Archives of Biochemistry and Biophysics 558 (2014) 1–9

Contents lists available at ScienceDirect

Archives of Biochemistry and Biophysics journal homepage: www.elsevier.com/locate/yabbi

Expression and induction of small heat shock proteins in rat heart under chronic hyperglycemic conditions V. Sudhakar Reddy, Ch. Uday Kumar, G. Raghu, G. Bhanuprakash Reddy ⇑ Biochemistry Division, National Institute of Nutrition, Hyderabad, India

a r t i c l e

i n f o

Article history: Received 13 February 2014 and in revised form 2 June 2014 Available online 17 June 2014 Keywords: sHsp Diabetes Diabetic cardiomyopathy Streptozotocin Phosphorylation p38MAPK p44/42MAPK Apoptosis Rat

a b s t r a c t The induction of small heat shock proteins (sHsp) is observed under various stress conditions to protect the cells and organisms from adverse events including diabetes. Diabetic cardiomyopathy is a common complication of diabetes. Therefore, in this study, we investigated the expression of sHsp under chronic hyperglycemic conditions in rat heart. Hyperglycemia was induced in WNIN rats by intraperitoneal injection of streptozotocin and maintained for a period of 12 weeks. Expression of sHsp, phosphorylation and translocation of phosphoforms of Hsp27 and aB-crystallin (aBC) from cytosolic fraction to cytoskeletal fraction was analyzed. While the expression of MKBP, HspB3, aBC was found to be increased in diabetic heart, expression of Hsp20 was decreased. Chronic hyperglycemia further induced phosphorylation of aBC at S59, S45, Hsp27 at S82, p38MAPK and p44/42MAPK. However, pS59-aBC and pS82-Hsp27 were translocated from detergent-soluble to detergent-insoluble fraction under hyperglycemic conditions. Furthermore, the interaction of pS82-Hsp27 and pS59-aBC with desmin was increased under hyperglycemia. However, the interaction of aBC and pS59-aBC with Bax was impaired by chronic hyperglycemia. These results suggest up regulation of sHsp (MKBP, HspB3 and aBC), phosphorylation and translocation of Hsp27 and aBC to striated sarcomeres and impaired interaction of aBC and pS59-aBC with Bax under chronic hyperglycemia. Ó 2014 Elsevier Inc. All rights reserved.

Diabetes has become a serious public health problem. The number of diabetic patients was 382 million in 2013 and is expected to reach 592 million in 2030 [1]. Diabetic cardiomyopathy is a common complication of diabetes mellitus and is one of the most common causes of morbidity and mortality in diabetic patients. Diabetic cardiomyopathy is a major risk factor for developing myocardial dysfunction in diabetic patients in the absence of hypertension and coronary heart disease [2–5]. Diabetes is associated with disturbed myofibrils, severe alterations in sarcomere microstructure and components of dystrophin associated protein complex [6]. However, the pathophysiological insults for the development of diabetic cardiomyopathy are poorly understood. Though the cause of diabetic cardiomyopathy is not fully understood, oxidative stress, cardiac inflammation, lipid accumulation, cardiac fibrosis and apoptosis are considered to be the major mechanisms implicated in diabetic cardiomyopathy [2,7]. Of these, oxidative stress induced by excessive production of reactive oxygen species (ROS)

and reactive nitrogen species (RNS)1 resulting from hyperglycemia causes cardiac fibrosis [8], altered signaling pathways, altered gene expressions [9] and myocardial cell death [10,11]. Heat shock proteins (Hsp) are a group of proteins that accumulate in the cells after a variety of physiological, environmental and pathological stresses. Small Hsp (sHsp) are proteins with monomeric molecular mass ranging from 15 to 30 kDa and with a conserved a-crystallin domain. Mammals contain 10 sHsp: Hsp27/ HSPB1, myotonic dystrophy kinase binding protein (MKBP)/HSPB2, HSPB3, aA-crystallin (aAC)/HSPB4, aB-crystallin (aBC)/HSPB5, Hsp20/HSPB6, cvHsp/HSPB7, Hsp22/H11/H2IG1/HSPB8, HSPB9 and sperm outer dense fiber protein (ODF)/HSPB10 [12]. Small Hsp acts as molecular chaperones by preventing aggregation or misfolding of proteins and allow their correct refolding under stress conditions [13–15]. These proteins are also involved in several fundamental cellular processes like cytoskeletal architecture, intracellular transport of proteins and protection against programmed cell death [15]. The heat shock response is mediated by a group of heat shock transcription factors (HSF). The mammals

⇑ Corresponding author. Address: National Institute of Nutrition, Hyderabad 500 007, India. Fax: +91 40 27019074. E-mail address: [email protected] (G.B. Reddy).

1 Abbreviations used: ROS, reactive oxygen species; RNS, reactive nitrogen species; TdT, terminal deoxynucleotidyl transferase; HRP, horse radish peroxidase; Hsp, heat shock proteins.

Introduction

http://dx.doi.org/10.1016/j.abb.2014.06.008 0003-9861/Ó 2014 Elsevier Inc. All rights reserved.

