Cysteamine Prevents Vascular Leakage Through Inhibiting Transglutaminase in Diabetic Retina

235 1

y-j lee and others Cysteamine prevention of 235:1 39–48 Research vascular leakage

Cysteamine prevents vascular leakage through inhibiting transglutaminase in diabetic retina

Yeon-Ju Lee1, Se-Hui Jung1, JongYun Hwang2, Sohee Jeon3, Eun-Taek Han4, Won Sun Park5, Seok-Ho Hong6, Young-Myeong Kim1 and Kwon-Soo Ha1

1Department of Molecular and Cellular Biochemistry, Kangwon National University School of Medicine, Chuncheon, Kangwon-do, Korea 2Department of Obstetrics and Gynecology, Kangwon National University School of Medicine, Chuncheon, Kangwon-do, Korea 3Department of Ophthalmology, Seoul St. Mary’s Hospital, Catholic University, Seoul, Korea 4Department of Medical Environmental Biology and Tropical Medicine, Kangwon National University School of Medicine, Chuncheon, Kangwon-do, Korea Correspondence 5Department of Physiology, Kangwon National University School of Medicine, Chuncheon, Kangwon-do, Korea should be addressed 6Department of Internal Medicine, Kangwon National University School of Medicine, Chuncheon, to K-S Ha Kangwon-do, Korea Email [email protected]

Abstract

Cysteamine (an aminothiol), which is derived from coenzyme A degradation and Key Words Endocrinology

of metabolized into taurine, has beneficial effects against cystinosis and neurodegenerative ff cysteamine diseases; however, its role in diabetic complications is unknown. Thus, we sought to ff diabetic retinopathy determine the preventive effect of cysteamine against hyperglycemia-induced vascular ff transglutaminase Journal leakage in the retinas of diabetic mice. Cysteamine and ethanolamine, the sulfhydryl ff vascular endothelial group-free cysteamine analogue, inhibited vascular endothelial growth factor (VEGF)- growth factor induced stress fiber formation and vascular endothelial (VE)-cadherin disruption in ff vascular leakage endothelial cells, which play a critical role in modulating endothelial permeability. Intravitreal injection of the amine compounds prevented hyperglycemia-induced vascular leakage in the retinas of streptozotocin-induced diabetic mice. We then investigated the potential roles of reactive oxygen species (ROS) and transglutaminase (TGase) in the cysteamine prevention of VEGF-induced vascular leakage. Cysteamine, but not ethanolamine, inhibited VEGF-induced ROS generation in endothelial cells and diabetic retinas. In contrast, VEGF-induced TGase activation was prevented by both cysteamine and ethanolamine. Our findings suggest that cysteamine protects against vascular leakage through inhibiting VEGF-induced TGase activation rather than ROS Journal of Endocrinology generation in diabetic retinas. (2017) 235, 39–48

Introduction

Cysteamine (β-mercaptoethylamine), an aminothiol with depletes cystine through reacting with its disulfide bridge, a primary amine group and a sulfhydryl group, is derived changes the enzymatic activity of several proteins as a from coenzyme A degradation and metabolized into result of binding to their thiol groups and enhances brain- taurine (Pinto et al. 2005, Besouw et al. 2013). Cysteamine derived neurotrophic factor secretion (Besouw et al. 2013).

10.1530/JOE-17-0109 Research y-j lee and others Cysteamine prevention of 235:1 40 vascular leakage

Due to cysteamine’s varied effects, it has been proposed that cysteamine prevents vascular leakage by inhibiting as a treatment for many diseases, including cystinosis hyperglycemia-induced TGase activation rather than ROS and neurodegenerative and neuropsychiatric disorders generation in the diabetic retinas. We found that the (Besouw et al. 2013). Cysteamine was initially utilized primary amine group of cysteamine was implicated in the to protect against X-rays and as a therapy for radiation prevention of VEGF-induced endothelial cell permeability sickness (Bacq et al. 1953). This compound was introduced and hyperglycemia-induced vascular leakage in the retinas as a potential treatment for cystinosis and is currently of streptozotocin-induced diabetic mice. The primary used to treat corneal cystine crystal accumulation in amine was also involved in hyperglycemia-induced TGase patients with cystinosis (Kaiser-Kupfer et al. 1987, Tsilou activation in diabetic retinas. However, the sulfhydryl et al. 2003). Cysteamine has been evaluated as a treatment group of cyateamine was required for inhibition of for neurodegenerative diseases, including Huntington’s ROS generation, but not essential for preventing TGase disease (Gibrat & Cicchetti 2011, Shannon & Fraint 2015) activation and vascular leakage. Our findings suggest and Parkinson’s disease (Gibrat & Cicchetti 2011, Cisbani that cysteamine prevents vascular leakage by inhibiting et al. 2015) and has been proposed to treat schizophrenia hyperglycemia-induced TGase activation in the retinas (Buckley et al. 2014), cystic fibrosis (De Stefano et al. of diabetic mice, and thus, has therapeutic potential for 2014) and nonalcoholic fatty liver disease (Dohil et al. preventing diabetic retinopathy and vascular leakage- 2011). However, its beneficial effects against diabetic associated diseases. complications, including diabetic retinopathy, remain unknown. Hyperglycemia and the resulting subsequent Materials and methods physiological changes develop into, among others, Cell culture microvascular and macrovascular diabetic complications (Bhatt et al. 2014). One of the major microvascular Human umbilical vein endothelial cells (HUVECs) were complications, diabetic retinopathy, is the leading isolated from the human umbilical cord vein according to the Declaration of Helsinki as previously described (Lim Endocrinology cause of blindness in adults (Wang et al. 2012, Gupta of et al. 2013, Bhatt et al. 2014), which progresses et al. 2014, Lee et al. 2016). Written informed consent from nonproliferative abnormalities, characterized was obtained from all umbilical cord donors. Cells from by vascular leakage, to severe proliferative diabetic several batches were grown at 37°C in a humidified 5% Journal

