Poly(ADP-Ribose)Polymerase Inhibition Counteracts Cataract Formation and Early Retinal Changes in Streptozotocin-Diabetic Rats

Home , PARP inhibitor

Viktor R. Drel,1 Weizheng Xu,2 Jie Zhang,2 Peter F. Kador,3,4 Tayyeba K. Ali,5 Jeho Shin,1 Ulrich Julius,6 Barbara Slusher,2 Azza B. El-Remessy,5 and Irina G. Obrosova1

PURPOSE. This study evaluated the role for poly(ADP-ribose) CONCLUSIONS. PARP activation is implicated in the formation of polymerase (PARP) in diabetes-induced cataractogenesis and diabetic cataract and in early retinal changes. These findings early retinal changes. provide a rationale for the development of PARP inhibitors for the prevention of diabetic ocular complications. (Invest Oph- METHODS. Control and streptozotocin (STZ)-diabetic rats were treated with or without the PARP inhibitors 1,5-iso- thalmol Vis Sci. 2009;50:1778–1790) DOI:10.1167/iovs.08-2191 quinolinediol (ISO; 3 mg kgϪ1 dϪ1 intraperitoneally) and 10-(4-methyl-piperazin-1-ylmethyl)-2H-7-oxa-1,2-diaza-benzo- rowing evidence suggests that the activation of poly(ADP- [de]anthracen-3–1 (GPI-15427, 30 mg kgϪ1 dϪ1 orally) for Gribose) polymerase (PARP), the enzyme that cleaves nic- ϩ 10 weeks after the first 2 weeks without treatment. Lens otinamide adenine dinucleotide (NAD ) with the formation of clarity was evaluated by indirect ophthalmoscopy and slit nicotinamide and poly(ADP-ribose) polymer, is an important lamp examination, and retinal changes were evaluated by event in the development of cardiovascular disease, cancer, 1,2 ϩ immunohistochemistry and Western blot analysis. In in vitro and diabetes mellitus. PARP activation contributes to NAD 1–3 studies, cultured human lens epithelial cells and bovine depletion and energy failure, changes in transcriptional reg- 1,2,4 5 retinal pericytes and endothelial cells were exposed to high ulation and gene expression, impaired signal transduction, 1,2 glucose or palmitate. and, in extreme cases, necrosis and apoptosis. In the past several years, it has been shown that PARP activation plays a RESULTS. PARP is expressed in lens, and poly(ADP-ribosyl)ated key role in diabetes-associated endothelial and myocardial dys- proteins are primarily localized in the 38- to 87-kDa range of function,1,6,7 peripheral and autonomic neuropathy,3,8,9 and the protein spectrum, with several minor bands at 17 to 38 nephropathy.10 kDa. The 38- to 87-kDa and the 17- to 38-kDa poly(ADP- The role of PARP in diabetic ocular complications deserves ribosyl)ated protein expression increased by 74% and 275%, thorough evaluation considering that PARP-1 is abundantly respectively, after 4 weeks of diabetes and by approximately expressed in lens11 and retina.12,13 PARP activation contributes 65% early after exposure of lens epithelial cells to 30 mM to the formation of pericyte ghosts and acellular capillaries,13 glucose. Both PARP inhibitors delayed, but did not prevent, the increased leukocyte adhesion to endothelial cells,13,14 and vas- formation of diabetic cataract. The number of TUNEL-positive cular endothelial growth factor (VEGF) formation15 and angio- nuclei in flatmounted retinas increased approximately 4-fold in genesis.16,17 The role of PARP activation in diabetes-associated STZ diabetic rats, and this increase was prevented by ISO and cataractogenesis remains unexplored. The consequences of GPI-15427. Both PARP inhibitors reduced diabetes-induced ret- retinal PARP activation at the early stages of diabetes and the inal oxidative-nitrosative and endoplasmic reticulum stress and relations of this mechanism to other stresses (oxidative-nitro- glial activation. GPI-15427 (20 ␮M) prevented oxidative-nitrosative stress, endoplasmic reticulum stress, neuroglial activa- sative stress and cell death in palmitate-exposed pericytes and tion, and premature neural retinal apoptosis) in the diabetic endothelial cells. retina have not been evaluated. Furthermore, although it has been established that PARP activation mediates high glucose– induced apoptosis in retinal endothelial cells,13 its role in

1 premature cell death caused by other factors in the diabetic From the Pennington Biomedical Research Center, Louisiana 18 State University System, Baton Rouge, Louisiana; 2MGI Pharma, Balti- milieu, particularly free fatty acids, is unknown. more, Maryland; 3Department of Pharmaceutical Sciences, College of The present study was aimed at evaluating the role or Pharmacy, and 4Department of Ophthalmology, College of Medicine, PARP activation in cataract formation and early retinal University of Nebraska Medical Center, Omaha, Nebraska; 5Program in changes associated with diabetes through the use of animal Clinical and Experimental Therapeutics, College of Pharmacy, Univer- and cell culture models and a pharmacologic approach with sity of Georgia, Augusta, Georgia; and 6Medical Clinic III, University two structurally unrelated PARP inhibitors, 1,5-isoquino- Hospital, Dresden, Germany. linediol (ISO) and 10-(4-methyl-piperazin-1-ylmethyl)-2H-7- Supported by Juvenile Diabetes Research Foundation Interna- oxa-1,2-diaza-benzo[de]anthracen-3-one (GPI-15427). tional Grants 1-2005-223 (IGO) and 2-2008-149 (ABE); National Insti- tutes of Health Grant R21DK070720 (IGO); and the American Heart Association (ABE). ATERIALS AND ETHODS Submitted for publication April 19, 2008; revised September 21, M M and November 4 and 19, 2008; accepted February 18, 2008. Disclosure: V.R. Drel, None; W. Xu, MGI Pharma (E), P; J. Zhang, Reagents MGI Pharma (E), P; P.F. Kador, None; T.K. Ali, None; J. Shin, None; Unless otherwise stated, all chemicals were of reagent-grade quality U. Julius, None; B. Slusher, MGI Pharma (E), P; A.B. El-Remessy, and were purchased from Sigma Chemical (St. Louis, MO). GPI-15427 None; , MGI Pharma (R) I.G. Obrosova was obtained from MGI Pharma (Baltimore, MD). Rabbit polyclonal The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertise- anti–nitrotyrosine (NT) antibody and mouse monoclonal anti–NT anti- ment” in accordance with 18 U.S.C. §1734 solely to indicate this fact. body, clone 1A6, were purchased from Upstate (Lake Placid, NY), and Corresponding author: Irina G. Obrosova, Pennington Biomedical mouse monoclonal anti–poly(ADP-ribose) antibody was purchased Research Center, Louisiana State University System, 6400 Perkins Road, from Trevigen, Inc. (Gaithersburg, MD). Mouse monoclonal anti–BiP Baton Rouge, LA 70808; [email protected]. (immunoglobulin heavy-chain binding protein)/GRP78 (78-kDa glu-