2

V.S. Reddy et al. / Archives of Biochemistry and Biophysics 558 (2014) 1–9

contain 3 different HSFs known as HSF1, HSF2 and HSF4. HSF1 is a major HSF, which mediates the regulation of several heat shock genes while HSF2 is involved in the differentiation and developmental processes. HSF4 is involved in the postnatal expression of Hsp [16]. The phosphorylation status of sHsp is important for determining their chaperone activity and cytoprotective functions [17,18]. Hsp27 is phosphorylated at serine positions 82 (S82), 72 (S72) and 15 (S15) while aBC at 59 (S59), 45 (S45) and 19 (S19). The p38 mitogen activated protein kinase (MAPK)/MAPK activated protein-2 is responsible for phosphorylation at S82, S72, S15 of Hsp27 and S59 of aBC [19,20]. While S45 of aBC is phosphorylated by p44/42 MAP kinase (ERK), kinase responsible for phosphorylation of S19 of aBC is unknown. Previously we have reported the elevated expression of a-crystallins, two prominent members of sHsp, in various tissues including heart in diabetic rats [21]. Recently we have also observed induction of some members of sHsp family and their phosphoregulation in retina [22] and lens [23] of diabetic rat model. However, the effect of chronic hyperglycemia on expression of sHsp family members, kinase mediated phosphoregulation and their involvement in cytoskeletal protection and apoptosis in experimental diabetic heart has not been examined. In the present study, for the first time, we investigated the response of sHsp family members in chronic hyperglycemia and their translocation from cytosol to striated sarcomeres in cardiac myofibers and role in apoptotic cell death in diabetic rat heart. Materials and methods Materials Streptozotocin (STZ), Tri-reagent, TritonX-100 (TritonX), acrylamide, bis-acrylamide, ammonium persulphate, b-mercaptoethanol, SDS, TEMED, PMSF, aprotinin, leupeptin, pepstatin, anti-actin antibody (Cat.No-A5060), horse radish peroxidase (HRP) conjugated anti-rabbit (A6154) and anti-mouse (A9044) secondary antibodies were purchased from Sigma Chemicals (St. Louis, MO, USA). Nitrocellulose membrane was obtained from Pall Corporation (Pensacola, FL, USA). Anti-Hsp27 (MA3-014), antiHsp20 (PA1-29447), anti-HSF1 (PA3-017), and specific antibodies recognizing three phosphorylated residues S59 (PA1-012), S45 (PA1-011), and S19 (PA1-010) of aBC were obtained from Thermo Scientific-Pierce (Rockford, IL, USA). Anti-aBC antibody was produced in the rabbit as reported previously [24]. Anti-MKBP (18821-485217), anti-Hsp22 (18-821-485201) were obtained from Genway (San Diego, CA, USA). Anti-p38MAPK (#9212S), anti-pp38MAPK (#9211S), anti-p44/42 MAPK (#4695), anti-p-p44/42 MAPK (#4370), anti-cleaved caspase-3 (#9661S), anti-pS82Hsp27 (#2401S), anti-Hsp70 (#4872) antibodies were purchased from Cell Signaling Technology, Inc. (Beverly, MA). Anti-HspB3 (sc-104935), anti-HSF2 (sc-13056), anti-HSF4 (sc-19864), antiBax (sc-6236), anti-Bcl-2 (sc-492), anti-desmin (sc-23879) and HRP conjugated anti-goat (sc-2020) secondary antibodies were purchased from Santa Cruz Biotechnology, Inc. (Dallas, Texas, USA). All primers were procured from Integrated DNA Technologies (Coralville, IA, USA). Alexafluor-488 conjugated anti-rabbit (A-11008) and Alexafluor-555 conjugated anti-mouse (A-21427) antibodies were obtained from Molecular Probes, Inc. (Eugene, OR, USA). Animal care and experimental conditions Three-month-old male WNIN (Wistar-NIN) rats with average body weight of 230 ± 14 g were obtained from National Center for Laboratory Animal Sciences, National Institute of Nutrition,

Hyderabad, India, and maintained at a temperature of 22 ± 2 °C, 50% humidity, and 12-h light/dark cycle as described previously and the heart tissues of the same animals were used [22]. The control rats (n = 10) received 0.1 M sodium citrate buffer, pH 4.5, as a vehicle, whereas the experimental rats received a single intraperitoneal injection of STZ (35 mg/kg bw) in the same buffer. At 72 h after STZ injection, fasting blood glucose levels were monitored and animals with blood glucose levels >150 mg/dL were considered for the experiment (n = 10). Control and diabetic animals were fed with AIN-93 diet ad libitum. Body weight and blood glucose concentration of each animal were measured weekly. At the end of 12 weeks, rats were fasted overnight and sacrificed by CO2 asphyxiation. Institutional and national guidelines for the care and use of animals were followed, and all experimental procedures involving animals were approved by the Institutional Animal Ethical Committee (IAEC) of the National Institute of Nutrition. Biochemical estimations Glucose and glycosylated hemoglobin (HbA1c) in blood were measured by the glucose oxidase -peroxidase (GOD–POD) method and ion-exchange resin, respectively, using commercially available kits (Biosystems, Barcelona, Spain). Quantitative real-time PCR (qRT-PCR) Total RNA was extracted from control and diabetic rat heart using Tri-reagent according to the manufacturer instructions. Isolated RNA was further purified by RNeasy Mini Kit (Qiagen, USA) and quantified by measuring the absorbance at 260 and 280 nm on ND1000 Spectrophotometer (NanoDrop Technologies, Delaware, USA). Two to 4 lg of total RNA was reverse transcribed using High Capacity cDNA Reverse Transcription kit (Applied Biosystems, Warrington, UK). Reverse transcription reaction was carried out using thermocylcer (ABI-9700) and the reaction conditions were as follows: initial temperature for 10 min at 25 °C, followed by 37 °C for 120 min and inactivation of reverse transcriptase at 84 °C for 5 min. Real-time PCR (ABI-7500) was performed in triplicates with 25 ng cDNA templates using SYBR green master mix (Applied Biosystems, Warrington, UK) with gene specific primers (Table. 1). Normalization and validation of data were carried using b-actin as an internal control and data were compared between control and diabetic samples according to comparative threshold cycle (2DDct) method as reported previously [22,25]. Whole tissue lysate preparation Heart tissue (100–200 mg) was homogenized in TNE buffer (pH 7.5) containing 20 mM Tris, 100 mM NaCl, 1 mM EDTA, 1 mM DTT and protease inhibitors. Homogenization was performed on ice using a glass homogenizer and the homogenate was centrifuged at 14,000g at 4 °C for 20 min. The protein concentrations were measured by Bradford reagent (Bio-Rad, Hercules, CA). SDS–PAGE and immunoblotting Equal amounts of protein were subjected to 12% SDS–PAGE and transferred to nitrocellulose membranes (0.22 lm pore size) by western blot transfer system (Bio-Rad, USA) at a voltage of 40 V for 2 h. Nonspecific binding was blocked with 5% nonfat dry milk powder in PBST (20 mM phosphate buffer; pH 7.2, 137 mM NaCl, 0.1% Tween 20) and incubated overnight at 4 °C with monoclonal anti-Hsp27 (mouse, 1:500), anti-p44/42 MAPK (rabbit, 1:1000), anti-p-p44/42 MAPK (rabbit, 1:2000), polyclonal anti-pS82Hsp27 (1:1000), anti-aBC (rabbit, 1:3000), anti-pS59, pS45, pS19aBC (rabbit, 1:2000), anti-HspB3 (goat; 1:1000), anti-Hsp20

V.S. Reddy et al. / Archives of Biochemistry and Biophysics 558 (2014) 1–9 Table 1 List of primers and their sequence used in the study. Gene

