Suppression of Retinal Peroxisome Proliferator-Activated Receptor γ in Experimental Diabetes and Oxygen-Induced Retinopathy: Role of NADPH Oxidase

Abstract

purpose. Recently, the authors have shown that NADPH oxidase is positively correlated with increased leukocyte adhesion and vascular leakage in diabetes and neovascularization in oxygen-induced retinopathy (OIR). Peroxisome proliferator-activated receptor gamma (PPARγ) agonists have been shown to prevent vascular inflammation and leakage in an experimental model of diabetes. The goal of this study was to investigate whether there is a link between NADPH oxidase and PPARγ that leads to vascular dysfunction in diabetic retina or OIR.

methods. Diabetes was induced with streptozotocin in wild-type mice or NOX2 knockout mice. One group of wild-type mice was treated with apocynin. Bovine retinal endothelial cells (BRECs) were treated with normal glucose (5 mM) or high glucose (25 mM) in the presence or absence of superoxide dismutase (SOD) or NADPH oxidase inhibitors (apocynin or diphenyleneiodonium [DPI]). Western blotting and immunofluorescence were used to evaluate PPARγ expression. Activation of nuclear factor (NF)κB was measured using the transcription factor assay kit and Western blot analysis of phospho-NFκB. PPARγ expression was also tested in OIR and lipopolysaccharide-induced retinal inflammation.

results. Retinal expression of PPARγ was suppressed in experimental models of diabetes, OIR, and retinal inflammation. This was associated with the activation of NFκB in the diabetic retina. These effects were prevented by apocynin or deletion of NOX2. PPARγ expression was also suppressed in endothelial cells treated with high glucose, and this was prevented by apocynin, DPI, and SOD.

conclusions. Suppression of PPARγ is involved in the pathogenesis of diabetic retinopathy and OIR. NADPH oxidase could be an upstream mediator of these changes.

Diabetic retinopathy is the most common cause of blindness in working adults. It is characterized by early retinal microvascular dysfunctions such as abnormal vascular flow, hyperpermeability, and the nonperfusion of capillaries. ¹ Hyperglycemia is a major risk factor that has been linked to the development of vascular dysfunction in diabetic retinopathy. Increased production of reactive oxygen species (ROS) ^{2 3} and inflammatory markers ^{4 5 6} have all been shown to be associated with high glucose treatment of endothelial cells. The retina’s large oxygen uptake and glucose oxidation make it more susceptible than any other tissue to oxidative stress. ^{6 7}  

Studies have shown that oxidative stress is implicated in the development of diabetic neuropathy, ⁸ nephropathy, ⁹ and retinopathy. ^{10 11} The major sources of ROS are NADPH oxidase(s), cytochrome P450, and nitric oxide synthase. In particular, studies have linked NADPH oxidase to vascular complications of diabetes such as atherosclerosis, ¹² hypertension, ^{13 14} nephropathy, ¹⁵ and retinopathy. ^{16 17} NADPH oxidase consists of two membranous subunits, NOX2 and p22^phox; three cytosolic subunits, p40^phox, p47^phox, and p67^phox; and the small GTP-binding protein rac-1. ^{1 18} Recent research has shown that the inhibition of NADPH oxidase prevents retinal neovascularization in oxygen-induced retinopathy (OIR) ¹⁹ and vascular leakage in the diabetic retina. ^{20 21} The mechanism by which NADPH oxidase mediates vascular damage in diabetic or OIR is still under investigation, but evidence indicates that this may occur through VEGF expression ^{19 22 23} or increased vascular inflammation. ^{16 20 21 24} In particular, leukocyte-endothelial interaction (leukostasis) is thought to be an early and key event in the pathogenesis of diabetic retinopathy ^{25 26} and OIR. ^{27 28}  