Downloaded from iovs.arvojournals.org on 09/26/2021 IOVS, April 2009, Vol. 50, No. 4 PARP, Diabetic Cataract, and Early Retinal Changes 1779

cose-regulated protein) antibody was purchased from BD Biosciences maldehyde in PBS for preparation of flatmounted retinas and quantita- (San Jose, CA). Rabbit polyclonal GRP94 antibody was purchased from tion of apoptosis. Rats from each experimental group were used for Abcam Inc. (Cambridge, MA). Secondary Alexa Fluor 488 goat anti– rapid dissection of the retina, which was immediately frozen in liquid rabbit and Alexa Fluor 488 goat anti–mouse antibodies, antifade re- nitrogen for subsequent Western blot analyses of nitrated and poly- agent (Prolong Gold), 4Ј,6-diamidino-2-phenylindole (DAPI), and hy- (ADP-ribosyl)ated proteins, BiP/GRP78, and GRP94. NT and poly(ADP- droxyethidine were purchased from Invitrogen (Eugene, OR). ribose) accumulations reflected the severity of oxidative-nitrosative Biotinylated anti–rabbit and anti–mouse antibody, avidin/biotin block- stress and the extent of PARP activation, respectively. GFAP is a marker ing kit, ABC kit (Vectastain Elite; Standard), and DAB substrate kit were of retinal glial activation, and BiP/GRP78 and GRP94 are markers of obtained from Vector Laboratories (Burlingame, CA). Mouse monoclo- endoplasmic reticulum stress response (a defense system for dealing nal anti–glial fibrillary acidic protein (GFAP) antibody and two in situ with the accumulation of unfolded proteins in the endoplasmic retic- apoptosis detection kits (ApopTag Plus Fluorescein and ApopTag Per- ulum [ER] lumen). oxidase; Chemicon International, Temecula, CA). A caspase assay kit (EnzChek Caspase-3) was purchased from Invitrogen (Carlsbad, CA). Specific Methods Used in Animal Studies Mounting medium (Micromount) was purchased from Surgipath Med- ical (Richmond, IL). Other reagents for immunohistochemistry were Immunohistochemical Studies. All flatmounted retinas purchased from Dako Laboratories (Santa Barbara, CA). were processed by a single investigator and evaluated blindly. The rate of apoptosis was quantified with an in situ apoptosis detection kit Animals (ApopTag Peroxidase; Chemicon International), as described previously,19,20 with a minor modification. NT, poly(ADP-ribose), GFAP, Experiments were performed in accordance with regulations specified BiP/GRP78, and GRP94 immunoreactivities in retinal sections were by the National Institutes of Health Principles of Laboratory Animal assessed by conventional immunohistochemistry.12,21 At least 10 fields Care (1985 revised version), the ARVO Statement for the Use of of each section were examined to select one representative image. Animals in Ophthalmic and Vision Research, and the Pennington Bio- Low-power observations of retinal sections stained for NT, poly(ADP- medical Research Center Protocol for Animal Studies. Male Wistar rats ribose), GFAP, BiP/GRP78, and GRP94 were made with a fluorescence (Charles River, Wilmington, MA), body weight 250–300 g, were fed a microscope (Axioskop; Carl Zeiss, Inc., Thornwood, NY). Color im- standard rat chow (PMI Nutrition International, Brentwood, MO) and ages were captured with a CCD camera at 1300 ϫ 1030 resolution had access to water ad libitum. Streptozotocin (STZ)-diabetes was (Axiocam HRc; Carl Zeiss, Inc.). Low-power images were generated induced as described.3,8 Blood samples for glucose measurements with a 40ϫ acroplan objective using the automatic capturing feature of were taken from the tail vein approximately 48 hours after STZ injec- the software (Axiovision, version 3.1.2.1; Carl Zeiss, Inc.). Low-power tion and the day before study termination. All rats with blood glucose observations of retinal sections stained for TUNEL-positive cells (Ap- levels Ն 13.8 mM were considered diabetic. The experimental groups opTag Plus Fluorescein In Situ Apoptosis Detection Kit; Chemicon were composed of control and diabetic rats treated with or without the Ϫ Ϫ International) were made with an imaging microscope (Axioplan 2; PARP inhibitors, ISO (3 mg kg 1 d 1 intraperitoneally) or GPI-15427 Ϫ Ϫ Carl Zeiss, Inc.). Fluorescent images were captured with a CCD camera (formulated as mesylate salt, 30 mg kg 1 d 1, in the drinking water), (CoolSNAP HQ; Photometrics, Tucson, AZ) at 1392 ϫ 1040 resolution. for 10 weeks after the first 2 weeks without treatment. At the end of Low-power images were generated with a 40ϫ acroplan objective with the 12-week study, lens changes were evaluated by indirect ophthal- image-acquisition software (RS Image 1.9.2; Photometrics). moscope and portable slit lamp (Kowa, Tokyo, Japan). Evaluations Western blot analyses of poly(ADP- were preceded by mydriasis with topical 1% tropicamide hydrochlo- Western Blot Analyses. ride. Cataracts were scored as follows: 1, no cataract (clear lenses); 2, ribosyl)ated and nitrated proteins, BiP/GRP78, and GRP94 in individual retinas (one retina from each rat) were performed as described previ- equatorial vacuoles; 3, cortical opacities; 4, mature cataract when the 12 whole lens becomes opaque. Control rats and rats with STZ diabetes of ously. Protein bands were visualized (BM Chemiluminescence Blot- 4 weeks’ duration were used for assessment of lens PARP and poly- ting Substrate; Roche, Indianapolis, IN). Membranes were then 12 ␤ (ADP-ribosyl)ated protein expression. stripped and reprobed with -actin antibody to confirm equal protein loading. Data were quantified by densitometry (Quantity One 4.5.0 Anesthesia, Euthanatization, and Tissue Sampling software; Bio-Rad Laboratories, Richmond, CA). Animals were sedated by CO and immediately killed by cervical 2 Cell Culture Studies dislocation. One eye from each rat was enucleated and fixed in normal buffered 4% formalin for further assessment of NT, poly(ADP-ribose), Human Lens Epithelial Cells. HLE cells, passages 6 to 10, GFAP, BiP/GRP78, and GRP94 immunoreactivities by conventional were supplied by the laboratory of Usha Andley at Washington Uni- immunohistochemistry. Several retinal sections from control and dia- versity (St. Louis, MO). In this laboratory, HLE cells were isolated from betic rats were used for obtaining representative pictures of apoptotic adult lenses obtained from MidAmerica Eye Bank (St. Louis, MO). The nuclei with an in situ apoptosis detection kit (ApopTag Plus Fluores- use of human tissue for research purposes conformed to the tenets of cein; Chemicon International). Another eye was fixed in 4% parafor- the Declaration of Helsinki. Lenses were processed less than 24 hours

TABLE 1. Initial and Final Body Weights and Blood Glucose Concentrations in Control and Diabetic Rats Maintained with and without PARP Inhibitor Treatment

Body Weight (g) Blood Glucose (mM)

Initial Final Initial Final

Control 291 Ϯ 8 565 Ϯ 76 5.7 Ϯ 0.55 5.5 Ϯ 1.38 Control ϩ GPI-15427 299 Ϯ 23 557 Ϯ 66 5.4 Ϯ 0.38 5.1 Ϯ 1.04 Control ϩ ISO 296 Ϯ 11 537 Ϯ 55 6.0 Ϯ 0.69 5.2 Ϯ 0.69 Diabetic 288 Ϯ 17 353 Ϯ 58* 25.4 Ϯ 5.50* 26.1 Ϯ 5.8* Diabetic ϩ GPI-15427 297 Ϯ 16 359 Ϯ 55* 26.2 Ϯ 3.81* 24.5 Ϯ 3.1* Diabetic ϩ ISO 298 Ϯ 17 339 Ϯ 80* 25.9 Ϯ 4.47* 25.3 Ϯ 4.9*

Data are mean Ϯ SD; n ϭ 12–20 per group. * P Ͻ 0.01 vs. controls.