retinopathy (Fong et al. 2004). In nonproliferative CO2 incubator in M199 culture media supplemented diabetic retinopathy, retinal blood vessels are damaged with 20% FBS, 3 ng/mL basic fibroblastic growth factor by pericyte loss, microaneurysm and vascular leakage (Millipore), 5 U/mL heparin (Sigma), 100 U/mL penicillin (Caldwell et al. 2003, Hammes 2005). Retinal vascular and 100 µg/mL streptomycin. Cells in passages three leakage in the early stages of diabetic retinopathy is to six were incubated for 6 h in low-serum medium predominantly caused by vascular endothelial growth supplemented with 1% FBS and antibiotics. This study was factor (VEGF)-mediated stress fiber formation and approved by the Institutional Review Board of Kangwon vascular endothelial (VE)-cadherin disruption (Spindler National University Hospital. et al. 2010, Lim et al. 2014). Recently, we demonstrated that reactive oxygen species (ROS)-mediated activation Measurement of intracellular ROS generation of transglutaminase (TGase) 2 plays a key role in VEGF-induced retinal vascular leakage in diabetic Intracellular ROS generation was determined using

mice (Lee et al. 2016). Though various drugs against 2′,7′-dichlorodihydrofluorescein diacetate (H2DCFDA, VEGF, oxidative stress and inflammation have been Invitrogen, Molecular Probes) staining as previously evaluated against VEGF-induced molecular events in described (Bhatt et al. 2016). Cells were treated with the diabetic retina, long-lasting therapies with minimal amine compounds for 30 min and then incubated

complications are required to treat diabetic retinopathy with 10 ng/mL VEGF (Millipore) and 10 µM H2DCFDA (Stewart 2016, Tolentino et al. 2016). in low-serum media (phenol red-free) for 10 min. Therefore, we aimed to determine the preventive Labeled cells were immediately analyzed by confocal effect of cysteamine on hyperglycemia-induced vascular microscopy (Fluoview-300; Olympus), and intracellular leakage in the retina of diabetic mice. We hypothesized ROS levels (fold) were determined by measuring

http://joe.endocrinology-journals.org © 2017 Society for Endocrinology Published by Bioscientifica Ltd. DOI: 10.1530/JOE-17-0109 Printed in Great Britain Downloaded from Bioscientifica.com at 10/11/2021 12:11:13PM via free access Research y-j lee and others Cysteamine prevention of 235:1 41 vascular leakage