Peroxisome proliferator-activated receptor gamma (PPARγ) is a member of a ligand-activated nuclear receptor superfamily and plays a critical role in a variety of biological processes, including adipogenesis, glucose metabolism, and angiogenesis. ²⁹ PPARγ may also represent a target for cardiovascular risk reduction. Synthetic PPARγ agonists such as pioglitazone and rosiglitazone increase insulin sensitivity, modify lipid profiles, decrease blood pressure, and reduce biomarkers of inflammation. ^{30 31 32} Previous work has shown that the PPARγ signaling pathway inhibits diabetes-induced vascular injury in retina ³³ and kidney ³⁴ through a mechanism involving the inhibition of leukocyte adhesion. The anti-inflammatory effect of PPARγ has been shown to be mediated through the inhibition of the transcription factor nuclear factor (NF)κΒ, which plays a crucial role in inflammation. ^{35 36 37}  

The goals of the present study were to characterize the changes in retinal expression of PPARγ in diabetic and OIRs and to determine whether NADPH oxidase plays a specific role in these changes. 

Methods

Animals

All experimental procedures were performed according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Experiments were performed on female C57Bl/6J mice 6 to 8 weeks old and age-matched NOX2-deficient mice backcrossed on a C57Bl/6 background (Jackson Laboratories, Bar Harbor, ME), each weighing 21 to 25 g. Six to eight mice were used for each experimental group (control wild-type [WT], control NOX2 knockout, diabetic WT, diabetic WT treated with apocynin, and diabetic NOX2 knockout). NOX2 knockout mice were genotyped before the experiment. 

Cell Culture

Primary cultures of bovine retinal endothelial cells (BRECs) passages 7 to 9 were used in these experiments as described. ³⁸ Parallel experiments were also conducted on primary cultured human umbilical vein endothelial cells (HUVECs) obtained from Clonetics (Walkersville, MD) and maintained in complete endothelial cell growth medium (EGM)-2 supplemented with single-use aliquots (EGM-2 SingleQuots; (Clonetics), which contained human fibroblast growth factor-basic (hFGF-B), VEGF, human recombinant epidermal growth factor (hEGF), human recombinant insulin-like growth factor (R³IGF)-1, ascorbic acid, hydrocortisone, heparin, 2% fetal bovine serum, gentamicin, and amphotericin. 

Normal and High-Glucose Treatment of Cell Culture

BRECs or HUVECs were grown until they were 80% to 90% confluent and were switched to serum-free medium overnight. The next day the cells were treated with 5 mM d-glucose (NG) or 25 mM d-glucose (HG) in the presence or absence of NADPH oxidase inhibitor, apocynin (100 μM; Sigma, St. Louis, MO), and 5 μM diphenyleneiodonium (DPI; Sigma), which is also a general inhibitor of flavoprotein-containing enzyme, a group that includes NADPH oxidase and several other enzymes such as nitric oxide synthase (NOS) ^{39 40} or 100 U cell-permeable superoxide dismutase polyethylene glycol (Sigma). Additional groups of cells were incubated in 5 mM d-glucose supplemented with 20 mM mannitol as an osmolarity control. Three days later, cells were harvested and processed for analysis of PPARγ expression. This experiment was replicated with at least two different batches of endothelial cells. 

Mouse Model of Diabetic Retinopathy

Wild-type (WT) and NOX2 knockout mice were made diabetic by multiple intraperitoneal injections of streptozotocin (STZ; 55 mg/kg; Sigma) dissolved in 0.1 M fresh citrate buffer (pH 4.5). The mean blood glucose level was 437 ± 53 in WT and 441 ± 61 in knockout mice. One group of diabetic WT mice received apocynin (10 mg/kg) in drinking water. After 5 weeks, diabetic and age-matched normal and knockout mice were processed for Western blot analysis and immunolocalization. One retina from each animal was immediately frozen in liquid nitrogen and stored at −80°C until further Western blot analysis, and the other eyeball was embedded in OCT for sectioning. 

Mouse Model of Acute Retinal Vascular Inflammation

Additional groups of WT and knockout mice were studied after injection with lipopolysaccharide (LPS) from Salmonella typhimurium (0.1 mg/kg; Sigma) as a model of acute retinal vascular inflammation. Retinas were collected and processed for analysis of PPARγ expression by Western blot 24 hours later. 