Downloaded from iovs.arvojournals.org on 09/26/2021 1780 Drel et al. IOVS, April 2009, Vol. 50, No. 4

TABLE 2. Lens Clarity in Control Rats and Rats with 12-Week Duration of STZ-Induced Diabetes Maintained with and without PARP Inhibitor Treatment

CC؉ GPI C ؉ ISO D D ؉ GPI D ؉ ISO

Total (n)121216161616 Clear lenses (n)1212141 8 5 Vacuolar stage (n)0 0 2 4 2 3 Opacities (n)0007 4 6 Mature cataract (n)0 0 0 4 2 2 Cataract score* 1.0 Ϯ 0.0 1.0 Ϯ 0.0 1.125 Ϯ 0.34 2.88 Ϯ 0.89† 2.0 Ϯ 1.16†‡ 2.31 Ϯ 1.08†‡

* Values are mean Ϯ SD. Clear lenses were scored as 1, those with vacuoles were scored as 2, those with cortical opacities were scored as 3, and those with mature cataract were scored as 4. † P Ͻ 0.1 vs. controls. ‡ P Ͻ 0.01 vs. untreated diabetic rats.

after death. The capsule epithelium from each lens was dissected and MD) in Eagle minimum essential medium (EMEM) containing 50 divided into two or three pieces. Primary cultures were initiated in mg/mL gentamicin and 20% fetal bovine serum. After conﬂuence was 35-mm tissue culture plates and subcultured after conﬂuence was attained, cells were passaged using Trypsin-EDTA (T3924; Sigma), achieved, usually in 7 to 10 days. Cells were cultured on tissue culture frozen, and shipped to Pennington Biomedical Research Center. There, plasticware (Corning, Corning, NY or Falcon Plastics, Cockeysville, HLE cells were cultured in 6-well plates (well diameter, 3.5 cm) at a cell

FIGURE 1. Left: representative West- ern blot analyses of PARP-1 (A) and poly(ADP-ribosyl)ated proteins (C, E) in the lenses of rats with STZ diabetes of 4 weeks’ duration. Right: PARP-1 (B) and poly(ADP-ribosyl)ated protein (D, F) contents (densitometry) in control and diabetic rats. Equal protein loading was conﬁrmed with ␤-actin antibody. Poly(ADP-ribosyl)ated protein content in control rats is taken as 100%. M, standards of PARP-1 (A) and poly(ADP-ribosyl)ated proteins

(C, E); C1,C2,C3, lenses from control rats; D1,D2,D3, lenses from diabetic rats. Mean Ϯ SD. n ϭ 3 per group. *P Ͻ 0.05 vs. controls; **P Ͻ 0.01 vs. controls.

Downloaded from iovs.arvojournals.org on 09/26/2021 IOVS, April 2009, Vol. 50, No. 4 PARP, Diabetic Cataract, and Early Retinal Changes 1781

density of approximately 5 ϫ 104/well, at 37°C, in a humidiﬁed atmo- were used per experimental condition. Cells were placed on round

sphere consisting of air/CO2 (19:1) for 48 hours in EMEM containing glass coverslips and coated with gelatin or fibronectin (for pericytes either 5 or 30 mM glucose. HLE cells were used for Western blot and endothelial cells, respectively). At 80% confluence, pericyte and analyses of PARP-1 and poly(ADP-ribosyl)ated proteins performed as endothelial cell cultures were transferred for 48 hours to the media described previously.21 without palmitate and without GPI-15427, with 0.6 mM palmitate and Bovine Retinal Pericytes and Endothelial Cells. Cell without GPI-15427, or with 0.6 mM palmitate and with 20 ␮M GPI- Preparation. Primary bovine retinal pericyte and endothelial cell 15427. cultures were established from fresh cow eyes as described previous- Assessment of Apoptosis. The rate of cell death was quantified ly.22,23 Passages 4 to 6 were used for all experiments. Purity of the at the end of exposure (ApopTag Plus Fluorescein In Situ Apoptosis cultures was confirmed by characteristic pericyte and endothelial cell Detection Kit; Chemicon International). Parallel cultures from each morphology and by the use of specific pericyte (␣-smooth muscle group were trypsinized and used for assessment of caspase activity. actin) and endothelial cell (von Willebrand factor) markers. On aver- DEVD (rhodamine 110 bis-(N-CBZ-L-aspartyl-L-glutamyl-L-valyl-L-aspar- age, in pericyte experiments, 98.8% Ϯ 1.4% of the isolated cells were tic acid amide)-specific protease, that is, primarily those of caspase-3 identified as pericytes. In endothelial cell experiments, 99.5% Ϯ 1.1% and caspase-7, activity measurements were based on monitoring in- of the isolated cells were identified as endothelial cells. creases in fluorescence caused by conversion of the nonfluorescent To dissect the effects of palmitate (prepared as described24) and bisamide substrate to the fluorescent monoamine and even more PARP inhibition, pericytes and endothelial cells were cultured in the fluorescent derivative of rhodamine 110 and were performed with a Dulbecco modified Eagle medium containing 20% serum, 100 U/mL caspase assay kit (EnzChek Caspase-3; Invitrogen). Increases in fluo- penicillin, 100 mg/mL streptomycin, and, for endothelial cells only, 50 rescence were measured spectrofluorometrically at ␭ excitation/496 ␮g/mL endothelial growth supplement. At least three 6-well plates nm and ␭ emission/520 nm (LS 55 Luminescence Spectrometer

FIGURE 2. Left: representative West- ern blot analyses of PARP-1 (A) and poly(ADP-ribosyl)ated proteins (C, E) in HLE cells cultured for 48 hours in 5 mM and 30 mM glucose. Right: PARP-1 (B) and poly(ADP-ribosyl)- ated protein (D, F) contents (densitometry) in HLE cells cultured for 48 hours in 5 mM and 30 mM glucose. Equal protein loading was conﬁrmed with ␤-actin antibody. Poly(ADP-ribosyl)ated protein content in HLE cells cultured in 5 mM glucose is taken as 100%. M, standards of PARP-1 (A) and poly(ADP-ribosyl)ated proteins (C, E). Mean Ϯ SD. n ϭ 3 per group. **P Ͻ 0.01 vs. controls.