Sequence

b-Actin

50 50 50 50 50 50 50 50 50 50 50 50 50 50 50 50 50 50 50 50 50 50 50 50 50 50 50 50

Hsp27 MKBP HSPB3

aAC aBC Hsp20 cvHsp Hsp22 HSPB9 ODF HSF1 HSF2 HSF4

GAG AAG AGC TAT GAG CTG CC 30 CTC AGG AGG AGC AAT GAT CT 30 GAT GAA CAT GGC TAC ATC TCT C 30 CTG ATT GTG TGA CTG CTT TG 30 ACT GTG GAC AAC CTG CTA GA 30 GGA GAT GTA GAC CTC ATT GAC T 30 TCT TGC AGA GGA CTC AGA CT 30 GTT TGT ACT GTC TGG TGA AAC TC 30 GAA AGA AGA TAT TTA TGC AGT GG 30 CCT CAA AGA TAT TGG ATA TGG TA 30 TGG AGT CTG ACC TCT TTT CTA C 30 AGA ACC TTG ACT TTG AGT TCC 30 AGA GGA AAT CTC TGT CAA GGT 30 GGA TAG ACA GAA CAC CCT 30 AGT TTA CTG TGG ACA TGA GAG AC 30 CTG GAC ATG TTC TGT GTG TG 30 AGA ACT GAT GGT AAA GAC CAA G 30 AGAAAG TGA GGC AAA TAC AGTC 30 GAA CCA AGT TTC CAG ATG AA 30 GAG AGG TAG GCA CTT ATT TTG TC 30 GGA CAG AGA ACT AAG ACA ATT GAG 30 CAG TAC AGC TTG TAG TCA CAC AG 30 CCA GCA GCA AAA AGT TGT CA 30 TGG TGA ACA CAG CAT CAG AGG AG 30 TGA TCC CTC CAG CCA GTA TC 30 CAG GTT GGA GGA GCC ATT TA 30 GTG GCC TGC TAA GAC CAG AC30 CGG TTG GCC TTA GGG TTC AGG GGG G 30

(rabbit, 1:10,000), anti-Hsp22 (rabbit, 1:1000), anti-p38MAPK (rabbit, 1:1000), anti-p-p38MAPK (rabbit, 1:1000), anti-HSF1 (rabbit, 1:10,000), anti-HSF2 (rabbit, 1:500), anti-HSF4 (goat, 1:500), anti-actin (rabbit, 1:500), anti-desmin (mouse, 1: 100), anti-Bcl-2 (rabbit, 1:100), anti-Bax (rabbit, 1:100), anti-cleaved caspase-3 (rabbit, 1:1000) antibodies diluted in PBST. After washing with PBST, membranes were incubated with anti-rabbit IgG (1:3500) or anti-mouse IgG (1:3500) or anti-goat IgG (1: 10,000) secondary antibodies conjugated to HRP. The immunoblots were developed using ECL detection kit (RPN2232, GE Health Care, Buckinghamshire, UK) by Image analyzer (G-Box iChemi XR, Syngene, UK) and images were analyzed and quantitated using image J software (available in the public domain at http:// rsbweb.nih.gov/ij/). Detergent soluble assay For analyzing detergent solubility, heart tissue was homogenized in TNE buffer containing 0.5% TritonX. Homogenate was centrifuged at 14,000g at 4 °C for 20 min. Following centrifugation, homogenate was separated as supernatant containing detergent soluble fraction and the pellet containing insoluble protein fraction. The pellet was washed with PBS, rehomogenized, sonicated and dissolved in Lammelli buffer and these samples were then analyzed by immunoblotting as described above.

3

blocking solution (3% horse serum, 3% BSA, 0.3% TritonX) in PBS. Slides were washed 3 times with PBS and incubated with antipS82-Hsp27 (1:25), anti-pS59-aBC (rabbit, 1:100) antibodies in PBS overnight at 4 °C. Sections were washed 3 times with PBS and the binding of primary antibodies was visualized by Alexafluor-488 conjugated anti-rabbit (1:1000) IgG antibody for 1 h. Sections were mounted in medium containing 4, 6-diamidino2-phenylindole (DAPI; #1500, Vector Laboratories, Burlingame, CA, USA) and visualized using a Leica laser microscope (LMD6000, Leica microsystems, Germany).

TUNEL assay For TUNEL assay, an in situ Cell death detection kit (11684817910; Roche Diagnostics GmBH, Mannheim, Germany) was used according to the manufacturer’s instructions. Briefly, the sections were deparafinized and dehydrated using xylene and ethanol gradings and permeablized using hot 0.1 M citrate buffer pH 6.0 and incubated with the TUNEL reaction mixture containing TdT and fluorescein labeled dUTP for 1 h at 37 °C. Sections were mounted in medium containing DAPI and images were captured with a Leica laser microscope. For negative control, TdT was not included in the reaction mixture.

Co-immunoprecipitation (Co-IP) Co-IP was carried out using Co-IP kit (26149PR; Thermo Scientific-Pierce, Rockford, IL, USA) according to the manufacturer instructions. Briefly, heart was homogenized in TNE buffer containing 0.5% TritonX and homogenate was centrifuged at 14,000g at 4 °C for 20 min. Following centrifugation, homogenate was separated as supernatant containing detergent-soluble fraction and the pellet containing insoluble protein fraction. The pellet was washed with PBS and solubilized in immunoprecipitation (IP) lysis buffer (25 mM Tris, 150 mM NaCl, 1 mM EDTA, 1% NP-40, 1% SDS, 0.5% TritonX, 5% glycerol; pH 7.4). Anti-desmin (20 lg) or anti-aBC (20 lg) or anti-pS59-aBC (20 lg) antibodies were immobilized onto spin columns containing coupling resin. One mg of protein was precleared using 80 ll control agarose resin slurry and centrifuged at 1000g for 1 min. The precleared protein sample was added to spin column and incubated at 4 °C overnight with gentle mixing. The spin column was centrifuged and washed 3 times with 200 ll IP lysis/wash buffer and eluted with elution buffer. The samples were loaded onto SDS–PAGE followed by immunoblotting using anti-pS82-Hsp27, anti-aBC, pS59-aBC, anti-Bax, and anti-desmin antibody.

Statistical analysis The data are expressed as the mean ± SE. Statistical significance between control and diabetic groups were determined by the Student’s t test. Values of p < 0.05 were considered significant.

