Experimental Eye Research 83 (2006) 807e816 www.elsevier.

com/locate/yexer

Expression modication of uncoupling proteins and MnSOD in retinal endothelial cells and pericytes induced by high glucose: The role of reactive oxygen species in diabetic retinopathy
Yan Cui a,b, Xun Xu a,*, Hongsheng Bi b, Qi Zhu a, Jianfeng Wu b, Xin Xia a, Qiushi Ren c, Patrick C.P. Ho d
Department of Ophthalmology, Shanghai First Peoples Hospital, 85 Wu Jin Road, Shanghai 200080, P.R. China b Jinan Shierming Eye Hospital, 48 Jinan Ying Xiong Shan Road, Jinan 250002, P.R. China c Department of Biomedical Engineering, Shanghai Jiao-Tong University, 800 Dong Chuan Road, Shanghai 200240, P.R. China d 8/F Kailey Tower, 16 Stanley Street, Central, Hong Kong, P.R. China Received 9 November 2005; accepted in revised form 31 March 2006 Available online 5 June 2006
a

Abstract Uncoupling proteins (UCPs) are mitochondrial transporters present in the inner membrane of mitochondria. They belong to the family of anion mitochondrial carriers. UCPs could act as proton carriers activated by metabolites and create a shunt between complexes of the respiratory chain and ATP synthase. The increased leakiness of the mitochondrial inner membrane to protons may be to minimize superoxide production by limiting the maximum DmH. The purpose of this study was to detect UCP expression in retinal capillary cells and their modication in high levels of glucose. The role of reactive oxygen species (ROS) of mitochondria and UCPs in pathogenesis of diabetic retinopathy was investigated. Bovine retinal capillary endothelial cells and pericytes were cultured with selective culture media, respectively. Passage cells were cultured in three different glucose concentrations (5, 23, 30 mM) until passage four. ROS changes in mitochondria of these cells in different glucose concentrations were detected with scanning laser confocal microscopy (SLCM). The mitochondria membrane potential (Dj), cell death rate and apoptosis rate were measured with owing cytometry. UCP expression in retinal capillary cells was detected by immunocytochemistry. Expression and modication of MnSOD and uncoupling proteins (UCPs) in different concentrations of glucose were detected by means of semi-quantitative RTePCR. ROS in mitochondria of both endothelial cells and pericytes increased as the glucose concentration of media increased. Dj and cell death rate of endothelial cells increased also. ROS was correlated to Dj and cell death rate positively in endothelial cells. No difference in Dj and cell death rate among different glucose levels was found in pericytes. Apoptosis rate of endothelial cells and pericytes in high glucose levels was higher than that in lower glucose levels. UCP1 and UCP2 were expressed in cultured retinal capillary cells whereas UCP3 was not. At high levels of glucose, expression of UCP1, UCP2 and MnSOD increased to accommodate ROS production compensatively. The compensative mechanism disappeared when glucose concentration was too high (30 mM). The results of this study showed that increasing mitochondrial ROS could be induced by high glucose concentration. Those proteins related to antioxidation mechanism, such as MnSOD and UCPs, could exert compensative action to a certain extent. This compensative action was insufcient when the glucose concentration was too high. 2006 Elsevier Ltd. All rights reserved.
Keywords: diabetic retinopathy; reactive oxygen species; mitochondria; uncoupling proteins; MnSOD

1. Introduction Oxidative stress induced by hyperglycemia is an important pathway of diabetic microvascular complications. An increasing number of studies place mitochondrial ROS overproduction at the heart of pathogenesis of diabetic

* Corresponding author. Tel.: 86 21 6324 0090x3162; fax: 86 21 6324 0825. E-mail address: cy_wl@hotmail.com (X. Xu). 0014-4835/$ - see front matter 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.exer.2006.03.024