Mouse Model of Oxygen-Induced Retinopathy

Experimental retinal neovascularization has been developed by incubating a group of mice at postnatal day (P) 7 in high oxygen (75%) for 5 days, followed by 5 days in room air (normoxia). One group of mice was treated with apocynin (intraperitoneal, 10 mg/kg/d) from P12 to P16. Mice were then killed on P17, and PPARγ expression was tested in retina using immunofluorescence and Western blotting. Additional groups of oxygen-treated mice were tested at P14 and compared with age-matched controls. 

Western Blot Analysis

For analysis of PPARγ and phospho (p)-NFκB, one retina from each mouse in different groups and treated endothelial cells were homogenized in a modified RIPA buffer (20 mM Tris-HCl [pH 7.4], 2.5 mM ethylenediaminetetraacetic acid, 50 mM NaF, 10 mM Na₄P₂O₇, 1% Triton X-100, 0.1% sodium dodecyl sulfate, 1% sodium deoxycholate, 1 mM phenylmethylsulfonyl fluoride). Homogenates (50 μg protein) were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis using ready precast gel (Bio-Rad, Hercules, CA), transferred to nitrocellulose membrane, and reacted with anti-PPARγ (1:200) and anti–p-NFκB (p65) 1:100 (Santa Cruz Biotechnology, Santa Cruz, CA) followed by horseradish peroxidase-linked secondary antibody and enhanced chemiluminescence (Amersham Pharmacia, San Francisco, CA). Membranes were stripped and reprobed for β-actin or NFκB to demonstrate equal loading, and results were analyzed with the use of densitometry. 

PPARγ Immunolocalization

Retinal frozen sections (12-μm thick) from diabetic and age-matched control mice were prepared for PPARγ immunolocalization. Retinal sections were fixed with 4% paraformaldehyde for 5 minutes, followed by washing with PBS and blocking with 3% normal goat serum for 30 minutes. Sections were incubated with the endothelial cell-specific marker biotinylated Griffonia simplicifolia isolectin B4 (GSI; Vector Laboratories, Burlingame, CA) and anti-PPARγ polyclonal antibody (Santa Cruz Biotechnology) 1:50 overnight at 4°C, followed by avidin-conjugated Texas red (Vector Laboratories) and Oregon green-labeled anti–rabbit antibody (Molecular Probes, Eugene, OR) to identify the expression and localization of PPARγ in retinal sections using confocal microscopy (LSM 510; Carl Zeiss, Thornwood, NY). Specificity of the reaction was confirmed by omitting the primary antibody and isolectin B4. For densitometry analysis, we collected two representative images from each retinal section (five sections per mouse) from six different mice in each group. Collected images were analyzed by computer-assisted morphometry for fluorescence intensity. 

Measurement of Retinal NFκB Activity

Frozen retina was homogenized in complete lysis buffer for the preparation of whole cell extract using a nuclear extract kit (Active Motif, Carlsbad, CA). Homogenate was centrifuged at 6160g for 10 minutes, and the supernatant was portioned into aliquots and stored at −80°C. Protein concentration was determined, and the 20 μg whole-cell extract was used for the determination of NFκB activity using the NFκB p65 transcription factor assay kit (TransAM; Active Motif) as described. ⁴¹ Each of the standards and samples was run in duplicate, and the amount of activated NFκB was normalized per microgram of retinal protein. 

Statistical Analysis

Group differences were evaluated using ANOVA followed by Tukey post hoc test. Results were considered significant when P < 0.05. For in vivo studies, age-matched control was compared with diabetic WT mice treated with or without apocynin and with diabetic or nondiabetic mice lacking NOX2 (n = 6–8). For in vitro studies, four dishes were prepared for each treatment group, and each experiment was replicated with at least two different batches of endothelial cells. Data are represented as mean ± SE from at least six animals in each group and three experiments from the in vitro study. 