Downloaded from iovs.arvojournals.org on 09/26/2021 1782 Drel et al. IOVS, April 2009, Vol. 50, No. 4

equipped with microplate reader; Perkin Elmer, Waltham, MA). After CoolSNAP HQ) at 1392 ϫ 1040 resolutions. Fluorescence was quanti- spectrofluorometry, the cells were counted, and caspase activity was fied with ImageJ 1.32 software (developed by Wayne Rasband, Na- calculated in relative fluorescence units per 104 cells during 30 min- tional Institutes of Health, Bethesda, MD; available at http://rsb.info. utes of reaction. nih.gov/ij/index.html.). Seven to 10 images were quantified per Superoxide Detection. For superoxide production measure- experimental condition, and the average per cell was calculated. ments, pericytes and endothelial cells were cultured in 6-well plates in media containing 0, 0.2, 0.4, 0.6, or 0.8 mM palmitate. Culture media Statistical Analysis were then aspirated, and the cells were washed with PBS. Two milli- Results are expressed as mean Ϯ SD. Data were subjected to equality ␮ ␮ liters of serum-free medium containing 50 Lof10 M hydroethidine of variance F test and then to log transformation, if necessary, before was added per well at 37°C for 30 minutes. Then the cells were one-way analysis of variance. Where overall significance (P Ͻ 0.05) washed with PBS and trypsinized, and ethidium fluorescence, an index was attained, individual between-group comparisons were made using of superoxide generation, was measured with the use of spectroflu- the Student-Newman-Keuls multiple-range test. Significance was de- ␭ ␭ orometry at excitation/465 nm and emission/630 nm (LS 55 fined at P Յ 0.05. When between-group variance differences could not Luminescence Spectrometer equipped with microplate reader; Perkin be normalized by log transformation (data sets for body weights and Elmer). After spectrofluorometry, the cells were counted, and fluores- plasma glucose), the data were analyzed by nonparametric Kruskal- 4 cence intensity was expressed per 10 cells. Wallis one-way analysis of variance, followed by Bonferroni/Dunn or Immunocytochemical Assessment of Nitrotyrosine and Fisher PLSD test for multiple comparisons. Poly(ADP-ribose). Coverslips with pericyte or endothelial cells were washed in PBS and fixed in 4% paraformaldehyde for 10 minutes. Fixed cells were washed in PBS and preincubated with 0.2% Triton X-100 in RESULTS PBS for 15 minutes. Coverslips were blocked with 1% BSA containing 10% goat serum for 1 hour. Then the cells were treated with either Initial (before STZ administration) body weights were similar mouse monoclonal anti–poly(ADP-ribose) antibody (1:100 dilution) or in control and diabetic rats treated with or without ISO or rabbit polyclonal anti–NT antibody (1:200 dilution). Secondary Alexa GPI-15427. Final body weights were similarly reduced in un- Fluor 488 goat anti–mouse antibody or Alexa Fluor 488 goat anti–rabbit treated and PARP inhibitor–treated diabetic rats compared antibody was applied in working dilutions of 1:200. Primary antibody with the control group (Table 1). Initial blood glucose concen- was omitted in negative controls. Coverslips were mounted in antifade trations were 4.5-, 4.6-, and 4.5-fold higher in untreated, ISO-, reagent (Prolong Gold; Invitrogen) and placed on a slide. Images of and GPI-15427–treated diabetic rats, respectively, than in non- immunostained cells were captured with a CCD camera (Photometric diabetic controls. Similarly, final blood glucose concentrations

FIGURE 3. (A) Representative microphotographs of retinal poly(ADP-ribose) immunostaining in control and diabetic rats maintained with or without PARP inhibitor treatment. n ϭ 6 to 10 per group. Magniﬁcation, ϫ40. (B) Representative Western blot analyses of retinal poly(ADP- ribosyl)ated proteins in control and diabetic rats maintained with and without PARP inhibitor treatment. (C) Retinal poly(ADP-ribosyl)ated protein contents (densitometry) in control and diabetic rats maintained with and without PARP inhibitor treatment. Equal protein loading was conﬁrmed with ␤-actin antibody. Poly(ADP-ribosyl)ated protein content in control rats is taken as 100%. Mean Ϯ SD. n ϭ 6 per group. **P Ͻ 0.01 vs. controls; #P Ͻ 0.05 vs. controls; ##P Ͻ 0.01 vs. untreated diabetic group.

Downloaded from iovs.arvojournals.org on 09/26/2021 IOVS, April 2009, Vol. 50, No. 4 PARP, Diabetic Cataract, and Early Retinal Changes 1783

FIGURE 4. (A) Representative microphotographs of retinal nitrotyrosine immunostaining in control and diabetic rats maintained with or without PARP inhibitor treatment. n ϭ 6 to 10 per group. Magniﬁcation, ϫ40. (B) Representative Western blot analyses of retinal nitrosylated proteins in control and diabetic rats maintained with and without PARP inhibitor treatment. (C) Retinal nitrosylated protein contents (densitometry) in control and diabetic rats maintained with and without PARP inhibitor treatment. Equal protein loading was conﬁrmed with ␤-actin antibody. Nitrosylated protein content in control rats is taken as 100%. Mean Ϯ SD. n ϭ 6 per group. **P Ͻ 0.01 vs. controls; ##P Ͻ 0.01 vs. untreated diabetic group.

were 4.7-, 4.5-, and 4.6-fold higher in untreated, ISO-, and and ISO, respectively, compared with 44% in diabetic rats. GPI-15427–treated diabetic rats than in nondiabetic controls. Mature cataract (the whole lens was opaque) was detected in PARP inhibition did not affect weight gain or blood glucose 12.5% of eyes of diabetic rats treated with ISO or GPI-15427 concentration in nondiabetic rats. compared with 25% of eyes in the untreated diabetic group. The average cataract score was 2.88-fold higher in rats with PARP-1 expression was similar in the lenses of nondiabetic STZ diabetes of 12 weeks’ duration than in nondiabetic con- control rats and rats with STZ diabetes of 4 weeks’ duration trols (Table 2). Twenty-ﬁve percent of eyes of untreated dia- (Figs. 1A, 1B). Poly(ADP-ribosyl)ated proteins were detected betic rats displayed the vacuolar stage of cataract, 44% had primarily in the 38- to 87-kDa range of the lens protein spec- cortical opacities, and 25% had mature cataract. PARP inhibi- trum (Figs. 1C, 1D), with several minor bands at 17 to 38 kDa tion counteracted, but did not prevent, diabetes-associated (Figs. 1E, 1F). Both 38- to 87-kDa and 17- to 38-kDa poly(ADP- cataractogenesis. Clear lenses were detected in 50% and 31% of ribosyl)ated protein expression was increased in rats with eyes in diabetic rats treated with GPI-15427 and ISO, respec- 4-week STZ diabetes compared with nondiabetic controls. tively, but only in 6% of eyes in untreated diabetic rats. Vacu- In a similar fashion, PARP-1 expression was indistinguish- olar stage was diagnosed in 12% and 19% of lenses in diabetic able between HLE cells cultured in 5 mM or 30 mM glucose for rats treated with GPI-15427 and ISO, respectively, and in 25% 48 hours (Figs. 2A, 2B). Poly(ADP-ribosyl)ated proteins were of lenses in the untreated diabetic group. Opacities were re- abundantly expressed in the 17- to 38-kDa range of the lens vealed in 25% and 38% of diabetic rats treated with GPI-15427 protein spectrum, whereas the upper part of the spectrum

FIGURE 5. Representative microphotographs of retinal glial ﬁbrillary acidic protein immunostaining in control and diabetic rats maintained with and without PARP inhibitor treatment. n ϭ 6 to 10 per group. Magniﬁcation, ϫ40.