808

Y. Cui et al. / Experimental Eye Research 83 (2006) 807e816

microvascular complications. Normalizing mitochondrial superoxide production can block other pathways of hyperglycemic damage (Brownlee, 2001; Nishikawa et al., 2000). Mitochondrial electron transport chain is the main source of reactive oxygen species (ROS), such as superoxide anion and hydrogen peroxide. ROS production is associated with the activity of respiratory complexes I and III and with ubisemiquinone generated in the course of electron transport reactions in the respiratory chain. Mild uncoupling of respiration diminishes mitochondrial ROS formation by complexes I and III, because ROS formation depends on the mitochondrial proton gradient and the mitochondrial potential. That is, a mild uncoupling of respiration may participate in antioxidant defense and the uncoupling proteins (UCPs) may be the effectors of such a defense mechanism. Uncoupling proteins (UCPs) belong to the mitochondrial anion carrier gene family and, as suggested by their name, can uncouple ATP production from mitochondrial respiration by causing proton leak (Adams, 2000; Erlanson-Albertsson, 2003). As an elevated proton electrochemical potential gradient (DmH) favors superoxide production, limiting the magnitude of this gradient should decrease superoxide production (Skulachev, 1996). The increased leakiness of the mitochondrial inner membrane to protons may be to minimize superoxide production by limiting the maximum DmH. Causing proton leak is the essential function of UCPs, which are mitochondrial transporters present in the inner membrane of mitochondria (Erlanson-Albertsson, 2002). Five different UCPs have already been identied: UCP1e4 and UCP5, or Brain mitochondrial carrier protein 1 (BMCP1). These proteins are expressed in different tissues and play different roles in cellular metabolism. UCPs are related to ROS production (Skulachev, 1998). Which kinds of UCPs isomers are expressed on retinal capillary cells? Does UCPs and MnSOD expression change in high glucose? To answer these questions, in this study we cultured bovine retinal endothelial cells and pericytes in different concentrations of glucose. Mitochondrial ROS quantication, its effects and possible causes were studied. The expression of UCPs and MnSOD in the retinal capillary cells and their modication in high glucose concentration were also studied. The role of ROS and UCPs on the pathogenesis of diabetic retinopathy was investigated.

endothelial cells and pericyte culture. For endothelial cell culture, the pellet was maintained in low-glucose Dulbeccos modied Eagle medium (DMEM; GIBCO, Grand Island, NY) supplemented with 10% human serum (Sigma), 100 mg/ ml heparin (Sigma) and 10 mmol/l HEPES (GIBCO). The culture plate was coated with Fibronectin (FN, GIBCO, Invitrogen, CA) beforehand. The medium for pericytes was low-glucose DMEM supplemented with 10% FBS and the plate did not need to be pretreated. After reaching conuency, bovine endothelial cells and pericytes were maintained in DMEM containing 10% human serum and 10% FBS, respectively, for passage. The passage endothelial cells and pericytes were incubated with different glucose concentrations, i.e. 5 mM, 23 mM and 30 mM. Equimolar mannitol was supplemented for osmotic controls. Lab-Tek II Chamber Slide System (Nalge Nunc International, USA) was used for later confocal microscopy examination. The endothelial cells and pericytes generated after four passages were used in the latter experiments including identication with Von Willebrand polyclonal antibody (Santa Cruz) and Actin, Smooth Muscle Ab-7 (Neo Markers). UCP expression of endothelial cells and pericytes was detected by an immunocytochemistry test. The process followed the staining protocol of Histostain-Plus Kits (Zymed Laboratories Inc., South San Francisco, CA). The primary antibody included UCP1, -2 and -3 (20 mg/ml rabbit polyclonal afnity puried IgG, Alpha Diagnostic International, TX, USA). Finally, color was developed with DAB chromogen (Reagent D of Histostain-Plus Kits). For negative controls, primary antibody was omitted.

2.2. Localization and quantication of ROS Localization and quantication of ROS production was investigated in cultured cells double-labeled with 20 ,70 -dichlorodihydrouorescein diacetate (H2DCFDA; Molecular Probes, Eugene, OR) to detect ROS production and MitoTracker Red CM-H2XRos (MTR; Molecular Probes) to visualize mitochondria. MTR is a cationic lipophilic compound that readily accumulates in mitochondria that possess a membrane potential, and it becomes uorescent when oxidized. MTR has a longer emission wavelength (608 nm), which is well resolved from the 525-nm emission wavelength of green uorescein analogues such as DCF. Conuent bovine retinal endothelial cells and pericytes cultured in different glucose concentrations were rinsed with PBS two times. After adding H2DCFDA and MTR into culture media with nal concentrations of 2 mM and 500 nM, cells were incubated at 37  C for 1 h, then washed twice with fresh pre-warmed medium and imaged using a Zeiss LSM 510 laser scanning confocal microscope. Green uorescence of DCF was excited at 488 nm and excitation of MTR was achieved at 543 nm using a helium/neon laser. To determine the site of high glucose-induced intracellular ROS production, endothelial cells and pericytes cultured in 23 mM and 30 mM glucose were rst incubated with carbonyl cyanide m-chlorophenylhydrazone (CCCP; Sigma), the

2. Materials and methods 2.1. Cell culture, identication and UCP expression The primary culture approach of retinal capillary endothelial cells and pericytes was similar to the previous reports (Capetandes and Gerritsen, 1990; Kim et al., 2002). In brief, the retina of fresh bovine eyes without pigment cells was homogenized and digested in 0.05% collagenase I (Sigma Chemical, St. Louis, MO) with 0.025% BSA at 37  C for 45 min. The digestive material was ltered with 88-mm mesh and collected. After centrifugation, the pellet tissue was used for both

