemental Fig. S8D).

Erianin attenuated TNF-α–induced BRB damage in vitro

Data in Fig. 7A showed that HREC monolayers incubated with TNF-α (20 ng/ml) for 6 or 18 h had an obviously lower TEER value, but erianin (10, 50 nM) attenuated this decrease of TEER value. TNF-α (20 ng/ml) increased the leakage of FITC-dextran through the confluent monolayer of HRECs when cells were incubated with TNF-α for 18 h, and erianin (10, 50 nM) reduced this increase (Fig. 7B). Cellular expression of occludin and claudin1 was decreased in HRECs when cells were treated with TNF-α (20 ng/ml) for 18 h, but erianin (10, 50 nM) rescued this reduction (Fig. 7C).

Figure 7.

Erianin attenuated TNF-α–induced BRB damage in vitro. A) TEER value is detected when HRECs were incubated with TNF-α (20 ng/ml) for 6 or 18 h, respectively (n = 5). B) FITC-dextran leakage assay is conducted when HRECs were incubated with TNF-α (20 ng/ml) for 18 h (n = 3). C) Erianin reversed the decreased expression of occludin and claudin1 in TNF-α–treated HRECs. The quantitative densitometric analysis of occludin and claudin1 is shown below (n = 4–5). D) TER value is detected when ARPE19 cells were incubated with TNF-α (20 ng/ml) for 6 or 18 h, respectively (n = 5). E) FITC-dextran leakage assay is conducted when ARPE19 cells were incubated with TNF-α (20 ng/ml) for 18 h (n = 4). F) Erianin reversed the decreased expression of occludin and claudin1 in TNF-α–treated ARPE19 cells. The quantitative densitometric analysis of occludin and claudin1 is shown below (n = 5–7). Data represent means ± sem. *P < 0.05, **P < 0.01, ^$P < 0.05, ^$$$P < 0.001 compared with control; ^#P < 0.05, ^##P < 0.01, ^&&P < 0.01, ^&&&P < 0.001 compared with TNF-α.

Data in Fig. 7D showed that ARPE19 cells incubated with TNF-α (20 ng/ml) for 6 or 18 h had an obviously lower TER value, but erianin (10, 50 nM) attenuated this decrease of TER value. TNF-α (20 ng/ml) increased the leakage of FITC-dextran through the confluent monolayer of ARPE19 cells when cells were incubated with TNF-α for 18 h, and erianin (10, 50 nM) reduced this increase (Fig. 7E). Cellular expression of occludin and claudin1 was decreased in ARPE19 cells when cells were treated with TNF-α (20 ng/ml) for 18 h, but erianin (10, 50 nM) rescued this reduction (Fig. 7F).

The results in Supplemental Fig. S9 showed that erianin (10, 50 nM), mannitol (25 mM), and d-glucose (25 mM) all had no obvious effect on the expression of cellular occludin and claudin1 in both HRECs and ARPE19 cells.

DISCUSSION

The BRB is in charge of the homeostatic regulation of the microenvironment in the retina. BRB controls fluid and molecular movement between the retinal tissues and the ocular vascular beds, and it also prevents the leakage of macromolecules or other potentially harmful agents into the retina (30, 31). BRB is composed of an inner BRB (iBRB) and an outer BRB (oBRB). The iBRB is formed by capillary endothelial cells and the tight junctions (TJs) between neighboring cells (32). Meanwhile, astrocytes, müller cells, and pericytes are also thought to contribute to maintain the proper function of the iBRB (33). The oBRB is formed by TJs between cells of the retinal pigment epithelium and plays an important role in transporting nutrients from the blood to the outer retina (34). Both iBRB and oBRB TJs mainly include occludin, the claudin family, zonula occludens proteins (ZO-1,2,3), and junction adhesion molecules (35, 36). Nearly all the retinal diseases, including DR, retinopathy of prematurity, and age-related macular degeneration, were accompanied with the disruption of BRB integrity (37, 38). Moreover, it has already been reported that the hallmark of NPDR is the BRB breakdown (9, 10). In this study, erianin was found to rescue the decreased expression of TJs, including occludin and claudin1, and attenuate the BRB leakage in retinas from mice with STZ-induced diabetes. These results imply that erianin had the protective activity toward BRB disruption during the progression of NPDR. Moreover, erianin (10 mg/kg) alone was also found to increase retinal protein expression of occludin and claudin1 in nondiabetic normal control mice, and the relevant mechanism needs further study.

Hyperglycemia is a typical feature of DM, and it has been reported that hyperglycemia caused oxidative stress and inflammatory injury during DR development and eventually induced vascular dysfunction (9). Microglial cells are the resident immune cells in retinas. A previous study has shown that microglial cells contributed to the survival of photoreceptors in retinal detachment because of the regulation of macrophage infiltration and photoreceptor phagocytosis (39). Hyperglycemia has already been reported to induce the activation of microglial cells during the progression of DR (23). Moreover, microglial cells have been reported to be in charge of regulating the inflammatory responses during DR development (40). Thus, it can be seen that regulating microglial reactivity has been suggested to be a promising strategy for treating various retinal diseases (41). In this study, we found that erianin reduced the activation of microglial cells both in vivo and in vitro. Transcription factor NF-κB is responsible for the expression of a variety of proinflammatory cytokines and plays important roles in triggering microglia-mediated inflammation in various neuropathological diseases (42). Next results showed that erianin inhibited the transcriptional activation of NF-κB and further reduced the expression of proinflammatory cytokine TNF-α both in vivo and in vitro. These above results indicate that the anti-inflammatory activity of erianin during NPDR development is due to its inhibition on NF-κB activation in microglial cells. Moreover, erianin alone had no obvious effect on the activation of the NF-κB signaling pathway in retinas from normal nondiabetic mice and in BV2 cells treated with d-glucose.