Morphological analysis and immunohistochemistry The harvested tissue was immediately placed in 4% paraformaldehyde in phosphate buffer (pH-7.2), fixed overnight, embedded in paraffin blocks, and cut into 4 lm sections and used for Haemotoxylin and Eosin (H&E) staining, Masson’s Trichome staining, immunostaining with specific antibodies, or terminal deoxynucleotidyl transferase (TdT) dUTP nick-end labeling (TUNEL) assay. Sections for immunostaining were deparafinized by incubating in xylene for 5 min followed by dehydration in decreasing grades of ethanol. Deparafinized sections were boiled in 0.01 M Na-citrate pH 6.0 for 10 min at 60 °C and blocked with

Results Biochemical and physiological parameters The mean blood glucose levels, body weights, food intake and HbA1c levels were previously reported [22]. While the body weights of diabetic animals significantly decreased (178 ± 2.87 g vs 334.1 ± 8.07 g), the mean blood glucose levels of diabetic rats were significantly higher (390.1 ± 48.7 mg/dL vs 91.5 ± 6.17 mg/ dL) when compared with control rats.

4

V.S. Reddy et al. / Archives of Biochemistry and Biophysics 558 (2014) 1–9

Histology Hematoxylin and eosin (H&E) staining of the diabetic heart showed structural abnormalities such as perinuclear vacuolization, increase in interstitial space between cardiac myofibrils and disarrayed myofibrils when compared with controls (Fig. 1A) whereas Masson’s Trichome staining showed increased interstitial fibrosis in diabetic heart when compared with controls (Fig. 1B). Altered expression of sHsp To determine the response of sHsp in diabetic rat heart, we analyzed the expression pattern of all sHsp and HSF in heart by qRTPCR and immunoblotting. Out of 10 sHsp analyzed, we could detect Hsp27, MKBP, HspB3, aBC, Hsp20, cvHsp and Hsp22 (Figs. 2A and 3A). The MKBP levels were significantly increased both at mRNA (2.0 ± 0.35-fold, p < 0.05) and protein level (921 ± 180%, p < 0.05) under chronic hyperglycemic conditions (Figs. 2A and 3A). The HspB3 levels were also significantly increased at mRNA (1.94 ± 0.134 fold, p < 0.05; Fig. 2A) and protein level (156 ± 21.8%, p < 0.01; Fig. 3A) when compared with control. While aBC levels were also significantly increased both at mRNA (4.2 ± 0.23-fold, p < 0.01; Fig. 2A) and protein level (224 ± 23%, p < 0.01; Fig. 3A), Hsp20 levels were significantly decreased at mRNA (0.41 ± 0.06-fold, p < 0.05) and protein (29.81 ± 0.23%, p < 0.01) level in diabetic rats (Figs. 2A and 3A). However, Hsp27, cvHsp and Hsp22 levels were unaltered in diabetic rats when compared with controls (Figs. 2A and 3A). Under stress conditions, sHsp act synergistically with the stress-induced Hsp70 in protecting cells. Hence, we have also analyzed the expression of Hsp70. Interestingly; the Hsp70 levels (23.9 ± 10.31%, p < 0.05) were significantly decreased under chronic hyperglycemic conditions (Fig. 3A).

Fig. 2. qRT-PCR analysis of sHsp (Panel A) and HSFs (Panel B) in heart of control and diabetic rats. Expression of sHsp and HSFs in heart was analyzed by qRT-PCR. Relative expression pattern was analysed by comparative threshold cycle (2DDct) method. Data were represented as fold change over control on an arbitrary scale after normalization with actin and represent mean ± SEM of three independent experiments (⁄p < 0.05; ⁄⁄p < 0.01). Chronic hyperglycemia significantly increased the expression of MKBP, HSPB3 and aBC and decreased Hsp20 transcript levels.

Decreased expression of HSF Since chronic hyperglycemia led to differential expression of some sHsp (MKBP, HspB3, aBC), the expression of regulatory proteins of heat shock response including HSF1, HSF2 and HSF4 were examined by quantitative RT-PCR and immunoblotting (Figs. 2B and 3B). Surprisingly, the expression levels of HSF1 were significantly decreased both at mRNA (0.43 ± 0.109 fold, p < 0.05) and protein level (5.21 ± 1.31%, p < 0.001) whereas no changes in the HSF2 at mRNA and protein levels were observed when compared with controls (Figs. 2B and 3B). However, HSF4 protein was not detected in control and diabetic rats (Fig. 3B). Kinase mediated phosphoregulation of Hsp27 and aBC

Fig. 1. Representative histology of control and diabetic rat heart. Panel A: Tissue sections of heart were stained with H&E. Arrow ‘a’ indicates perinuclear vacuolization; arrow ‘b’ indicates increase in interstitial space between cardiac myofibers. Panel B: Tissue sections of heart were stained with Masson’s Trichome. Arrow ‘c’ indicates fibrosis area stained in blue. All the pictures were taken at 63 magnification. Scale bars, 50 lm. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

The phosphorylation status of sHsp represents an important factor in determination of various cellular functions of sHsp including chaperone activity. Most sHsp have multiple phosphorylable serine residues. However, we mainly focused on the phosphorylation on S82 of Hsp27 and S59, S45, S19 of aBC by immunoblotting. Chronic hyperglycemia significantly decreased the pS82-Hsp27 (10.19 ± 1.36%, p < 0.001) and pS59-aBC (19.68 ± 5.57%, p < 0.01) levels whereas increased the pS45-aBC levels (186.3 ± 18.47%, p < 0.05) but the pS19-aBC levels remained unaltered (Fig. 4A). While the S82 of Hsp27 and S59 of aBC are phosphorylated by p38MAPK, S45 of aBC is phosphorylated by p44/42 MAPK. Therefore, we assessed the activation of p38MAPK and p44/42 MAPK

V.S. Reddy et al. / Archives of Biochemistry and Biophysics 558 (2014) 1–9

5

Fig. 3. Immunoblot analysis of sHsp and HSFs in heart of control and diabetic rats. Panel A: Representative immunoblots of sHsp and respective quantitative bars demonstrating the sHsp levels. Panel B: Representative immunoblots of HSF and respective quantitative bars demonstrating the HSF levels. Data were normalized for actin expression and represented as percent of control. Data represent mean ± SEM of three independent experiments (⁄p < 0.05; ⁄⁄p < 0.01). Chronic hyperglycemia significantly increased the expression of MKBP, HSPB3 and aBC and decreased the levels of Hsp20, Hsp70 and HSF1.