Y. Cui et al. / Experimental Eye Research 83 (2006) 807e816

809

protonophoric uncoupler with nal concentration 0.5 mM at 37  C for 5 min. The foregoing steps were then repeated. Images were background-corrected before analysis and the average uorescence intensity per mm2 of cell area was calculated using the analysis software package supplied by Zeiss. 2.3. Assessment of mitochondrial membrane potential (Dj), cell death rate and apoptosis rate Mitochondrial membrane potential (Dj) changes were detected using the mitochondrial membrane potential-sensitive uorescent dye 5,50 ,6,60 -tetrachloro-1,10 ,3,30 -tetraethylbenzimidazolyl-carbocyanine iodide (JC-1; Molecular Probes). The endothelial cells and pericytes cultured in different glucose were collected, pelleted, washed and resuspended in PBS. For each sample, cells were suspended in 1 ml warm PBS at approximately 1 106 cells/ml containing 3 mM JC-1 and the cells incubated at 37  C, 5% CO2 for 30 min. After loading, the cells were washed twice with PBS and analyzed by ow cytometry (Beckman Coulter XL-4) (excitation at 488 nm, with a 525 nm band-pass lter to collect green emission and a 590 nm band-pass lter to collect orange emission). For the control tube, CCCP (Sigma) was added with nal concentration 50 mM and the cells incubated at 37  C for 5 min. Using the CCCP-treated sample, standard compensation was performed. Mitochondrial membrane potential was indicated by the orange/green uorescence intensity ratio. Apoptosis and cell death rate were measured using Alexa Fluor 488-conjugated annexin V and propidium iodide (PI) labeling (Molecular Probes), respectively. The endothelial cells and pericytes cultured in different glucose concentrations were harvested and washed twice in cold PBS and suspended in 0.5 ml annexin-binding buffer with Alexa Fluor 488conjugated annexin V (1:10 nal dilution) and 10 mg/ml PI at room temperature for 15 min. Fluorescence was measured by a ow cytometer (Beckman Coulter XL-4) as soon as possible. The emission wavelength was 530 nm and 675 nm, respectively. Apoptotic cells show green uorescence and dead cells show both red and green uorescence. Final analysis was accomplished on a PC computer with WinMDI 2.9 software. Six samples in each group were tested. Correlation between Dj, cell death rate and ROS was analyzed. 2.4. Detection of UCPs and MnSOD mRNA by semi-quantitative RTePCR Total RNAs were extracted from each 3.5 mm culture plate of passage 4 endothelial cells and pericytes using a TRIzol Max Kit (Invitrogen, CA, USA). Five micrograms of RNA was reverse transcribed using a commercial kit (Promega, Madison, WI). Samples of cDNA were subjected to UCP1, UCP2, UCP3 and MnSOD amplication with b-actin as a housekeeping gene. The primers were designed using Primer Express software (PerkineElmer) as shown in Table 1. Reactions were performed in 50 ml volume containing 5 ml buffer, 1 ml of 10 mM of dNTP, 25 mmol each amplication primer, 5 U Taq polymerase and 10 ml RT product. Subsequent cycles

Table 1 Sequences of primers and Tm values for RTePCR of bovine b-actin, UCP1e3 and MnSOD Gene (GenBank accession #) b-Actin (AY141970) UCP1 (RNUCPB24) UCP2 (AF127029) UCP3 (NM_174210) MnSOD (S67818) Primer sequence 50 -TCGTGATGGACTCCGGTGAC-30 and 50 -AGCACCGTGTTGGCGTAGAG-30 50 -CTGGGAAGAAGACGGAACAC-30 and 50 -ACATTTGCCAGGGTCTACAC-30 50 -CAGTTCTACACCAAGGGCTC-30 and 50 -GGGAGGTCGTCTGTCATTAG -30 50 -ATCATTACCCGGATTTTGGC-30 and 50 -CCCGTTTCATCTGCTCGTAG-30 50 -GGACAAATCTGAGCCCTAAC-30 and 50 -CTCCTTATTGAAGCCGAGCC-30 Length (bp) 445 Tm (  C) 61.9

221

59.9

332

59.9

550

59.9

161

59.9

of PCR were performed using the following conditions: 95  C for 5 min, then 30 cycles of 1 min denaturation at 94  C, 1 min annealing at each Tm (Table 1), 1 min extension at 72  C, and a 10-min terminal extension at 70  C. PCR products were electrophoresed on 1.5% agarose gel and imaged by Alphaimager 2200 (Alpha Innotech, CA, USA). Quantitative analysis was achieved by measuring the integrated density values (IDV) of PCR products in gel photographs using AlphaEase FC 3.12 Image software. 2.5. Statistics Data are reported as the mean SE. Data in two groups were analyzed with an independent-samples t-test. Multiple comparisons used one-way ANOVA. The relationship between two factors was analyzed by bivariate correlate analysis. The minimum level of signicance was set at P < 0.05. 3. Results 3.1. Cell culture, identication and UCP expressions Pericytes and endothelial cells grew well in the selective media with their morphology characteristics (Wong et al., 1987; Porta et al., 1994). In these selective media, preparations became pure gradually after 3e4 passages. Pericytes were homogeneously positive for smooth muscle actin antigen and negative for Von Willebrand (Fig. 1A,C), whereas a nely granular cytoplasmic staining of Von Willebrand was detected in endothelial cells and was negative for smooth muscle actin antigen (Fig. 1B,D). At the fourth passage, pericytes were cultured in different glucose concentrations for 24e25 days, while endothelial cells were cultured for 45e46 days. UCP1 and UCP2 were detected positively by the immunocytochemical method in both endothelial cells (Fig. 2A,B) and pericytes (Fig. 2D,E). For these two kinds of cells, UCP3 expression was negative (Fig. 2C,F).