Downloaded from iovs.arvojournals.org on 09/26/2021 1784 Drel et al. IOVS, April 2009, Vol. 50, No. 4

displayed less manifest poly(ADP-ribosyl)ation (Figs. 2C, 2D). Poly(ADP-ribosyl)ated protein expression in the upper and lower parts of the lens protein spectrum was increased by approximately 65% early (48 hours) after exposure of HLE cells to 30 mM glucose. Retinal poly(ADP-ribose) immunoreactivity was increased in diabetic rats compared with nondiabetic controls, and this increase was essentially prevented by ISO and GPI-15427 (Fig. 3A). Poly(ADP-ribose) positive nuclei were localized primarily in the ganglion cell layer but were also detectable in other parts of the retina. Poly(ADP-ribosyl)ated protein expression, quantified by Western blot analysis, was increased by 41% in untreated diabetic rats compared with nondiabetic controls but remained essentially unchanged from the control level in diabetic rats treated with ISO or GPI-15427 (Figs. 3B, 3C). Increased nitrotyrosine immunoreactivity in all retinal layers was detected in untreated diabetic rats compared with nondiabetic controls, and this increase was essentially prevented by ISO and GPI-15427 (Fig. 4A). Nitrated protein expression, quantified by Western blot analysis, was increased by 53% in untreated diabetic rats compared with controls (Figs. 4B, 4C). ISO and GPI-15427 counteracted the accumulation of diabetes-associated retinal nitrotyrosine. Retinal GFAP immunoreactivity was increased in diabetic rats compared with controls, and this increase was counteracted by PARP inhibitors (Fig. 5). Representative images of TUNEL-positive cells (TUNEL fluorescence) in the retinal sections of control and diabetic rats are shown in Figure 6A. The number of TUNEL-positive nuclei in flatmounted retinas was increased approximately 4-fold in diabetic rats compared with the control group, and this increase was completely prevented by ISO or GPI-15427 (Figs. 6B, 6C). Immunoreactive BIP/GRP78 (Fig. 7A) and GRP94 (Fig. 8A) were identified in the retinas of control rats. GRP94 immunoreactivity displayed a uniform distribution among all retinal layers, whereas BIP/GRP78 distribution was less homoge- neous, and expression in inner plexiform and outer nuclear layers was faint. BiP/GRP78 expression was increased by 37% in untreated diabetic rats compared with controls, indicative of ER stress (Figs. 7B, 7C) ISO and GPI-15427 counteracted diabetes-induced BiP/GRP78 expression. Conversely, GRP94 expression showed a minor (9%) induction in untreated diabetic rats compared with nondiabetic controls (Figs. 8B, 8C) and remained in the nondiabetic range in ISO- and GPI-15427– treated diabetic rats. A 48-hour exposure to palmitate caused a dose-dependent increase in superoxide production in retinal pericytes and endothelial cells (Figs. 9A, 9B), with enhanced oxidative stress after exposure to 0.6 mM palmitate. Poly(ADP-ribosyl)ated protein fluorescence was increased in retinal pericytes (Figs. 10A, 10B) and endothelial cells (Figs. 10C, 10D) cultured with the addition of 0.6 mM palmitate compared with those cultured without palmitate. The PARP inhibitor GPI-15427 prevented the accumulation of poly(ADP- ribosyl)ated proteins in palmitate-exposed retinal microvascu- lar cells Nitrotyrosine fluorescence was increased in 0.6 mM palmitate-exposed cultured retinal pericytes (Figs. 11A, 11B) and endothelial cells (Figs. 11C, 11D). Palmitate-induced nitrosative stress in both cell types was reduced though not completely blunted by GPI-15427.

FIGURE 6. (A) Representative microphotographs of TUNEL-positive cells (arrows, point to TUNEL fluorescence) in the retinal sections of control and diabetic rats. (B) Representative microphotographs of positive cell counts per retina in control and diabetic rats maintained TUNEL-positive cells (TUNEL immunostaining) in the flatmounted ret- with and without PARP inhibitor treatment. Mean Ϯ SD. n ϭ 6to10 inas of control and diabetic rats maintained with and without PARP per group. **P Ͻ 0.01 vs. controls; ##P Ͻ 0.01 vs. untreated diabetic inhibitor treatment. Magnification: (A) ϫ40; (B) ϫ20. (C) TUNEL- group.

Downloaded from iovs.arvojournals.org on 09/26/2021 IOVS, April 2009, Vol. 50, No. 4 PARP, Diabetic Cataract, and Early Retinal Changes 1785

FIGURE 7. (A) Representative microphotographs of retinal BiP/GRP78 immunostaining in control and diabetic rats maintained with or without PARP inhibitor treatment. n ϭ 6 to 10 per group. Magniﬁcation, ϫ40. (B) Representative Western blot analyses of retinal BiP/GRP78 in control and diabetic rats maintained with and without PARP inhibitor treatment. (C) Retinal BiP/GRP78 contents (densitometry) in control and diabetic rats maintained with and without PARP inhibitor treatment. Equal protein loading was conﬁrmed with ␤-actin antibody. BiP/GRP78 content in control rats is taken as 100%. Mean Ϯ SD. n ϭ 6 per group. **P Ͻ 0.01 vs. controls; #P Ͻ 0.05 vs. untreated diabetic group.

FIGURE 8. (A) Representative microphotographs of retinal GRP94 immunostaining in control and diabetic rats maintained with or without PARP inhibitor treatment. n ϭ 6 to 10 per group. Magniﬁcation, ϫ40. (B) Representative Western blot analyses of retinal GRP94 in control and diabetic rats maintained with and without PARP inhibitor treatment. (C) Retinal GRP94 contents (densitometry) in control and diabetic rats maintained with and without PARP inhibitor treatment. Equal protein loading was conﬁrmed with ␤-actin antibody. GRP94 content in control rats is taken as 100%. Mean Ϯ SD. n ϭ 6 per group.

Downloaded from iovs.arvojournals.org on 09/26/2021 1786 Drel et al. IOVS, April 2009, Vol. 50, No. 4

FIGURE 9. Dose-dependent increase in superoxide ﬂuorescence in cultured retinal pericytes (A) and endothelial cells (B) exposed to 0, 0.2, 0.4, 0.6, and 0.8 mM palmitate. RFU, relative ﬂuorescence units. Mean Ϯ SD. n ϭ 4 per group. *P Ͻ 0.05 and **P Ͻ 0.01 vs. cells cultured in 5 mM glucose.

FIGURE 10. Left: representative microphotographs of poly(ADP-ribose) fluorescence (green) in retinal pericytes (A) and endothelial cells (C) cultured for 48 hours without 0.6 mM palmitate or GPI-15427 (C), with 0.6 mM palmitate and without GPI-15427 (P), and with 0.6 mM palmitate and with GPI-15427 (PϩGPI). Magnification, ϫ100. Blue fluorescence corresponds to 4Ј,6-diamidino-2-phenylindole-stained nuclei. Right: poly(ADP-ribose) fluorescence (relative fluorescence units [RFU] per cell) in retinal pericytes (B) and endothelial cells (D) cultured without 0.6 mM palmitate or GPI-15427 (C), with 0.6 mM palmitate and without GPI-15427 (P), and with 0.6 mM palmitate and with GPI-15427 (PϩGPI). Mean Ϯ SD. n ϭ 4 per group. **P Ͻ 0.01 vs. cells cultured without palmitate and without GPI-15427; ##P Ͻ 0.01 vs. cells cultured with palmitate and without GPI-15427.