Fig. 4. Kinase mediated phosphoregulation of Hsp27 and aBC in heart of control and diabetic rats. Panel A: Representative immunoblots of phosphorylated Hsp27, aBC and respective quantitative bars demonstrating the phosphorylated Hsp27, aBC levels in heart of control and diabetic rats. Panel B: Representative immunoblots of p38MAPK, pp38MAPK p44/42MAPK, p-p44/42MAPK and respective quantitative bars demonstrating the p38MAPK, p-p38MAPK, p44/42MAPK, p-p44/42MAPK levels in heart of control and diabetic rats. Data were normalized for actin expression and represented as percent of control. Data represent mean ± SEM of three independent experiments (⁄p < 0.05; ⁄⁄ p < 0.01). Chronic hyperglycemia significantly decreased pS59-aBC, pS82-Hsp27 whereas increased the pS45-aBC levels and induced the activation of p38MAPK and p44/ 42MAPK pathway by increasing the levels of p-p38MAPK and p-p44/42MAPK.

6

V.S. Reddy et al. / Archives of Biochemistry and Biophysics 558 (2014) 1–9

by immunoblotting. Though, chronic hyperglycemia did not affect the levels of p38MAPK and p44/42 MAPK it significantly increased the levels of p-p38MAPK (788 ± 135%, p < 0.05), p-p44 (652 ± 108%, p < 0.05) and p-p42 MAPK (525 ± 97%; p < 0.05) (Fig. 4B). Translocation of pS82-Hsp27 and pS59-aBC from soluble to insoluble fractions Although, under hyperglycemia, levels of pS82-Hsp27 and pS59-aBC were found to be decreased in soluble fraction, they were translocated from detergent-soluble (cytosol) to detergentinsoluble (cytoskeletal) fraction (Fig. 5A). Hence, the total (detergent-soluble and detergent-insoluble) pS82-Hsp27 and pS59-aBC levels were significantly high in diabetic heart. Levels of desmin were increased in detergent-soluble fraction and also translocated to detergent-insoluble fraction (Fig. 5A). Moreover, co-immunoprecipitation studies have also shown the strong interaction of pS82-Hsp27 and pS59-aBC with desmin (Fig. 5B). We have also investigated the translocation of phosphorylated sHsp from cytosol to striated sarcomeres by immunofluorescence. The intense staining of pS82-Hsp27 and pS59-aBC at striated sarcomeres in diabetic heart when compared to control indicates the translocation from cytosol to striated sarcomeres under hyperglycemic conditions (Fig. 6A and B). Chronic hyperglycemia impairs aBC interaction with proapoptotic protein Bax We analysed the apoptotic cell death under chronic hyperglycemic conditions in heart. The TUNEL-positive cells were significantly higher in diabetic heart (82.66 ± 6.48%, p < 0.001; Fig. 7A) when compared with control. Further, the expression of important regulators of apoptosis such as Bax, cleaved caspase-3 and Bcl-2 was assessed by immunoblotting. Chronic hyperglycemia significantly increased the expression of Bax (264 ± 26.5%, p < 0.05) and cleaved caspase-3 (651.3 ± 126%, p < 0.05) whereas decreased the Bcl-2 levels (65 ± 8.07%, p < 0.05) when compared with the control (Fig. 7B). Interestingly, the interaction of aBC and pS59-aBC with Bax was reduced in diabetic heart when compared with controls (Fig. 7C and D). Discussion This study reported the following novel findings: (i) chronic hyperglycemia increased the expression of MKBP, HspB3, aBC while decreased the Hsp20, Hsp70, HSF1 (ii) hyperglycemia

induced the phosphorylation of Hsp27 on S82 and aBC on S59 and S45, (iii) hyperglycemia induced the translocation of pS82Hsp27 and pS59-aBC and (iv) hyperglycemia induced the cell death and reduced the interaction of sHsp (aBC) and pS59-aBC with proapoptotic protein Bax. Studies indicate that incidence of mortality from cardiac diseases is two to four fold greater in diabetic patients than the patients suffering with cardiac diseases without diabetes. Chronic hyperglycemia is an independent risk factor that causes myocardial dysfunction. Few previous studies including ours reported the expression of single sHsp either Hsp27 or aBC in heart under hyperglycemic conditions [21,26]. However, this is the first study that report expression of all members of sHsp family, HSFs, phosphorylation status of sHsp and their translocation from cytosol to cytoskeleton, involvement of aBC and its pS59 in apoptosis under chronic hyperglycemic conditions in STZ induced diabetic rat model. The increased fibrosis, disarrayed myofibrils, perinuclear vacuolization in 12 week diabetic heart indicates the occurrence of cardiomyopathy-like features in diabetic rats. MKBP and HspB3 were found to be the smallest members of sHsp family and were expressed mainly in heart and skeletal muscle [27] and they assemble as a tetramer with subunit ratio of 3:1 [28]. Some of the previous studies have shown the upregulation of MKBP in Myotonic Dystrophy (DM) patients [29] and also transiently during development in neonatal myocardium [30]. The increased expression of aBC is consistent with the previous studies in STZ induced diabetic rat [21] and in high fat fed pre-diabetic mice [31]. However, oxidative stress also might be triggering the increased expression of MKBP, HSPB3 and aBC in diabetic heart. Several past studies have reported no change in the expression level of Hsp20 in high fat fed insulin resistance rat model and db/db mice but ischemic post conditioning (IPC) in db/db mice resulted in decreased expression of Hsp20 [32]. However, the reason for decreased expression of Hsp20 under chronic hyperglycemia is unknown. The decreased or unaltered expression of HSF1 and HSF2 in diabetic heart, however, is not in concurrence with expression of MKBP, HspB3 and aBC but with that Hsp20 and Hsp70 expression. Previous studies reported that the upregulation of Hsp60 was accompanied by downregulation of HSF1 in rat myocardial cells [33]. However, it should be noted that glycogen synthase kinase 3 (GSK3) and MAPKAP Kinase 2 which are potent inhibitors of HSF1, are reported to be elevated in diabetic patients [34–36]. Furthermore, the reduction in the membrane fluidity due to glycation and oxidative stress may possibly explain the decreased expression of HSF1 [37,38]. Thus results of the present study indicate that Hsp70 is mediated by the HSF1 whereas some of the sHsp such

Fig. 5. Translocation of pS82-Hsp27, pS59-aBC from soluble (cytosol) to insoluble (cytoskeleton) fraction. Panel A: The heart samples were fractionated into TritonX soluble (cytosol) and insoluble (cytoskeleton) fractions and subjected to immunoblotting. Chronic hyperglycemia induced the translocation of pS82-Hsp27, pS59-aBC and desmin from soluble to insoluble fraction. Panel B: Interaction of sHsp with sarcomere marker desmin as analysed by co-immunoprecipitation in insoluble fractions of control and diabetic rat heart. pS82-Hsp27, pS59-aBC showed increased interaction with desmin under chronic hyperglycemic conditions. IB, immunoblotting antibody; IP, immunoprecipitation; IP (), no antibody control immunoprecipitation.