single-cell fluorescence intensities from ten randomly Cells were acid washed for 30 min using PBS (pH 2.7) selected cells per experiment. containing 25 mM glycine and 3% BSA and then fixed and permeabilized as described previously. Cells were incubated with a FITC-conjugated goat anti-mouse Measurement of in situ TGase activity antibody (1:200, Sigma, Cat. No. F0257) for 2 h and with In situ TGase transamidating activity was determined 1 µg/mL DAPI in PBS for 5 min. VE-cadherin and nuclei by confocal microscopic assay as previously described were visualized using confocal microscopy (K1-Fluo, (Lee et al. 2016). Briefly, cells were incubated with 1 mM Nanoscope Systems, Taejon, Korea). 5-(biotinamido)pentylamine for 1 h at 37°C, fixed with 3.7% formaldehyde in PBS for 30 min and permeabilized Generation of diabetic mouse model with 0.2% Triton X-100 in PBS for 30 min. After incubation with blocking solution (2% BSA in 20 mM Tris (pH 7.6), Six-week-old male C57BL/6 mice were obtained from DBL 138 mM NaCl, and 0.1% Tween-20) for 30 min, cells (EumSeong, Korea). Diabetic mice were generated by a single were treated with FITC-conjugated streptavidin (1:200, intraperitoneal injection of streptozotocin (150 mg/kg v/v) in blocking solution for 1 h. Single-cell fluorescence body weight, Sigma) freshly prepared in 100 mM citrate intensities of stained cells were determined for 10 buffer (pH 4.5) as previously described (Bhatt et al. 2013). randomly selected cells per experiment. Mice with non-fasting blood glucose levels greater than 19 mM, polyuria and glucosuria were considered diabetic. Two weeks after streptozotocin injection, diabetic mice Immunofluorescence and VE-cadherin internalization were subjected to the measurements of ROS levels, Actin filaments and VE-cadherin were visualized as in vivo TGase activity and vascular leakage in retinas. No previously described (Lim et al. 2014). In order to mice were treated with insulin. All animal experiments visualize actin filaments, cells were treated with 10 ng/mL conformed to the NIH guidelines (Guide for the Care VEGF for 1 h at 37°C, fixed with 3.7% formaldehyde and Use of Laboratory Animals) and were approved by

Endocrinology and permeabilized with 0.2% Triton X-100. Cells were the Kangwon Institutional Animal Care and Use Ethics of incubated with Alexa Fluor 488 phalloidin (1:200, Committee. Invitrogen, Molecular Probes) for 1 h at room temperature,

Journal and actin filaments were observed with a confocal Measurement of vascular leakage in retinas microscope. VE-cadherin was visualized in cells preincubated Microvascular leakage in mouse retinas was investigated with TGase inhibitors for 30 min and treated with by fluorescein angiography as previously described Lim( 10 ng/mL VEGF for 90 min at 37°C. Following fixation et al. 2014). Briefly, 24 h after intravitreal injection of and permeabilization, cells were incubated overnight 2 µL 500 mM cysteamine or 100 mM ethanolamine, with a monoclonal VE-cadherin antibody (1:200, mice were deeply anesthetized and 1.25 mg 500 kDa Santa Cruz Biotechnology, Cat. No. sc-9989) at 4°C. FITC-dextran (Sigma) was injected into the left ventricle. Cells were then probed with a FITC-conjugated goat After 5 min, mice were killed by cervical dislocation. The anti-mouse antibody (1:200, Sigma, Cat. No. F0257) eyes were enucleated, and immediately fixed with 4% in PBS for 2 h, and VE-cadherin was visualized using paraformaldehyde for 45 min. The retinas were dissected confocal microscopy. VE-cadherins were represented in the Maltese cross configuration and flat-mounted by line profiles displaying the distribution of relative onto glass slides. The superficial vessels of the retinas fluorescence intensities (RFIs) obtained using the (n = 6 per group) were observed by confocal microscopy, Fluoview software (FV-300), indicated by dotted lines and vascular leakage was quantitatively analyzed using crossing two cell–cell contacts. Fluoview software (FV-300). Internalization of VE-cadherin was visualized as previously described (Li et al. 2012). Briefly, cells Measurement of ROS levels in mouse retinas were incubated with a monoclonal antibody against VE-cadherin extracellular domain (1:200, Millipore, ROS levels in mouse retinas were determined by confocal clone BV6) for 1 h and treated with 10 ng/mL VEGF microscopy as previously described (He et al. 2014). Briefly, for 90 min in the presence of 100 µM chloroquine. mice were killed by cervical dislocation and the eyes

were enucleated and quickly frozen in optimal cutting Statistical analyses temperature (O.C.T.) compound (Sakura Finetek, Torrance, Data processing was performed using Origin 6.1 CA, USA). Unfixed cryosections (10 μm) were prepared software (OriginLab, Northampton, MA, USA). Statistical using a microtome-cryostat (Leica Biosystems), washed significance was determined using Student’s t-tests. A P with PBS and incubated with 5 μM dihydroethidium value <0.05 was considered statistically significant, and (DHE; Invitrogen, Molecular Probes) in PBS for 30 min data were expressed as the mean ± s.d. of at least three at 37°C. The stained sections were observed by confocal independent experiments. microscopy and ROS levels of mouse retinal cryosections (n = 6 per group) were quantitatively analyzed using fluorescence n( = 6 per group). Results