810

Y. Cui et al. / Experimental Eye Research 83 (2006) 807e816

Fig. 1. Identication of cultured capillary cells. Pericytes were homogeneously positive for smooth muscle actin antigen (A) and negative for Von Willebrand (C), whereas a nely granular cytoplasmic staining of Von Willebrand was detected in endothelial cells (B) and negative for smooth muscle actin antigen (D).

3.2. Localization of high glucose-induced ROS to mitochondria Confocal microscopy was used to localize the high concentration glucose-induced increase of ROS production to

mitochondria. The intensity of DCF uorescence was more intense in endothelial cells (Fig. 3B,E,H) and pericytes (Fig. 4B,E,H) that had been exposed to high levels of glucose. Quantitative analysis of the DCF uorescence indicated an increase in the relative uorescence intensity per mm2 of cell area

Fig. 2. UCP1 and UCP2 were detected positively by immunocytochemical method in both endothelial cells and pericytes. For these two kinds of cells, UCP3 expression was negative. (A), (B) and (C) are for endothelial cells UCP1, UCP2 and UCP3, respectively whereas (D), (E) and (F) are for pericytes UCP1, UCP2 and UCP3, respectively.

Y. Cui et al. / Experimental Eye Research 83 (2006) 807e816

811

Fig. 3. Localization of high glucose-induced ROS production to the mitochondria using confocal microscopy. Endothelial cells were incubated in different glucose in the presence of H2DCFDA (2 mM) and MTR (500 nM) for 1 h. (A), (B) and (C) were for 5 mM glucose; (D), (E) and (F) were for 23 mM glucose; (G), (H) and (I) were for 30 mM glucose. Red uorescence was MTR, which indicated mitochondria; green uorescence was DCF, which indicated ROS sites; yellow uorescence was those regions where MTR and DCF uorescence coincide.

in higher glucose as compared to the lower glucose both in endothelial cells (n 6, P 0.00) and in pericytes (n 6, P 0.01) (Fig. 5A). In contrast, there was no difference in the MTR relative uorescence intensity per mm2 of cell area in three-glucose concentration both in endothelial cells (n 6, P 0.20) and in pericytes (n 6, P 0.14) (Fig. 3A,D,G; Fig. 4A,D,G; Fig. 5B). To determine whether DCF uorescence (i.e. ROS production) was localized to mitochondria, the two images were superimposed. Thus, those regions where MTR and DCF uorescence coincide would appear yellow due to the combination of red and green pixels and provide direct evidence that mitochondria are a site of ROS production. There are signicantly more distinct subcellular yellow structures in cells exposed to higher glucose as compared to lower glucose, which indicates that mitochondria are the major source of ROS production after exposure to highconcentration glucose (Fig. 3C,F,I; Fig. 4C,F,I). DCF uorescence intensity of cells cultured in 23 mM or 30 mM glucose that were rst incubated with CCCP decreased to the level of baseline conditions (5 mM glucose) (n 6, P > 0.05). CCCP

completely prevented the ROS-inducing effect by high glucose (Fig. 5A). 3.3. Mitochondria membrane potential (Dj), cell death rate and apoptosis rate The lipophilic cationic compound 5,50 ,6,60 -tetrachloro1,10 ,3,30 -tetraethylbenzimidazolyl-carbocyanine iodide (JC-1) possesses the unique ability to differentially label mitochondria with low and high membrane potential. In mitochondria with high membrane potential, JC-1 forms aggregates emitting in the high orange wavelength of 590 nm. In mitochondria with low membrane potential, JC-1 forms monomers, which emit in the green wavelength (525e530 nm). FL1 (green uorescence from JC-1) and FL2 (orange uorescence from JC-1) were collected after compensation for spectral overlap. The ratio of orange/green uorescence was regarded as an index for the mitochondrial membrane potential (Dj). The mitochondrial membrane potential (Dj) of endothelial cells increased with increasing glucose in media (n 6,

812

Y. Cui et al. / Experimental Eye Research 83 (2006) 807e816

Fig. 4. Localization of high glucose-induced ROS production to the mitochondria using confocal microscopy. Pericytes were incubated in different concentrations of glucose in the presence of H2DCFDA (2 mM) and MTR (500 nM) for 1 h. (A), (B) and (C) were for 5 mM glucose; (D), (E) and (F) were for 23 mM glucose; (G), (H) and (I) were for 30 mM glucose. Red uorescence was MTR, which indicated mitochondria; green uorescence was DCF, which indicated ROS sites; yellow uorescence was those regions where MTR and DCF uorescence coincide.