Results

Effect of Apocynin or Deletion of NOX2 on Retinal Expression of PPARγ in Diabetic Mice

Recently, we have shown that the inhibition of NADPH oxidase by apocynin or the deletion of NOX2 leads to a significant reduction in ROS formation in the diabetic retina. ²¹ To determine whether there is a relationship between NADPH oxidase activation and PPARγ expression in diabetic retina, we tested the effect of apocynin treatment or deletion of NOX2 on the retinal level of PPARγ in diabetic mice using Western blot analysis. Results of this study demonstrated a noticeable decrease in PPARγ levels in retinas of diabetic mice compared with control WT or NOX-knockout (0.49 ± 0.051 vs. 1.49 ± 0.048 and 1.33 ± 0.03, respectively). However, PPARγ expression was restored by the deletion of NOX2 (1.64 ± 0.3) or apocynin treatment (1.55 ± 0.19; Fig. 1A ). These observations were confirmed by the immunofluorescence technique using a specific endothelial cell marker (GSI) and anti-PPARγ. PPARγ immunoreactivity was stronger in the retinas of control WT and knockout mice than in those of diabetic WT mice, particularly in retinal vessels and surrounding retinal cells shaped similarly to Müller glial cells. However, the deletion of NOX2 or apocynin treatment restored the normal level of PPARγ expression and its distribution in the diabetic retina (Figs. 1B 1C) . No immunofluorescence reaction was observed in retinal sections treated only by the secondary antibody, indicating the specificity of the reaction. 

Effect of HG on PPARγ Expression in Cultured Endothelial Cells

Because our immunofluorescence demonstrated that PPARγ is localized in the endothelial cells, we tested the effect of HG on PPARγ in cultured endothelial cells. Our experiment showed a significant decrease in the level of PPARγ in BRECS by HG compared with the NG or mannitol (1420.0 ± 177 vs. 2466.0 ± 67 and 2742 ± 153, respectively). This effect was prevented by SOD (2806 ± 267) and NADPH oxidase inhibitors (apocynin, 2314 + 57; DPI, 2446 ± 54; Fig. 2 ). Similar results were noticed in HUVECs incubated under the same conditions (data not shown). 

Effect of Apocynin or Deletion of NOX2 on Diabetes-Induced NFκB Activation

Previous studies have demonstrated the link between NADPH oxidase and vascular inflammation in diabetic retinopathy. ^{20 21} In the present study, we tested whether this link occurs through the NFκB-dependent signaling pathway. Our experiments demonstrated a significant activation of retinal NFκB, as shown by the increases in the level of p-NFκB in diabetic retina compared with the control (p-NFκB, 4.5 ± 1.3 vs. 0.19). This effect was prevented by apocynin (p-NFκB, 1.6 ± 0.4) or by the deletion of NOX2 (p-NFκB, 0.98 ± 0.3; Fig. 3A ). This result was confirmed by the NFκB p65 transcription factor assay kit (TransAM). The assay showed a significant increase in the activity of NFκB in diabetic retina compared with control WT and NOX knockout mice. This increase was blocked in mice treated with apocynin or lacking NOX2 (Fig. 3B) . 

PPARγ Expression in a Mouse Model of OIR

We also tested the changes in PPARγ expression in relation to retinal neovascularization in a mouse model of OIR. Western blot analysis demonstrated a significant suppression of PPARγ expression in OIR at P17 (1.7 ± 0.26 vs. 2.1 ± 0.26) that was restored by apocynin treatment (2.7 ± 0.44; Fig. 4A ). The decrease in PPARγ expression started early during the course of OIR at P14 and was more obvious than at P17 if compared with the age-matched control (2.2 ± 0.3 vs. 15.3 ± 7; Fig. 4B ). In addition, immunolocalization using double labeling of retinal sections with endothelial cell marker (GSI) and anti-PPARγ also showed the restoration of PPARγ in the retina of OIR by apocynin (Fig. 4C) . 

PPARγ Expression in Acute Retinal Inflammation

We have previously reported that injection of LPS into mice is associated with increased retinal expression of ICAM-1 and leukostasis and that the deletion of NOX2 prevents these effects. ²¹ Here, we tested whether this was associated with any changes in the expression of PPARγ. Our experiments demonstrated a significant decrease in the retinal expression of PPARγ in LPS-injected mice compared with control (2001 ± 127 vs. 3121 ± 205) that was prevented in mice lacking NOX2 (3139 ± 208; Fig. 5 ). 