V.S. Reddy et al. / Archives of Biochemistry and Biophysics 558 (2014) 1–9

7

Fig. 6. Immunofluorescence of pS82-Hsp27 and pS59-aBC in control and diabetic heart Panel A: Tissue sections of control (a, b, c, d) and diabetic (a0 , b0 , c0 , d0 ) heart were labeled with pS82-Hsp27 antibody (a), DAPI (b), merged (c) and at higher magnification (d). Panel B: Tissue sections of control (a, b, c, d) and diabetic (a0 , b0 , c0 , d0 ) heart were labeled with pS59-aBC antibody (green; a), DAPI (blue; b), merged (c) and at higher magnification (d). Arrows represent pS82-Hsp27 and pS59-aBC staining at sarcomere striations in diabetic heart. All the pictures were taken at 63 magnification, Scale bars (a, a0 , b, b0 , c, c0 ), 50 lm, (d, d0 ), 20 lm. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

as MKBP, HspB3 and aBC may be mediated by alternative mechanism. Further, activation of the HSF1 pathway led to cytoprotection and improved insulin signaling in cells, consistent with effects observed with Hsp70 treatment [39]. It is also reported that there is a deficiency of Hsp70 in skeletal muscle in type 2 diabetes [40]. This study reports that Hsp70 levels were also low in diabetic heart. Of the 10 sHsp, Hsp27 and aBC are major sHsp that show protective effect against various stress or pathological insults [41,42]. The properties of Hsp27 and aBC are often regulated by their phosphorylation state. The phosphorylated Hsp27 and aBC have been shown to provide enhanced protection against disruption of cytoskeletal elements including actin, desmin, titin and microtubules [43,44]. In the present study, the increased immunostaining of pS82-Hsp27 and pS59-aBC at striations in diabetic heart when compared with either low or absence of immunostaining in control rats indicate the phosphorylation may mediate translocation of sHsp to striated sarcomeres. This is further corroborated by increased interaction of pS82-Hsp27 and pS59-aBC with desmin under hyperglycemic conditions. Furthermore, we showed the increased levels of desmin in detergent-soluble and detergentinsoluble fraction of diabetic heart (Fig. 5A). Previous studies reported the increased expression of desmin in the heart of prediabetic mice [31], failing human myocardium [45]. Hence, the sHsp were translocated from detergent soluble fraction to detergent insoluble fractions and interacted with desmin as analysed by co-immunoprecipitation under chronic uncontrolled hyperglycemia. Several past studies have also demonstrated that Hsp27 and aBC interact with desmin, other intermediate filaments and prevent their aggregation [46,47]. Translocation of phosphorylated Hsp27 and aBC to striated sarcomeres under various stress or

pathological insults including muscle lengthening contractions [48], ischemia [49,50], after cardioplegia and cardiopulmonary bypass [51] is also well reported. In support of this, one of our previous studies also reported the translocation of the pS59-aBC from detergent-soluble fraction to detergent-insoluble fraction in diabetic rat retina [22]. Activation of p38MAPK and p44/42 MAPK pathway by increasing the levels of p-p38MAPK and p44/42 MAPK in chronic hyperglycemia might be mediating the phosphorylation of pS82-Hsp27, pS59-aBC and pS45-aBC respectively. The reason for increased expression of pS45-aBC remains unknown. Previously, we have reported the response of sHsp in diabetic retina in the same set-up of the study. There were some similarities between retina and heart. For example, the expression pattern of aBC, Hsp20 and phosphorylation of aBC on S45 and S59 in heart and retina showed similar trend under chronic hyperglycemia. However, Hsp27 and Hsp22 have shown differential effects: while Hsp27 levels were decreased in soluble fraction with concomitant increase in insoluble fraction of retina, it was unaltered in heart. Similarly, Hsp22 levels were increased in retina while unaltered in heart. These results indicate tissue specific response of sHsp under hyperglycemia and further studies are warranted to understand the significance tissue specific response of sHsp. In the present study, chronic hyperglycemia strongly induced cell death as assessed by increased number of TUNEL positive cells, increased levels of proapoptotic proteins such as cleaved caspase3, Bax and decreased levels of antiapoptotic protein Bcl-2 in diabetic heart when compared with controls. aBC acts as antiapoptotic regulator and it has been shown to protect the cells from osmotic [52], thermal [53] and oxidative stress [54]. Furthermore, it protected the cardiomyocytes from myocardial infarction [55] and H2O2 exposure [56]. aBC prevents the cell death by inhibiting

8

V.S. Reddy et al. / Archives of Biochemistry and Biophysics 558 (2014) 1–9

Fig. 7. Apoptotic cell death in heart of control and diabetic rats. Panel A: Apoptosis is detected by TUNEL staining of tissue sections of control (a, b, c) and diabetic heart (a0 , b0 , c0 ). The green stain indicates the apoptotic cell death and blue stain (DAPI) indicates nuclei. Scale bars, 50 lm and bars demonstrating the number of TUNEL-positive cells in control and diabetic heart. Data represent mean ± SEM of three independent experiments (⁄⁄⁄p < 0.001). Panel B: Representative immunoblots of Bax, Bcl-2 and cleaved caspase-3 and respective quantitative bars demonstrating the Bax, Bcl-2 and cleaved caspase-3 levels in heart of control and diabetic rats. Data were normalized for actin expression and represented as percent of control. Data represent mean ± SEM of three independent experiments (⁄p < 0.05). Chronic hyperglycemia significantly increased the cleaved caspase-3, Bax whereas decreased the Bcl-2 levels. Interaction of aBC (Panels C) and pS59-aBC (Panels D) with Bax as analysed by co-immunoprecipitation. Chronic hyperglycemia reduced the interaction of aBC and pS59-aBC with Bax in rat heart. IB, immunoblotting antibody; IP, immunoprecipitation; IP (), no antibody (control immunoprecipitation). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