MAPKs are a group of serine or threonine protein kinases that are highly conserved across eukaryotic species. In multicellular organisms, MAPKs are necessary for development, differentiation, learning, memory, and secretion of paracrine and autocrine factors (43, 44). There are 3 well-defined MAPK pathways: the ERK1/2 pathway, the JNK pathway, and the p38 kinase pathway. It is reported that ERK1/2 was engaged in many CNS diseases through regulating neuroinflammation (45). Our previous study showed that hyperglycemia induced the activation of the ERK1/2–HIF-1α–VEGF pathway in microglial cells, which contributed to retinal neoangiogenesis during PDR development (23). In this study, we found that erianin reduced the phosphorylated activation of ERK1/2, MEK1/2, and cRaf both in vivo and in vitro. Moreover, erianin alone had no obvious effect on the activation of the ERK1/2 signaling pathway in retinas from normal nondiabetic mice and in BV2 cells treated with d-glucose. Meanwhile, the ERK1/2 inhibitor U0126 abrogated the transcriptional activation of NF-κB induced by d-glucose in BV2 cells. These results indicate that the anti-inflammatory activity of erianin during NPDR development is acquired by abrogating the ERK1/2–NF-κB signaling pathway.

Members of the GLUT family, belonging to the sugar porter subfamily of the major facilitator superfamily, serve as monosaccharide transporters in mammals and are in charge of the uptake or release of sugars from cells and the circulation system (46). Among 14 members in this family, GLUT1 and GLUT4 are essential for maintaining glucose homeostasis (47). It has been reported that the activation of some immune cells was correlated with the enhanced glucose uptake mediated by GLUT1 (48, 49). In this study, we found that erianin reduced cellular glucose uptake in BV2 cells when cells were treated with d-glucose (25 mM). As the major facilitator superfamily transporter, GLUT1 transports substrate by using the alternating access mechanism. A working model for GLUT1 has been previously reported by Deng et al. (50), the predicted conformations including outward-open, ligand-bound and occluded, inward-open, and ligand-free and occluded, which were all required for a complete transport cycle. Next, the results of molecular docking analysis imply the potential interaction between erianin with the inward-open structure of GLUT1, which may impede the complete transport cycle and thus lead to the reduced cellular uptake of glucose. STF31 is a widely applied inhibitor for GLUT1 (51, 52). In this study, STF31 reduced the d-glucose–induced phosphorylation of ERK1/2 and the subsequent transcriptional activation of NF-κB in BV2 cells. These results indicate that the erianin-provided inhibition on glucose uptake contributed to its abrogation on the ERK1/2–NF-κB signal pathway. Additionally, our results showed that U0126 and QNZ, a well-known NF-κB inhibitor, both reduced the increased GLUT1 mRNA expression induced by 25 mM d-glucose. These results imply that the ERK1/2–NF-κB signal pathway also plays a positive feedback role in regulating the d-glucose–mediated GLUT1 expression in microglial cells.

Further results showed that the integrity of iBRB and oBRB was disrupted when HRECs and ARPE19 cells were cocultured with d-glucose–stimulated microglial cells, and this alteration is irrelevant with d-glucose itself. It has been proved that the proinflammatory cytokine TNF-α altered the leukocyte adherence, damaged the TJ complex, and finally induced BRB breakdown (53). Thus, we guess that the release of TNF-α may be responsible for the BRB damage mediated by d-glucose–stimulated microglial cells. Next, results proved our guess that TNF-α itself reduced the expression of occludin and claudin1 and increased the permeability of both iBRB and oBRB in vitro. However, erianin improved all those alternations induced by TNF-α in both HRECs and ARPE19 cells. Moreover, erianin alone had no obvious effect on the expression of occludin and claudin1 in both HRECs and ARPE19 cells. These results indicate that erianin maintained BRB integrity not only by inhibiting d-glucose–induced inflammatory response in microglia but also by directly attenuating proinflammatory cytokine TNF-α–mediated BRB damage.

CONCLUSIONS

This study demonstrates that the natural product erianin alleviates BRB disruption initiated by microglia during the progression of DR. Erianin inhibited GLUT1-mediated glucose uptake into microglial cells and thus abrogated the activation of the ERK1/2–NF-κB inflammatory signaling pathway and reduced the production of proinflammatory cytokine TNF-α. Moreover, erianin also directly attenuated TNF-α–mediated BRB damage by rescuing the reduction of TJs.

Supplementary Material

This article includes supplemental data. Please visit http://www.fasebj.org to obtain this information.

Glossary

BRB

blood-retinal barrier

DM

diabetes mellitus

DR

diabetic retinopathy

FBS

fetal bovine serum

GLUT

glucose transporter

HREC

human retinal endothelial cell

Iba1

ionized calcium-binding adapter molecule 1

iBRB

inner BRB

NPDR

nonproliferative DR

oBRB

outer BRB

OX42

cluster of differentiation 11b/c

PDR

proliferative DR

STZ

streptozotocin

TEER

transendothelial electrical resistance

TER

transepithelial electrical resistance

TJ

tight junction

AUTHOR CONTRIBUTIONS

L. Ji conceived the study; T. Zhang, H. Ouyang, and X. Mei designed and performed the experiments and analyzed the results; B. Lu helped carry out the animal experiments; T. Zhang and L. Ji drafted the manuscript; K. Chen and Z. Wang contributed to the manuscript revision; and all authors read and approved the final manuscript.