Discussion

This study examined changes in the expression of PPARγ in diabetic and ischemic retina and the role NADPH oxidase plays, if any, in mediating these changes. There are two main findings of this work. The first is that PPARγ is expressed in normal retina, particularly in blood vessels and Müller cells, and that it is suppressed in experimental models of diabetes, OIR, and acute retinal inflammation and in endothelial cells treated with HG. The second is that the inhibition of NADPH oxidase by apocynin or the deletion of NOX2 restores normal levels of retinal PPARγ, associated with decreased levels of activated NFκB in diabetic retina. To the best of our knowledge, this is the first report characterizing the expression of PPARγ in retina under normal and pathologic conditions associated with vascular dysfunction. Furthermore, it is the first report to show the role of NADPH oxidase in mediating changes in the levels of PPARγ in diabetic retina and OIR. 

In our previous study on a mouse model of STZ-induced diabetes, we showed increases in retinal intracellular adhesion molecule (ICAM)-1 expression, leukocyte adhesion, and vascular permeability. These effects of diabetes have been prevented by apocynin or deletion of NOX2. ²¹ We also showed that apocynin blocks retinal neovascularization in OIR. ¹⁹ Here, we tested the changes in PPARγ expression during diabetic retinopathy and OIR and whether NADPH oxidase plays any role in mediating these changes. 

Activation of PPARγ using specific agonists such as thiazolidinediones (TZDs) has been used for the treatment of diabetes for the past several years. Treatment with TZDs has been shown to have a protective role on the vasculature of patients with diabetes. These cardiovascular protective effects are distinct from and additive to any beneficial vascular consequence of glucose lowering. ⁴² PPARγ is found in endothelial cells and in vascular smooth muscle cells, where its activation exerts anti-inflammatory and antiproliferative effects, suggesting that they may be of benefit in ameliorating chronic vascular inflammation. ⁴³ However, little is known about its role in diabetic retinopathy. The present study demonstrated that PPARγ is expressed in the vessels and glial cells of normal retina and is abrogated in the experimental models of diabetes, ischemic retinopathy, and retinal inflammation, which are known to exhibit significant vascular injury. This finding is consistent with what has been reported in diabetes. For example, PPARγ has been shown to be decreased in the subcutaneous tissue of obese subjects with type 2 diabetes ⁴⁴ and in peritoneal macrophages from rats with alloxan-induced diabetes. ⁴⁵ In addition, PPARγ expression was decreased in aorta and heart tissues of long-term glucose-fed rats and was restored by combination therapy of antioxidants and a-lipoic acid. ⁴⁶ Streptozotocin-induced diabetes has also been reported to suppress adipose tissue PPARγ expression by 75% in normal mice with partial restoration during insulin treatment. ⁴⁷ However, this is the first report to show the suppression of PPARγ in diabetic retina. 

Furthermore, treating endothelial cells with HG caused a significant decrease in PPARγ expression. Suppression of PPARγ in diabetes was associated with the activation of NFκB. Because vascular inflammation is crucial in the pathogenesis of vascular dysfunction, such as hyperpermeability ²⁶ and neovascularization ²⁷ associated with diabetic or ischemic retinopathy, we suggest that PPARγ may exert its protective effect by way of an anti-inflammatory pathway. This suggestion is consistent with the recent report by Muranaka et al., ³³ who showed the inhibition of ICAM-1 expression, leukocyte adhesion, and retinal vascular leakage in experimental diabetes by the PPARγ agonist rosiglitazone and the increase in the same parameters by deletion of the gene encoding PPARγ. Moreover, rosiglitazone has been shown to inhibit retinal neovascularization in OIR by a mechanism downstream from VEGF. ⁴⁸ This probably occurs through targeting ICAM-1 because VEGF-induced angiogenesis is blocked in ICAM-1–deficient mice. ^{49 50} In addition, Miyahara et al. ⁵¹ suggest that ICAM-1 is involved in VEGF-induced leukocyte-endothelial cell interactions and subsequent blood-retinal barrier (BRB) breakdown in the diabetic retina. 