P 0.01), as did the cell death rate of endothelial cells (n 6, P 0.02). But for pericytes, no difference in Dj (n 6, P 0.65) and cell death rate (n 6, P 0.94) was detected in media of different glucose concentrations (Fig. 6A,B). As for apoptosis rate, both endothelial cells and pericytes showed the same features. The apoptosis rate in high glucose concentrations (23 mM and 30 mM) was higher than that in low glucose concentration (5 mM) (n 6, P < 0.05). There was no signicant difference between apoptosis rates in 23 mM and 30 mM glucose (n 6, P > 0.05) (Fig. 6C). For endothelial cells, Dj was positively correlated to ROS (r 0.98, P 0.00). Cell death rate was positively correlated to ROS (r 0.97, P 0.00). For pericytes, no correlation relationship between Dj or cell death rate and ROS was found.

and pericytes. Neither endothelial cells nor pericytes expressed UCP3 mRNA (Fig. 7A,C). Secondly, UCP1, UCP2 and MnSOD mRNA expression levels were affected by glucose concentration. The changing pattern of UCP1 and UCP2 mRNA expression of endothelial cells and pericytes were similar. The expression level was greatest when the glucose concentration was 23 mM. As glucose concentration increased to 30 mM, UCP1 and UCP2 mRNA expression level decreased to the level of low glucose. MnSOD mRNA expression level of endothelial cells in 23 mM glucose was 2.22-fold the level in 5 mM glucose. In 30 mM glucose, it decreased to the level in 5 mM glucose (Fig. 7B). MnSOD mRNA expression level of pericytes in 23 mM glucose was 2.55-fold the level in 5 mM glucose. In 30 mM glucose, it decreased to a certain extent. But the level was signicantly greater than expression level in 5 mM glucose. It was 1.86-fold the level in 5 mM glucose (Fig. 7D). 4. Discussion Uncoupling proteins (UCPs) are mitochondrial transporters present in the inner membrane of mitochondria. They belong

3.4. UCPs and MnSOD mRNA expression in different glucose concentrations The RTePCR data revealed two main ndings. First, UCP1 and UCP2 mRNA was expressed in passage 4 endothelial cells

Y. Cui et al. / Experimental Eye Research 83 (2006) 807e816

813

Fig. 5. Quantitative analysis of the DCF and MTR uorescence intensity per mm2 of cell area in endothelial cells and pericytes. Data in (A) indicated an increased tendency of DCF intensity per mm2 of cell area following increasing glucose concentration, both in endothelial cells (n 6, P 0.00) and in pericytes (n 6, P 0.01). DCF uorescence intensity of cells cultured in 23 mM or 30 mM glucose that was rst incubated with CCCP decreased to the level of baseline conditions (5 mM glucose) (n 6, P > 0.05). (B) There was no difference in the MTR relative uorescence intensity per mm2 of cell area in three-glucose concentrations both in endothelial cells (n 6, P 0.20) and in pericytes (n 6, P 0.14).

to the family of anion mitochondrial carriers. UCPs could act as proton carriers activated by metabolites and create a shunt between complexes of the respiratory chain and ATP synthase. The increased leakiness of the mitochondrial inner membrane to protons may occur to minimize superoxide production by limiting the maximum DmH. Five different UCPs have already been identied: UCP1e4 and UCP5, or Brain mitochondrial carrier protein 1 (BMCP1). These proteins are expressed in different tissues and play different roles in cellular metabolism. Brain-specic UCP4 and UCP5 has been suggested to play a role in apoptosis in the brain (Mattson and Kroemer, 2003; Mattson and Liu, 2003). UCP1, UCP2 and UCP3 are involved in the limitation of free radical level in cells. Immunocytochemistry of UCP expression in endothelial cells and pericytes showed that UCP1 and UCP2 expression was positive in both kinds of cells, and UCP3 was negative. In these two kinds of cells, perhaps UCP1 and UCP2 regulate the production of ROS through proton leakage in different situations.