the proapoptotic actions of caspase-3 and Bax through their direct interaction [56,57]. In spite of increased amount of the aBC and Bax, their interaction was dramatically reduced under chronic hyperglycemic conditions. While some studies have shown that phosphorylation of aBC on S59 protected the cardiac myocytes [58], others showed negative effect on apoptosis [59]. Interestingly, in the present study, pS59-aBC showed reduced interaction with Bax. Hence, chronic hyperglycemia also might impair the protective function of aBC. Such observations were made in type-1 and type-2 diabetic rat retina [60,61]. In conclusion, we report differential expression of sHsp family members in a comprehensive manner and site specific phosphorylation of aBC under chronic hyperglycemia. The results also suggest that sHsp may have tissue-specific function while fulfilling general purpose under various stress conditions. Members of sHsp family may express either acutely or indistinctly and may act individually or in concert to enable cells to resist stress conditions. Together the results suggest induction, phosphorylation, translocation of sHsp in diabetic cardiomyopathy and further studies are required to ascertain their role in cell death.

Acknowledgements This work was supported by grants from Department of Science and Technology and Department of Biotechnology, Government of India to G.B.R. V.S.R. was supported by a research fellowship from the University Grants Commission, Government of India. We thank Sneha Jakhotia for the help in editing the manuscript.

References [1] Diabetes Atlas 6th Edition (2013). [2] D. Westermann, T. Walther, K. Savvatis, F. Escher, M. Sobirey, A. Riad, M. Bader, H.P. Schultheiss, C. Tschope, Diabetes 58 (2009) 1373–1381. [3] E. Adeghate, Mol. Cell. Biochem. 261 (2004) 187–191. [4] A. Avogaro, S. Vigili de Kreutzenberg, C. Negut, A. Tiengo, R. Scognamiglio, Am. J. Cardiol. 93 (2004) 13A–16A. [5] E. Picano, J. Am. Coll. Cardiol. 42 (2003) 454–457. [6] P.R. Taub, I. Ramirez-Sanchez, T.P. Ciaraldi, S. Gonzalez-Basurto, R. CoralVazquez, G. Perkins, M. Hogan, A.S. Maisel, R.R. Henry, G. Ceballos, F. Villarreal, Clin. Sci. (Lond.) 125 (2013) 383–389. [7] I. Falcao-Pires, A.F. Leite-Moreira, Heart Fail. Rev. 17 (2012) 325–344. [8] M. Aragno, R. Mastrocola, G. Alloatti, I. Vercellinatto, P. Bardini, S. Geuna, M.G. Catalano, O. Danni, G. Boccuzzi, Endocrinology 149 (2008) 380–388. [9] L. Cai, Y.J. Kang, Cardiovasc. Toxicol. 1 (2001) 181–193. [10] L. Cai, Y. Wang, G. Zhou, T. Chen, Y. Song, X. Li, Y.J. Kang, J. Am. Coll. Cardiol. 48 (2006) 1688–1697. [11] L. Cai, Free Radic. Biol. Med. 41 (2006) 851–861. [12] G. Kappe, E. Franck, P. Verschuure, W.C. Boelens, J.A. Leunissen, W.W. de Jong, Cell Stress Chaperones 8 (2003) 53–61. [13] A.P. Arrigo, Med. Sci. (Paris) 21 (2005) 619–625. [14] M. Haslbeck, Cell. Mol. Life Sci. 59 (2002) 1649–1657. [15] Y. Sun, T.H. MacRae, Cell. Mol. Life Sci. 62 (2005) 2460–2476. [16] L. Pirkkala, P. Nykanen, L. Sistonen, FASEB J. 15 (2001) 1118–1131. [17] H. Ecroyd, S. Meehan, J. Horwitz, J.A. Aquilina, J.L. Benesch, C.V. Robinson, C.E. Macphee, J.A. Carver, Biochem. J. 401 (2007) 129–141. [18] H.A. Koteiche, H.S. McHaourab, J. Biol. Chem. 278 (2003) 10361–10367. [19] J. Landry, H. Lambert, M. Zhou, J.N. Lavoie, E. Hickey, L.A. Weber, C.W. Anderson, J. Biol. Chem. 267 (1992) 794–803. [20] H.E. Hoover, D.J. Thuerauf, J.J. Martindale, C.C. Glembotski, J. Biol. Chem. 275 (2000) 23825–23833. [21] P.A. Kumar, A. Haseeb, P. Suryanarayana, N.Z. Ehtesham, G.B. Reddy, Arch. Biochem. Biophys. 444 (2005) 77–83. [22] V.S. Reddy, G. Raghu, S.S. Reddy, A.K. Pasupulati, P. Suryanarayana, G.B. Reddy, Invest. Ophthalmol. Vis. Sci. 54 (2013) 7674–7682. [23] V.S. Reddy, C.U. Kumar, G.B. Reddy, Biochem. Biophys. Res. Commun. 446 (2014) 602–607.