Oxidative stress plays a crucial role in the pathogenesis of vascular dysfunction in diabetes. The sources and mechanism of reactive oxygen species effects on diabetic vasculature continue to be defined. Because NADPH oxidase-derived ROS are a major factor in triggering vascular dysfunctions in diabetes, ^{20 52} OIR, ¹⁹ and LPS-induced endotoxemia, ⁵³ we tested whether PPARγ is involved in this process. Our experiments showed that the inhibition of NADPH oxidase using apocynin or the deletion of NOX2 restores the suppressed retinal PPARγ in diabetic, OIR, and LPS-injected mice, indicating that NADPH oxidase has a negative regulatory effect on PPARγ. The findings of our in vitro experiments were consistent with the previously mentioned in vivo study findings. These experiments showed that HG suppresses PPARγ expression in cultured BRECs and HUVECs but that it is prevented by SOD and NADPH oxidase inhibitors (apocynin and DPI), indicating that superoxide generation by NADPH plays an important role in mediating the effect of HG on PPARγ expression in endothelial cells. 

Restoration of the retinal level of PPARγ by NADPH oxidase inhibition was associated with the abrogation of inflammatory signaling, as shown by the decreases in NFκB activation. Of note, the effect of apocynin on retinal PPARγ and NFκB was similar to that of NOX2 deletion, which indicates its specificity as an NADPH oxidase inhibitor. Our previous findings in diabetic and LPS-injected mice demonstrated the inhibition of ICAM-1 expression and the abrogation of leukocyte adhesion by apocynin or the deletion of NOX2, and this was associated with the preservation of BRB in diabetic mice. ²¹ Furthermore, apocynin treatment blocked retinal neovascularization in OIR. ¹⁹ Interestingly, the decrease in PPARγ expression follows the same pattern of the increase in retinal expression of NOX2 in OIR. Although the decrease in PPARγ expression was more obvious by P14, the increase in NOX2 expression was also more prominent by P14 than at P17. ¹⁹ These data suggest an inverse relationship between the levels of NOX2 and PPARγ that could be involved in the pathogenesis of retinal neovascularization. 

Correlating our previous findings with the current data led us to suggest that the restoration of PPARγ could be an effective therapeutic strategy in preventing retinal vascular inflammation, a crucial event in the pathogenesis of diabetes or OIR. The inhibitory effect of NADPH oxidase on PPARγ gives a novel insight for how NADPH oxidase mediates vascular inflammation in diabetic retinopathy and in other vascular complications of diabetes such as atherosclerosis. Hwang et al. ⁵⁴ have reported that PPARγ ligands reduce superoxide anion generation in vascular endothelial cells by inhibiting NADPH oxidase. Hence, PPARγ restoration by NADPH oxidase inhibitors might lead to further inhibition of ROS generation by NADPH oxidase. More experiments are needed to elucidate how NADPH oxidase modulates PPARγ expression and activity in diabetic retinopathy. 

Because NFκB has been suggested to be the major redox-sensitive transcriptional regulator of endothelial adhesion molecules such as ICAM-1, and vascular cell adhesion molecule-1, we tested the effect of NADPH oxidase inhibition on its activation. Activation of NFκB is associated with the phosphorylation and degradation of the inhibitor κB (IκB) and with the nuclear translocation of NFκB subunit p65. ⁵⁵ In resting cells, NFκB is inactive because IκB proteins retain it in the cytoplasm and prevent DNA binding. ⁵⁶ Our data show an activation of NFκB in retinas of diabetic mice that was abrogated in mice lacking NOX2 or treated with apocynin. These findings clearly indicate that vascular inflammation associated with diabetic retinopathy is mediated through NADPH oxidase-dependent activation of NFκB, which leads to increased ICAM-1 expression and leukocyte-endothelial interaction. These data, together with the decreased PPARγ level in diabetic retina, support the link between PPARγ and retinal vascular inflammation in diabetic retinopathy. 

In summary, our findings indicate that suppression of PPARγ is downstream from NADPH oxidase activation in diabetic and ischemic retinopathies. Suppression of PPARγ leads to activation of the NFκB signaling pathway, including the upregulation of ICAM-1, increased leukocyte-endothelial interaction, and retinal vascular dysfunction. Targeting this signaling pathway could be beneficial in preventing retinal vascular damage induced by hyperglycemia. 

 Supported by the Scientist Development Grant AHA00104 from the American Heart Association.