Downloaded from iovs.arvojournals.org on 09/26/2021 IOVS, April 2009, Vol. 50, No. 4 PARP, Diabetic Cataract, and Early Retinal Changes 1787

FIGURE 11. Left: representative microphotographs of nitrotyrosine fluorescence (green) in retinal pericytes (A) and endothelial cells (C) cultured for 48 hours without 0.6 mM palmitate or GPI-15427 (C), with 0.6 mM palmitate and without GPI-15427 (P), and with 0.6 mM palmitate and with 20 ␮M GPI-15427 (PϩGPI), Magnification, ϫ100. Blue fluorescence corresponds to 4Ј,6-diamidino-2-phenylindole-stained nuclei. Right: nitrotyrosine fluorescence (relative fluorescence units [RFU] per cell) in retinal pericytes (B) and endothelial cells (D) cultured for 48 hours without 0.6 mM palmitate or GPI-15427 (C), with 0.6 mM palmitate and without GPI-15427 (P), and with 0.6 mM palmitate and with 20 ␮M GPI-15427 (PϩGPI). Fluorescence per cell in pericytes or endothelial cell cultured without palmitate and without GPI-15427 is taken as 100%. Mean Ϯ SD. n ϭ 4 per group. **P Ͻ 0.01 vs. cells cultured without palmitate or GPI-15427; ##P Ͻ 0.01 vs. cells cultured with palmitate and without GPI-15427.

Palmitate exposure (0.6 mM) was associated with aug- human subjects with type 2 diabetes29 point to the key role for mented cell death manifested by increased numbers of TUNEL- the first enzyme of the sorbitol pathway, aldose reductase (AR), positive cells (Figs. 12A–D) and elevated caspase activity (Figs. several AR inhibitors (ARIs) have been withdrawn from dia- 12E, 12F) in retinal pericytes and endothelial cells. PARP inhi- betic complication–related clinical trials because of low effi- bition counteracted palmitate-induced increases in TUNEL pos- cacy (carboxylic acid–derived ARIs) or adverse effects (hydan- itivity and caspase activation in both cell types. toin ARIs). Evidence of participation of nonenzymatic glycooxidation, another important mechanism in diabetic com- 30 DISCUSSION plications, in cataractogenesis in animals and humans with diabetes is controversial.25,31,32 Long-term consumption of vi- The findings described herein provide the first evidence of tamin C and vitamin E supplements reduced the development early PARP activation in the lenses of diabetic rats and high of age-related lens opacities in humans.33,34 Note, however, glucose-exposed HLE cells and of the contribution of this that the results of clinical trials of conventional antioxidants in mechanism to the formation of diabetic cataract. Multiple human subjects with diabetic complications have been incon- mechanisms have been implicated in diabetes-associated and clusive.35,36 In the present experimental study, PARP inhibi- other types of cataractogenesis,25 but no anticataract agent is tors delayed rather than prevented diabetic cataract formation available for use in humans. Although numerous findings in (i.e., efficacy was comparable to that of conventional antioxi- diabetic animal models25–28 and gene polymorphism studies in dants in other reports).25,37 Note, however, that ISO and GPI-

Downloaded from iovs.arvojournals.org on 09/26/2021 1788 Drel et al. IOVS, April 2009, Vol. 50, No. 4

FIGURE 12. (A, C) Representative microphotographs of TUNEL-positive cells in retinal pericytes (A) and endothelial cells (C) cultured for 48 hours without 0.6 mM palmitate or GPI-15427 (C), with 0.6 mM palmitate and without GPI-15427 (P), and with 0.6 mM palmitate and with 20 ␮M GPI-15427 (PϩGPI). Magniﬁcation, ϫ100. Blue ﬂuorescence corresponds to 4Ј,6-diamidino-2-phenylindole-stained nuclei. (B, D) Percentage of TUNEL-positive cells in retinal pericyte (B) and endothelial cell (D) cultured for 48 hours without 0.6 mM palmitate or GPI-15427 (C), with 0.6 mM palmitate and without GPI-15427 (P), and with 0.6 mM palmitate and with 20 ␮M GPI-15427 (PϩGPI). Mean Ϯ SD. n ϭ 4 to 8 per group. **P Ͻ 0.01 vs. cells cultured without palmitate and without GPI-15427; #P Ͻ 0.01 vs. cells cultured with palmitate and without GPI-15427. (E, F) Caspase activities in retinal pericytes (A) and endothelial cells (C) cultured for 48 hours without 0.6 mM palmitate or GPI-15427 (C), with 0.6 mM palmitate and without GPI-15427 (P), and with 0.6 mM palmitate and with 20 ␮M GPI-15427 (PϩGPI). Caspase activity in cells cultured without 0.6 mM palmitate and without GPI-15427 is taken as 100%. Mean Ϯ SD. n ϭ 3 to 6 per group. **P Ͻ 0.01 vs. cells cultured without palmitate and without GPI-15427; #P Ͻ 0.05 and ##P Ͻ 0.01 vs. cells cultured with palmitate and without GPI-15427.

15427 have been used at lower doses; therefore, dose-response nitrosative stress in the retina of STZ diabetic rats and studies are needed for full assessment of the anticataractogenic palmitate-exposed cultured retinal pericytes and endothelial potential of PARP inhibitors. Taking into consideration the cells. Until recently, PARP activation was regarded as a phe- multiple consequences of PARP activation in tissue sites for nomenon arising from free radical- and peroxynitrite-induced diabetic complications, it is probably important to control this DNA single-strand breakage.1 However, recent studies reveal enzyme activity from an early stage of diabetes to prevent the that in some tissues of diabetic animals PARP activation may development of cataract. lead to rather than result from oxidative-nitrosative stress10,21 Our ﬁndings also demonstrate that PARP inhibition coun- and that PARP activation does not necessarily require DNA teracts numerous changes characteristic of early diabetic reti- single-strand breakage and may occur because of enzyme phos- nopathy. In particular, PARP inhibition alleviated oxidative- phorylation by ERK.38 In the diabetic rat retina, poly(ADP-

Downloaded from iovs.arvojournals.org on 09/26/2021 IOVS, April 2009, Vol. 50, No. 4 PARP, Diabetic Cataract, and Early Retinal Changes 1789