V.S. Reddy et al. / Archives of Biochemistry and Biophysics 558 (2014) 1–9 [24] M.S. Kumar, M. Kapoor, S. Sinha, G.B. Reddy, J. Biol. Chem. 280 (2005) 21726– 21730. [25] G.B. Reddy, V. Vasireddy, M.N. Mandal, M. Tiruvalluru, X.F. Wang, M.M. Jablonski, G. Nappanveettil, R. Ayyagari, Invest. Ophthalmol. Vis. Sci. 50 (2009) 3456–3463. [26] H. Chen, X.J. Wu, X.Y. Lu, L. Zhu, L.P. Wang, H.T. Yang, H.Z. Chen, W.J. Yuan, Acta Pharmacol. Sin. 26 (2005) 806–812. [27] Y. Sugiyama, A. Suzuki, M. Kishikawa, R. Akutsu, T. Hirose, M.M. Waye, S.K. Tsui, S. Yoshida, S. Ohno, J. Biol. Chem. 275 (2000) 1095–1104. [28] J. den Engelsman, S. Boros, P.Y. Dankers, B. Kamps, W.T. Vree Egberts, C.S. Bode, L.A. Lane, J.A. Aquilina, J.L. Benesch, C.V. Robinson, W.W. de Jong, W.C. Boelens, J. Mol. Biol. 393 (2009) 1022–1032. [29] A. Suzuki, Y. Sugiyama, Y. Hayashi, N. Nyu-i, M. Yoshida, I. Nonaka, S. Ishiura, K. Arahata, S. Ohno, J. Cell. Biol. 140 (1998) 1113–1124. [30] K.M. Shama, A. Suzuki, K. Harada, N. Fujitani, H. Kimura, S. Ohno, K. Yoshida, Cell Struct. Funct. 24 (1999) 1–4. [31] D. Cruz-Topete, E.O. List, S. Okada, B. Kelder, J.J. Kopchick, J. Proteomics 74 (2011) 716–727. [32] S.G. Zhu, L. Xi, R.C. Kukreja, J. Cell. Mol. Med. 16 (2012) 950–958. [33] R. Buriro, Y.J. Lv, I. Ali, S. Tang, Z.J. Liu, M. Zhang, A. Adem, J. Hartung, E.D. Bao, Genet. Mol. Res. 12 (2013) 3003–3016. [34] X. Wang, M.A. Khaleque, M.J. Zhao, R. Zhong, M. Gaestel, S.K. Calderwood, J. Biol. Chem. 281 (2006) 782–791. [35] S.E. Nikoulina, T.P. Ciaraldi, S. Mudaliar, P. Mohideen, L. Carter, R.R. Henry, Diabetes 49 (2000) 263–271. [36] B. He, Y.H. Meng, N.F. Mivechi, Mol. Cell. Biol. 18 (1998) 6624–6633. [37] P.L. Hooper, J.J. Hooper, Diabetes Technol. Ther. 7 (2005) 204–208. [38] P. Tong, T. Thomas, T. Berrish, D. Humphriss, L. Barriocanal, M. Stewart, M. Walker, R. Wilkinson, K.G. Alberti, Lancet 345 (1995) 357–358. [39] J.H. Lee, J. Gao, P.A. Kosinski, S.J. Elliman, T.E. Hughes, J. Gromada, D.M. Kemp, Biochem. Biophys. Res. Commun. 430 (2013) 1109–1113. [40] J. Chung, A.K. Nguyen, D.C. Henstridge, A.G. Holmes, M.H. Chan, J.L. Mesa, G.I. Lancaster, R.J. Southgate, C.R. Bruce, S.J. Duffy, I. Horvath, R. Mestril, M.J. Watt, P.L. Hooper, B.A. Kingwell, L. Vigh, A. Hevener, M.A. Febbraio, Proc. Natl. Acad. Sci. U.S.A. 105 (2008) 1739–1744. [41] Y. Sun, T.H. MacRae, FEBS J. 272 (2005) 2613–2627.

9

[42] A.P. Arrigo, S. Simon, B. Gibert, C. Kretz-Remy, M. Nivon, A. Czekalla, D. Guillet, M. Moulin, C. Diaz-Latoud, P. Vicart, FEBS Lett. 581 (2007) 3665–3674. [43] F. Bennardini, A. Wrzosek, M. Chiesi, Circ. Res. 71 (1992) 288–294. [44] N. Golenhofen, A. Arbeiter, R. Koob, D. Drenckhahn, J. Mol. Cell. Cardiol. 34 (2002) 309–319. [45] A. Heling, R. Zimmermann, S. Kostin, Y. Maeno, S. Hein, B. Devaux, E. Bauer, W.P. Klovekorn, M. Schlepper, W. Schaper, J. Schaper, Circ. Res. 86 (2000) 846– 853. [46] M.D. Perng, L. Cairns, I.P. van den, A. Prescott, A.M. Hutcheson, R.A. Quinlan, J. Cell. Sci. 112 (Pt 13) (1999) 2099-2112. [47] J.L. Elliott, M. Der Perng, A.R. Prescott, K.A. Jansen, G.H. Koenderink, R.A. Quinlan, Philos. Trans. R. Soc. Lond. B Biol. Sci. 368 (2013) 20120375. [48] T.J. Koh, J. Escobedo, Am. J. Physiol. Cell Physiol. 286 (2004) C713–C722. [49] P. Eaton, W. Fuller, J.R. Bell, M.J. Shattock, J. Mol. Cell. Cardiol. 33 (2001) 1659– 1671. [50] N. Golenhofen, W. Ness, R. Koob, P. Htun, W. Schaper, D. Drenckhahn, Am. J. Physiol. 274 (1998) H1457–H1464. [51] R.T. Clements, N.R. Sodha, J. Feng, S. Mieno, M. Boodhwani, B. Ramlawi, C. Bianchi, F.W. Sellke, J. Thorac. Cardiovasc. Surg. 134 (2007) 1461–1470. [52] N. Golenhofen, W. Ness, E.F. Wawrousek, D. Drenckhahn, Histochem. Cell. Biol. 117 (2002) 203–209. [53] A. Aoyama, E. Frohli, R. Schafer, R. Klemenz, Mol. Cell. Biol. 13 (1993) 1824– 1835. [54] P. Mehlen, X. Preville, P. Chareyron, J. Briolay, R. Klemenz, A.P. Arrigo, J. Immunol. 154 (1995) 363–374. [55] A. Mitra, T. Basak, K. Datta, S. Naskar, S. Sengupta, S. Sarkar, Cell Death Dis. 4 (2013) e582. [56] R. Chis, P. Sharma, N. Bousette, T. Miyake, A. Wilson, P.H. Backx, A.O. Gramolini, Am. J. Physiol. Heart Circ. Physiol. 303 (2012) H967–H978. [57] Y.W. Mao, J.P. Liu, H. Xiang, D.W. Li, Cell Death Differ. 11 (2004) 512–526. [58] L.E. Morrison, H.E. Hoover, D.J. Thuerauf, C.C. Glembotski, Circ. Res. 92 (2003) 203–211. [59] M.C. Kamradt, F. Chen, S. Sam, V.L. Cryns, J. Biol. Chem. 277 (2002) 38731– 38736. [60] M.K. Losiewicz, P.E. Fort, Invest. Ophthalmol. Vis. Sci. 52 (2011) 5034–5042. [61] Y.H. Kim, M.Y. Choi, Y.S. Kim, J.M. Han, J.H. Lee, C.H. Park, S.S. Kang, W.S. Choi, G.J. Cho, Neurobiol. Dis. 28 (2007) 293–303.