 Submitted for publication March 11, 2008; revised August 18 and September 9, 2008; accepted December 9, 2008.

 Disclosure: A. Tawfik, None; T. Sanders, None; K. Kahook, None; S. Akeel, None; A. Elmarakby, None; M. Al-Shabrawey, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

PPARγ expression in diabetic retina. (A) Western blot analysis showed the suppression of PPARγ expression in diabetic wild-type mice (D) compared with the control WT (C) and NOX2 knockout (NOX^−/−) mice. Deletion of NOX2 (D-NOX^−/−) or apocynin treatment (D+apo) restored retinal levels of PPARγ in the diabetic mice. *P < 0.05 vs. C and NOX^−/−; #P < 0.05 vs. D (n = 6). (B) Immunofluorescence of retinal sections using endothelial cell marker (red) and anti-PPARγ (green) shows more PPARγ immunoreactivity in retinas of the normal WT (C) and NOX^−/− mice than in the WT diabetic (D) mice. Diabetic mice lacking NOX2 (D-NOX^−/−) or treated with apocynin (D+apo) showed marked restoration of PPARγ. Note that PPARγ is expressed in different layers of retina, particularly in retinal vessels (arrows) and other cells that look morphologically similar to the Müller glial cells (arrowhead). (C) Densitometry analysis of the reaction intensity shows that the decrease in PPARγ immunoreactivity by diabetes (D) was significant compared with the control WT (C) or NOX knockout mice (NOX^−/−). This decrease was significantly restored by the deletion of NOX2 (D-NOX^−/−) or by apocynin treatment (D+apo). *P < 0.05 vs. C and NOX^−/−; #P < 0.05 vs. D (n = 5).

Figure 2.

 

PPARγ expression in cultured BRECs. Western blot analysis showed the suppression of PPARγ expression in high glucose-treated BRECs (HG) compared with the normal glucose (NG) and mannitol (M)-treated cells. Superoxide dismutase (HG+SOD), apocynin (HG+apo), and DPI (HG+DPI) prevented the effect of HG on PPARγ expression in cultured BRECs. *P < 0.05 vs. NG and M; #P < 0.05 vs. HG (n = 3).

Figure 3.

 

Assay of NFκB activity in retina. (A) Western blot analysis of p-NFκB shows a significant increase in the retinal level of p-NFκB by diabetes (D) compared with control (C). Deletion of NOX2 (D-NOX^−/−) or apocynin treatment (D+apo) prevented the effects of diabetes on the retinal p-NFκB. (B) Measurement of NFκB activity by TransAM NFκB p65 transcription factor assay kit demonstrated a significant increase in the amount of active NFκB in the retinas of diabetic wild-type mice compared with control WT (C) and NOX2 knockout (NOX^−/−) mice. Deletion of NOX2 (D-NOX^−/−) or apocynin treatment (D+apo) prevented the effect of diabetes on NFκB activity. *P < 0.05 vs. C and NOX^−/−; #P < 0.05 vs. D (n = 6).

Figure 4.

 

Analysis of PPARγ expression in OIR. Western blot analysis of PPARγ in retinas of P17 (A) and P14 (B) mice. PPARγ expression increased in OIR compared with age-matched control (C). Note the early onset of PPARγ expression decrease in OIR by P14. Administration of apocynin (OIR+apo) restored the normal expression of PPARγ in OIR. *P < 0.05 vs. C; #P < 0.05 vs. OIR (n = 6). (C) Immunofluorescence of retinal sections shows that PPARγ (green) is expressed in retinal vessels (arrow) and surrounding retinal cells (arrowhead) of normal mice (C). There is a marked decrease in the PPARγ immunoreactivity in OIR compared with the control (C). Restoration of PPARγ in retinal vessels and related cells was noticed in the apocynin treated mice (OIR+apo).

Figure 5.

 

Western blot analysis of PPARγ in LPS-injected mice. There is a significant decrease in the retinal expression of PPARγ in the LPS-injected wild-type mice (LPS-WT) compared with the control (C). PPARγ level was restored in mice lacking NOX2 (LPS-NOX^−/−) (n = 7). *P < 0.05 vs. C; #P < 0.05 vs. LPS-WT.