ribosyl)ated proteins accumulated in cells containing DNA activation, and retinal neural and capillary cell apoptosis in breaks and in those with preserved DNA integrity.13 The latter diabetic rats. These findings, consistent with previous reports is consistent with current findings suggesting that the relations on PARP contribution to the formation of acellular capillaries between diabetes-associated oxidative-nitrosative stress and and pericyte ghosts, leukostasis, and VEGF formation, provide PARP activation in retina and retinal capillary cells are bidirec- a rationale for the development of PARP inhibitors to prevent tional rather than unidirectional. Increased formation of reac- and slow the progression of diabetic retinopathy and cataract tive oxygen and nitrogen species leads to PARP activation and formation. vice versa. PARP inhibition also counteracted diabetes-induced retinal Acknowledgments glial activation manifest in GFAP accumulation and neural retinal apoptosis. Similar effects on both phenomena have been The authors thank Usha P. Andley for providing human lens epithelial reported for two ARIs, sorbinil39 and ARI-809.27 It has been cells and valuable recommendations regarding their use. hypothesized that retinal neurodegenerative changes, includ- ing increased glial cell reactivity and microglial activation, References together with altered glutamate metabolism and premature 40 1. Jagtap P, Szabo C. Poly(ADP-ribose) polymerase and the therapeu- apoptosis, are critical components of diabetic retinopathy. tic effects of its inhibitors. Nat Rev Drug Discov. 2005;4(5):421– However, a recent study in the STZ diabetic mouse model has 440. shown that diabetes-induced degeneration of retinal capillaries 2. Schreiber V, Dantzer F, Ame JC, de Murcia G. Poly(ADP-ribose): can develop independently of neuronal loss or chronic GFAP novel functions for an old molecule. Nat Rev Mol Cell Biol. 2006; upregulation in glial cells.41 7(7):517–528. The frequency of early apoptosis in retinal capillary cells has 3. Obrosova IG, Li F, Abatan OI, et al. Role of poly(ADP-ribose) been reported to predict the development of the histologic polymerase activation in diabetic neuropathy. Diabetes. 2004; lesions of retinopathy in diabetes and galactosemia.42 In addi- 53(3):711–720. tion to high glucose,13,43 other factors in the diabetic milieu, 4. Ha HC, Hester LD, Snyder SH. Poly(ADP-ribose) polymerase-1 de- particularly fatty acids,18 cause premature apoptosis of retinal pendence of stress-induced transcription factors and associated gene expression in glia. Proc Natl Acad SciUSA.2002;99(5): pericytes and endothelial cells. In the present study, the num- 3270–3275. ber of apoptotic cells increased more than 3-fold in retinal 5. Veres B, Radnai B, Gallyas F Jr, et al. Regulation of kinase cascades pericytes and endothelial cells shortly (48 hours) after expo- and transcription factors by a poly(ADP-ribose) polymerase-1 in- sure to 0.6 mM palmitate, the concentration in the circulation hibitor, 4-hydroxyquinazoline, in lipopolysaccharide-induced in- of diabetic rodents. These findings are consistent with caspase flammation in mice. J Pharmacol Exp Ther. 2004;310(1):247–255. activation in both cell types. PARP inhibition counteracted 6. Garcia Soriano F, Virag L, Jagtap P, et al. Diabetic endothelial fatty acid–induced increases in TUNEL positivity and caspase dysfunction: the role of poly(ADP-ribose) polymerase activation. activation in retinal capillary cells. Taking into consideration Nat Med. 2001;7(1):108–113. that a PARP inhibitor treatment has previously been reported 7. Pacher P, Liaudet L, Soriano FG, Mabley JG, Szabo E, Szabo C. The to counteract high glucose-induced endothelial cell apopto- role of poly(ADP-ribose) polymerase activation in the development sis,13 it is reasonable to suggest that PARP activation expedites of myocardial and endothelial dysfunction in diabetes. Diabetes. the mechanism(s) involved in hyperglycemia- and elevated 2002;51(2):514–521. fatty acid-induced cell death. 8. Obrosova IG, Xu W, Lyzogubov VV, et al. PARP inhibition or gene deficiency counteracts intraepidermal nerve fiber loss and neuro- Recent reports suggest that the accumulation of unfolded or pathic pain in advanced diabetic neuropathy. Free Radic Biol Med. misfolded proteins that cause ER stress and the unfolded pro- 2008;44:972–981. ␤ tein response play important roles in diabetes-associated -cell 9. Gibson TM, Cotter MA, Cameron NE. Effects of poly(ADP-ribose) 44 dysfunction. The role for ER stress in diabetic complications polymerase inhibition on dysfunction of non-adrenergic non-cho- remains unexplored, though a recent study has implicated this linergic neurotransmission in gastric fundus in diabetic rats. Nitric phenomenon in lens epithelial cell apoptosis and cataract for- Oxide. 2006;15(4):344–350. mation in galactose-fed rats.45 The present study showed that 10. Szabo C, Biser A, Benko R, Bottinger E, Susztak K. Poly(ADP-ribose) modest retinal ER stress, manifested by the presence of ER- polymerase inhibitors ameliorate nephropathy of type 2 diabetic mediated chaperones BiP/GRP78 and GRP94 (two proteins Leprdb/db mice. Diabetes. 2006;55(11):3004–3012. containing ER stress response element in their promoters), is 11. Tamada Y, Fukiage C, Nakamura Y, Azuma M, Kim YH, Shearer TR. identifiable by immunohistochemistry and Western blot analy- Evidence for apoptosis in the selenite rat model of cataract. Bio- sis in nondiabetic and diabetic rats. Furthermore, a slight, but chem Biophys Res Commun. 2000;275(2):300–306. statistically significant, increase in BiP/GRP78 expression, pri- 12. Obrosova IG, Drel VR, Kumagai AK, Sza´bo C, Pacher P, Stevens MJ. Early diabetes-induced biochemical changes in the retina: compar- marily confined to inner and outer plexiform and ganglion cell ison of rat and mouse models. Diabetologia. 2006;49(10):2525– layers of diabetic rats, indicated that a weak induction of retinal 2533. ER stress was present at a very early stage of diabetes. Another 13. Zheng L, Szabo C, Kern TS. Poly(ADP-ribose) polymerase is in- ER chaperone, GRP94, showed only a trend toward an involved in the development of diabetic retinopathy via regulation of crease, suggesting that diabetes may affect a recently identified nuclear factor-␬B. Diabetes. 2004;53(11):2960–2967. BiP inducer X.46 ER stress has been implicated in oxidative 14. Sugawara R, Hikichi T, Kitaya N, et al. Peroxynitrite decomposition stress and apoptosis.47–49 Furthermore, inducible nitric oxide catalyst, FP15, and poly(ADP-ribose) polymerase inhibitor, PJ34, synthase, known to play an important role in diabetic retinop- inhibit leukocyte entrapment in the retinal microcirculation of athy,50 is involved in ER stress–mediated oxidative-nitrosative diabetic rats. Curr Eye Res. 2004;29(1):11–16. stress.49 Retinal ER stress and its relation to oxidative injury 15. Obrosova IG, Minchenko AG, Frank RN, et al. Poly(ADP-ribose) and premature apoptosis associated with early diabetic ret- polymerase inhibitors counteract diabetes- and hypoxia-induced retinal vascular endothelial growth factor overexpression. Int J inopathy deserve further study. Although PARP inhibition Mol Med. 2004;14(1):55–64. apparently reduced the induction of ER-mediated chaper- 16. Rajesh M, Mukhopadhyay P, Godlewski G, et al. Poly(ADP-ribose)- ones, the significance and mechanisms of this effect cannot polymerase inhibition decreases angiogenesis. Biochem Biophys be interpreted based on current knowledge and require Res Commun. 2006;350(4):1056–1062. specific studies. 17. Tentori L, Lacal PM, Muzi A, et al. Poly(ADP-ribose) polymerase In conclusion, PARP activation is involved in cataractogen- (PARP) inhibition or PARP-1 gene deletion reduces angiogenesis. esis, retinal oxidative-nitrosative and ER stresses, neuroglial Eur J Cancer. 2007;43(14):2124–2133.