It is well known that ROS can damage the structure and function of cells directly. As this study showed, ROS of endothelial cells and pericytes increased following an increase in glucose concentration of media. These ROS were mitochondrial in origin. DCF uorescence intensity of cells cultured in 23 mM or 30 mM glucose that were rst incubated with CCCP, the protonophoric uncoupler, decreased to the level of baseline conditions. CCCP completely prevented the ROS-inducing effect by high glucose. Thus we concluded mitochondrial ROS increased inducing by high glucose levels. ROS overproduction could activate the other pathways toward diabetic microvascular complications. The intracellular AGE formation probably reects increased triose phosphate levels resulting from inhibition of GAPDH by mitochondrial overproduction of ROS (Anderson et al., 1999). Resulting from inhibition of GAPDH, dihydroxyacetone phosphate levels increased and PKC was activated by increasing the de novo synthesis of DAG (Koya and King, 1998). Hyperglycemia increases hexosamine pathway ux by providing more fructose-6-phosphate for GFATdthe rate-limiting enzyme of the pathwaydand the increasing level of fructose-6-phosphate was due to inhibition of GAPDH by ROS (Du et al., 2000). Therefore, mitochondrial ROS overproduction is likely to be the central factor that induced other pathways of diabetic retinopathy. For endothelial cells, Dj and cell death rate increased following the increase of glucose concentration in media. Both Dj and cell death rate were positively correlated with mitochondrial ROS. When the electrochemical potential difference generated by the proton gradient across the inner mitochondrial membrane is high, the lifetime of superoxide generating electron transport intermediates such as ubisemiquinone is prolonged. There seems to be a threshold value above which superoxide production is markedly increased (Young et al., 2002). Hyperglycemia increases the proton gradient above this threshold value as a result of overproduction of electron donors by the TCA cycle (Du et al., 2001). This, as a result, causes a marked increase in the production of superoxide by pericytes and endothelial cells. In turn, the increase of ROS can damage mitochondrial membrane directly or indirectly and decrease the Dj and ROS production. So a kind of feedback accommodation is formed. For pericytes, there was no obvious difference in Dj and cell death rate among different glucose media at the fourth passage. The relationship between Dj and ROS could not be simply explained by the former feedback accommodation theory. Perhaps other factors are involved between them. The cell death rate features of pericytes reect the ageing phenomenon of pericytes after several passages. As the component cells of retinal capillary, pericytes and endothelial cells exhibit different characteristics. Apoptosis of endothelial cells and pericytes was increased at high levels of glucose after 3 days of culture (Beltramo et al., 2004). Several groups (Li et al., 1998; Naruse et al., 2000; Romeo et al., 2002) have reported that cultured retinal pericytes exposed to high levels of glucose (25e30 mM) for

814

Y. Cui et al. / Experimental Eye Research 83 (2006) 807e816

Fig. 6. The mitochondrial membrane potential (Dj), cell death rate and apoptosis rate of endothelial cells and pericytes in different glucose media. (A) Dj of endothelial cells increased with increasing glucose in media (n 6, P 0.01). But for pericytes, no change in Dj (n 6, P 0.65) was detected in media of different glucose concentrations. (B) Cell death rate of endothelial cells increased following the increase of glucose concentration in media (n 6, P 0.02). For pericytes, no difference in cell death rate (n 6, P 0.94) was detected in media of different glucose concentrations. (C) Apoptosis rate of endothelial cells and pericytes at high glucose concentration (23 mM and 30 mM) was higher than that at low glucose concentration

a period of 7 days or more show a higher rate of apoptosis than cells grown at 5.5 mM glucose. Cultured in high glucose for 24e25 days and 45e46 days, apoptosis rates of pericytes and endothelial cells in high glucose concentration (23 mM and 30 mM) were higher than that in low glucose concentration (5 mM). There was no signicant difference between apoptosis rates in 23 mM and 30 mM glucose. Variation of apoptosis rate of capillary cells exposed to high levels of glucose during different periods needs further study. Immunocytochemistry and RTePCR proved that UCP1 and UCP2 were expressed in endothelial cells and pericytes of retinal capillaries. UCP3 was not expressed. What is the signicance of the UCP expression? Mitochondria are the cellular organelles where respiration occurs. They contain two compartments bound by inner and outer membranes. The outer membrane is permeable to small metabolites, whereas the permeability of the inner membrane is controlled to maintain the high electrochemical gradient created by the mitochondrial respiratory chain that is necessary for energy conservation and ATP synthesis in mitochondria. Energy produced by mitochondrial respiration is used for ATP synthesis by oxidative phosphorylation. Although respiration is coupled with ADP phosphorylation, this coupling is less than perfect and may be partially or very partially loose. The uncoupling proteins are particular mitochondrial transporters of the inner membrane that appear to be controlling the level of respiration coupling. The term uncoupling protein was originally used for UCP1, which is mainly present in mitochondria of brown adipocytes. In these cells, UCP1 acts as a proton carrier activated by free fatty acids and creates a shunt between complexes of the respiratory chain and ATP synthase. Activation of UCP1 enhances respiration, and the uncoupling process results in a futile cycle and dissipation of oxidation energy as heat. In comparison to the established uncoupling and thermogenic activities of UCP1, UCP2 and UCP3 appear to be involved in the limitation of free radical levels in cells rather than in physiological uncoupling and thermogenesis (Skulachev, 1998). Respiration is associated with production of ROS because the oxygen molecule is capable of accepting an additional electron to create the superoxide ion, a more reactive form of oxygen (Raha and Robinson, 2000). Mitochondria can produce a large part of the total ROS made in cells. Mitochondrial ROS increased in retinal capillary cells in high-level glucose. Being an effector of ROS production regulation by proton leak mechanism, uncoupling protein expression was modied in different levels of glucose. It is known that mild uncoupling of respiration diminishes mitochondrial ROS formation. The explanation for the control of ROS production by respiration uncoupling is that ROS formation depends on the mitochondrial proton gradient and the mitochondrial potential. In other words, a mild uncoupling of respiration may participate in antioxidant defense and the UCPs are just the effectors of such a defense mechanism (Arsenijevic et al., 2000).