Downloaded from iovs.arvojournals.org on 09/26/2021 1790 Drel et al. IOVS, April 2009, Vol. 50, No. 4

18. Cacicedo JM, Benjachareowong S, Chou E, Ruderman NB, Ido Y. posterior subcapsular lens opacities. Am J Clin Nutr. 2002;75(3): Palmitate-induced apoptosis in cultured bovine retinal pericytes: 540–549. roles of NAD(P)H oxidase, oxidant stress, and ceramide. Diabetes. 35. Ziegler D. Thioctic acid for patients with symptomatic diabetic 2005;54(6):1838–1845. polyneuropathy: a critical review. Treat Endocrinol. 2004;3(3): 19. Barber AJ, Lieth E, Khin SA, Antonetti DA, Buchanan AG, Gardner 173–189. TW. Neural apoptosis in the retina during experimental and hu- 36. Economides PA, Khaodhiar L, Caselli A, et al. The effect of vitamin man diabetes: early onset and effect of insulin. J Clin Invest. E on endothelial function of micro- and macrocirculation and left 1998;102(4):783–791. ventricular function in type 1 and type 2 diabetic patients. Diabe- 20. El-Remessy AB, Al-Shabrawey M, Khalifa Y, Tsai NT, Caldwell RB, tes. 2005;54(1):204–211. Liou GI. Neuroprotective and blood-retinal barrier-preserving ef- 37. Kilic F, Mitton K, Dzialoszynski T, Sanford SE, Trevithick JR. fects of cannabidiol in experimental diabetes. Am J Pathol. 2006; Modelling cortical cataractogenesis, 14: reduction in lens damage 168(1):235–244. in diabetic rats by a dietary regimen combining vitamins C and E 21. Obrosova IG, Drel VR, Pacher P, et al. Oxidative-nitrosative stress and beta-carotene. Dev Ophthalmol. 1994;26:63–71. and poly(ADP-ribose) polymerase (PARP) activation in experimental diabetic neuropathy: the relation is revisited. Diabetes. 2005; 38. Kauppinen TM, Chan WY, Suh SW, Wiggins AK, Huang EJ, Swan- 54(12):3435–3441. son RA. Direct phosphorylation and regulation of poly(ADP-ribose) polymerase-1 by extracellular signal-regulated kinases 1/2. 22. Podesta` F, Romeo G, Liu WH, et al. Bax is increased in the retina of diabetic subjects and is associated with pericyte apoptosis in Proc Natl Acad SciUSA.2006;103(18):7136–7141. vivo and in vitro. Am J Pathol. 2000;156(3):1025–1032. 39. Asnaghi V, Gerhardinger C, Hoehn T, Adeboje A, Lorenzi M. A role 23. Sone H, Deo BK, Kumagai AK. Enhancement of glucose transport for the polyol pathway in the early neuroretinal apoptosis and glial by vascular endothelial growth factor in retinal endothelial cells. changes induced by diabetes in the rat. Diabetes. 2003;52(2):506– Invest Ophthalmol Vis Sci. 2000;41(7):1876–1884. 511. 24. Gao Z, Zhang X, Zuberi A, et al. Inhibition of insulin sensitivity by 40. Barber AJ. A new view of diabetic retinopathy: a neurodegenera- free fatty acids requires activation of multiple serine kinases in tive disease of the eye. Prog Neuropsychopharmacol Biol Psychi- 3T3–L1 adipocytes. Mol Endocrinol. 2004;18(8):2024–2034. atry. 2003;27(2):283–290. 25. Kador PF. Ocular pathology of diabetes mellitus. In: Tasman W, 41. Feit-Leichman RA, Kinouchi R, Takeda M, et al. Vascular damage in Jaeger EA, eds. Duane’s Ophthalmology. Vol. 3. Philadelphia: a mouse model of diabetic retinopathy: relation to neuronal and Lippicott Williams & Wilkins; 2007:1–84. glial changes. Invest Ophthalmol Vis Sci. 2005;46(11):4281–4287. 26. Lee AY, Chung SK, Chung SS. Demonstration that polyol accumu- 42. Kern TS, Tang J, Mizutani M, et al. Response of capillary cell death lation is responsible for diabetic cataract by the use of transgenic to aminoguanidine predicts the development of retinopathy: com- mice expressing the aldose reductase gene in the lens. Proc Natl parison of diabetes and galactosemia. Invest Ophthalmol Vis Sci. Acad SciUSA.1995;92(7):2780–2784. 2000;41(12):3972–3978. 27. Sun W, Oates PJ, Coutcher JB, Gerhardinger C, Lorenzi M. A 43. el-Remessy AB, Bartoli M, Platt DH, Fulton D, Caldwell RB. Oxida- selective aldose reductase inhibitor of a new structural class pre- tive stress inactivates VEGF survival signaling in retinal endothelial vents or reverses early retinal abnormalities in experimental dia- cells via PI 3-kinase tyrosine nitration. J Cell Sci. 2005;118(pt betic retinopathy. Diabetes. 2006;55(10):2757–2762. 1):243–252. 28. Drel VR, Pacher P, Ali TK, et al. Aldose reductase inhibitor ﬁdares- 44. Eizirik DL, Cardozo AK, Cnop M. The role for endoplasmic retic- tat counteractes diabetes-associated cataract formation, retinal ox- ulum stress in diabetes mellitus. Endocr Rev. 2008;29(1):42–61. idative-nitrosative stress, glial activation, and apoptosis. Int J Mol 45. Mulhern ML, Madson CJ, Kador PF, Randazzo J, Shinohara T. Med. 2008;21:667–676. Cellular osmolytes reduce lens epithelial cell death and alleviate 29. Lee SC, Wang Y, Ko GT, et al. Risk factors for cataract in Chinese cataract formation in galactosemic rats. Mol Vis. 2007;13:1397– patients with type 2 diabetes: evidence for the inﬂuence of the 1405. aldose reductase gene. Clin Genet. 2001;59(5):356–359. 46. Kudo T, Kanemoto S, Hara H, et al. A molecular chaperone inducer 30. Goh SY, Cooper ME. Clinical review: the role of advanced glyca- protects neurons from ER stress. Cell Death Differ. 2008;15(2): tion end products in progression and complications of diabetes. J Clin Endocrinol Metab. 2008;93(4):1143–1152. 364–375. 31. Swamy-Mruthinti S, Green K, Abraham EC. Inhibition of cataracts 47. Szegezdi E, Logue SE, Gorman AM, Samali A. Mediators of endo- in moderately diabetic rats by aminoguanidine. Exp Eye Res. 1996; plasmic reticulum stress-induced apoptosis. EMBO Rep. 2006;7(9): 62(5):505–510. 880–885. 32. Kador PF, Lee JW, Fujisawa S, Blessing K, Lou MF. Relative impor- 48. Boyce M, Yuan J. Cellular response to endoplasmic reticulum tance of aldose reductase versus nonenzymatic glycosylation on stress: a matter of life or death. Cell Death Differ. 2006;13(3):363– sugar cataract formation in diabetic rats. J Ocul Pharmacol Ther. 373. 2000;16(2):149–160. 49. Hsieh YH, Su IJ, Lei HY, Lai MD, Chang WW, Huang W. Differential 33. Jacques PF, Taylor A, Hankinson SE, et al. Long-term vitamin C endoplasmic reticulum stress signaling pathways mediated by supplement use and prevalence of early age-related lens opacities. iNOS. Biochem Biophys Res Commun. 2007;359(3):643–648. Am J Clin Nutr. 1997;66:911–916. 50. Zheng L, Du Y, Miller C, et al. Critical role of inducible nitric oxide 34. Taylor A, Jacques PF, Chylack LT Jr, et al. Long-term intake of synthase in degeneration of retinal capillaries in mice with strep- vitamins and carotenoids and odds of early age-related cortical and tozotocin-induced diabetes. Diabetologia. 2007;50(9):1987–1996.

Downloaded from iovs.arvojournals.org on 09/26/2021