(5 mM) (n 6, P < 0.05). There was no signicant difference between apoptosis rates in 23 mM and 30 mM glucose (n 6, P > 0.05) (Fig. 5C).

Y. Cui et al. / Experimental Eye Research 83 (2006) 807e816

815

Fig. 7. Expression of UCP1, UCP2 and MnSOD mRNA of bovine retinal endothelial cells and pericytes in different glucose concentrations (5 mM, 23 mM, 30 mM). Neither endothelial cells nor pericytes expressed UCP3 mRNA. The changing patterns of UCP1 and UCP2 mRNA expression of endothelial cells and pericytes were similar. The expression level was greatest when glucose concentration was 23 mM. As glucose concentration increased to 30 mM, UCP1 and UCP2 mRNA expression level decreased to the level of low glucose. (A,B) UCPs and MnSOD mRNA expression of endothelial cells in 5 mM, 23 mM and 30 mM glucose, respectively. MnSOD mRNA expression level of endothelial cells in 23 mM glucose was 2.22-fold the level in 5 mM glucose. In 30 mM glucose, it decreased to the level in 5 mM glucose. (C,D) Pericytes. MnSOD mRNA expression level of pericytes in 23 mM glucose was 2.55-fold the level in 5 mM glucose. In 30 mM glucose, it decreased to 1.86-fold the level in 5 mM glucose.

As this study showed, mRNA expression of UCP1, UCP2 and MnSOD increased as the glucose of media increased to 23 mM from 5 mM. But when the glucose concentration increased to 30 mM, this compensatory mechanism disappeared. Only mRNA expression of MnSOD of pericytes remained at a higher level than low glucose. Echtay et al. (2002a,b) have proposed a simple feedback cycle in which mitochondrial oxidative stress acutely and chronically upregulates the proton translocation activity of UCPs to lower the Dj and thus decrease superoxide production. When the compensatory mechanism cannot counteract ROS overproduction, relevant pathology changes will occur. This suggests that antioxidants targeted to mitochondria may be potential therapies for diabetes (Echtay et al., 2002a,b). Therapeutic strategies to counteract ROS through reinforcing complementary action of UCPs and MnSOD should be explored. Acknowledgements We thank staff of the central laboratory of Shandong Provincial Hospital for cell culture assistance. This study was supported by National Nature Science Funding of China (30471852), Project of Shanghai Science and Technology Committee (054119561) and National Basic Research Program of China (2005CB724302).

References
Adams, S.H., 2000. Uncoupling protein homologs: emerging views of physiological function. J. Nutr. 130, 711e714. Anderson, M.M., Requena, J.R., Crowley, J.R., Thorpe, S.R., Heinecke, J.W., 1999. The myeloperoxidase system of human phagocytes generates N3(carboxymethyl) lysine on proteins. J. Clin. Invest. 104, 103e113. Arsenijevic, D., Onuma, H., Pecqueur, C., Raimbault, S., Manning, B., Miroux, B., Goubern, M., Alves-Guerra, M.C., Couplan, E., Surwit, R., Bouillaud, F., Richard, D., Collins, S., Ricquier, D., 2000. Mice lacking uncoupling protein-2 survive to Toxoplasma gondii infection: a link with reactive oxygen species production and immunity. Nat. Genet. 26, 435e439. Beltramo, E., Berrone, E., Buttiglieri, S., Porta, M., 2004. Thiamine and benfotiamine prevent increased apoptosis in endothelial cells and pericytes cultured in high glucose. Diabetes Metab. Res. Rev. 20, 330e336. Brownlee, M., 2001. Biochemistry and molecular cell biology of diabetic complications. Nature 414, 813e820. Capetandes, A., Gerritsen, M.E., 1990. Simplied methods for consistent and selective culture of bovine retinal endothelial cells and pericytes. Invest. Ophthalmol. Vis. Sci. 31, 1738e1744. Du, X.L., Edelstein, D., Rossetti, L., Fantus, I.G., Goldberg, H., Ziyadeh, F., Wu, J., Brownlee, M., 2000. Hyperglycemia-induced mitochondrial superoxide overproduction activates the hexosamine pathway and induces plasminogen activator inhibitor-1 expression by increasing Sp1 glycosylation. Proc. Natl. Acad. Sci. USA 97, 12222e12226. Du, X.L., Edelstein, D., Dimmeler, S., Ju, Q., Sui, C., Brownlee, M., 2001. Hyperglycemia inhibits endothelial nitric oxide synthase activity by posttranslational modication at the AKT site. J. Clin. Invest. 108, 1341e1348.

816

Y. Cui et al. / Experimental Eye Research 83 (2006) 807e816 inhibition prevents glucose-induced apoptosis in cultured bovine retinal microvascular pericytes. Exp. Eye Res. 71, 309e315. Nishikawa, T., Edelstein, D., Du, X.L., Yamagishi, S., Matsumura, T., Kaneda, Y., Yorek, M.A., Beebe, D., Oates, P.J., Hammes, H.P., Giardino, I., Brownlee, M., 2000. Normalizing mitochondrial superoxide production blocks three pathways of hyperglycemic damage. Nature 404, 787e790. Porta, M., Molinatti, P.A., Dosso, A.A., Williams, F.M., Brooks, R.A., Kohner, E.M., 1994. Growth of bovine retinal pericytes and endothelial cells in high hexose concentrations. Diabetes Metab. 20, 25e30. Raha, S., Robinson, B.H., 2000. Mitochondria, oxygen free radicals, disease and ageing. TIBS 25, 502e508. Romeo, G., Liu, W.H., Asnaghi, V., Kern, T.S., Lorenzi, M., 2002. Activation of nuclear factor-kB induced by diabetes and high glucose regulates a proapoptotic program in retinal pericytes. Diabetes 51, 2241e2248. Skulachev, V.P., 1996. Role of uncoupled and non-coupled oxidations in maintenance of safely low levels of oxygen and its one-electron reductants. Q. Rev. Biophys. 29, 169e202. Skulachev, V.P., 1998. Uncoupling: new approaches to an old problem of bioenergetics. Biochim. Biophys. Acta 1363, 100e124. Wong, H.C., Boulton, M., Marshall, J., Clark, P., 1987. Growth of retinal capillary endothelia using pericyte conditioned medium. Invest. Ophthalmol. Vis. Sci. 28, 1767e1775. Young, T.A., Cunningham, C.C., Bailey, S.M., 2002. Reactive oxygen species production by the mitochondrial respiratory chain in isolated rat hepatocytes and liver mitochondria: studies using myxothiazol. Arch. of Biochem. and Biophys. 405, 65e72.

Echtay, K.S., Roussel, D., St-Pierre, J., Jekabsons, M.B., Cadenas, S., Stuart, J.A., Harper, J.A., Roebuck, S.J., Morrison, A., Pickering, S., Clapham, J.C., Brand, M.D., 2002a. Superoxide activates mitochondrial uncoupling proteins. Nature 415, 96e99. Echtay, K.S., Murphy, M.P., Smith, R.A., Talbot, D.A., Brand, M.D., 2002b. Superoxide activates mitochondrial uncoupling protein 2 from the matrix side: studies using targeted antioxidants. J. Biol. Chem. 277, 47129e 47135. Erlanson-Albertsson, C., 2002. Uncoupling proteins: a new family of proteins with unknown function. Nutr. Neurosci. 5, 1e11. Erlanson-Albertsson, C., 2003. The role of uncoupling proteins in the regulation of metabolism. Acta Physiol. Scand. 178, 405e412. Kim, J., Kim, K.S., Shinn, J.W., Oh, Y.S., Kim, H.T., Jo, I., Shinn, S.H., 2002. The effect of antioxidants on glycated albumin-induced cytotoxicity in bovine retinal pericytes. Biochem. Biophys. Res. Commun. 292, 1010e1016. Koya, D., King, G.L., 1998. Protein kinase C activation and the development of diabetic complications. Diabetes 47, 859e866. Li, W., Liu, X., He, Z., Yanoff, M., Jian, B., Ye, X., 1998. Expression of apoptosis regulatory genes by retinal pericytes after rapid glucose reduction. Invest. Ophthalmol. Vis. Sci. 39, 1535e1543. Mattson, M.P., Kroemer, G., 2003a. Mitochondria in cell death: novel targets for neuroprotection and cardioprotection. Trends Mol. Med. 9, 196e205. Mattson, M.P., Liu, D., 2003b. Mitochondrial potassium channels and uncoupling proteins in synaptic plasticity and neuronal cell death. Biochem. Biophys. Res. Commun. 304, 539e549. Naruse, K., Nakamura, J., Hamada, Y., Nakayama, M., Chaya, S., Komori, T., Kato, K., Kasuya, Y., Miwa, K., Hotta, N., 2000. Aldose reductase