Cysteamine and ethanolamine inhibit VEGF-induced Measurement of in vivo TGase activity in mouse retinas stress fiber formation and VE-cadherin disruption in In vivo TGase transamidating activity was determined HUVECs in mouse retinas by confocal microscopy as previously In order to investigate whether cysteamine can described (Lee et al. 2016). Briefly, 24 h after intravitreal prevent hyperglycemia-induced vascular leakage, we injection of 2 µL 500 mM cysteamine or 100 mM examined the effect of cysteamine on VEGF-mediated ethanolamine, mice were deeply anesthetized using stress fiber formation and VE-cadherin disruption in 2.5% avertin, and then 48 µL 100 mM 5-(biotinamido) HUVECs. Vascular leakage in the retinas of diabetic pentylamine was injected into the left ventricle. After mice is predominantly caused by hyperglycemia- 10 min, mice were killed by cervical dislocation. Eyes were induced stress fiber formation and adherens junction enucleated, immediately fixed with 4% paraformaldehyde disruption (Lim et al. 2014, Lee et al. 2016). We found for 45 min and retinas were dissected in the Maltese cross that VEGF activated the formation of stress fibers, which configuration and permeabilized with 0.2% Triton X-100 was prevented by cysteamine (Fig. 1B). Furthermore, Endocrinology in PBS for 30 min at room temperature. After a 30-min through visualization of VE-cadherin to determine of incubation with blocking solution, retinas were treated adherens junction integrity, we found that cysteamine with FITC-conjugated streptavidin (1:200, v/v) in blocking inhibited VEGF-induced adherens junction disassembly Journal solution for 1 h at room temperature. The superficial (Fig. 1C). The changes in VE-cadherin stability are vessels of stained retinas (n = 6 per group) were observed represented by line profiles displaying the distribution using confocal microscopy. In vivo TGase activities were of relative fluorescence intensities (RFIs), indicated by quantitatively determined using fluorescence intensities dotted lines crossing two cell–cell contacts (Fig. 1C). in the retinas of normal and diabetic mice. VE-cadherin disassembly was further investigated using VE-cadherin internalization assay. VEGF significantly increased intracellular VE-cadherin, which was reversed Measurement of VEGF levels in mouse retinas by cysteamine (P < 0.001, Fig. 1D). Thus, it is likely that VEGF levels in mouse retinas were determined using cysteamine prevents VEGF-induced endothelial cell ELISA method according to the manufacturer’s protocol permeability by inhibiting stress fiber formation and (Quantikine Mouse VEGF ELISA kit; R&D Systems). Two adherens junction disruption. weeks after streptozotocin injection, mice were killed by We then treated endothelial cells with ethanolamine, cervical dislocation and the eyes were rapidly enucleated. the sulfhydryl group-free cysteamine analog (Fig. 1A), Four retinas of each sample were sonicated for 15 s in to study the function of the primary amine group 130 µL of ice-cold lysis buffer containing 50 mM Tris– of cysteamine in the prevention of VEGF-induced HCl (pH 7.5), 1% Triton X-100, 150 mM NaCl, 1 mM endothelial cell permeability. Ethanolamine prevented EDTA, 0.1 mM phenylmethylsulfonyl fluoride, 10 µg/ VEGF-mediated stress fiber formation, VE-cadherin mL aprotinin and 10 µg/mL leupeptin. The lysates were disruption and VE-cadherin internalization (Fig. 1B, centrifuged at 14,000 rpm for 15 min at 4°C and retinal C and D). These results indicate that the primary VEGF concentrations were determined using an ELISA amine of cysteamine is implicated in the prevention of reader (Molecular Devices). VEGF-induced endothelial cell permeability.

Figure 1 Effects of cysteamine and ethanolamine on VEGF-induced stress fiber formation and VE-cadherin disruption in HUVECs. HUVECs were incubated with 500 µM cysteamine (CYE) or 100 µM ethanolamine (ETA) for 30 min and treated with 10 ng/mL VEGF. Stress fibers, VE-cadherin and VE-cadherin internalization were stained and visualized by confocal microscopy (n = 3). (A) The chemical structures of cysteamine and ethanolamine. (B) Representative images of stress fibers. Scale bar, 30 µM. (C) Representative images of VE-cadherin (top). Scale bar, 10 µM. Adherens junctions are represented by histograms of VE-cadherin (bottom), indicated by the dotted lines in the images. RFI, relative fluorescence intensity. (D) Internalization of VE-cadherin (green). Nuclei (blue) were stained with DAPI. Scale bar, 20 µM. A full colour version of this figure is available athttp://dx.doi.org/10.1530/JOE-17-0109

Cysteamine and ethanolamine inhibit using fluorescence angiography (Fig. 2A). High levels of hyperglycemia-induced vascular leakage in the retinas extravasation of FITC-dextran were observed in the retina of diabetic mice of diabetic mice (n = 6); this vascular leakage was blocked in the retinas of cysteamine-injected contralateral eyes (n = 6). In order to validate our in vitro findings, we investigated the We quantitatively analyzed vascular leakage by determining preventive role of cysteamine against hyperglycemia-induced the fluorescence intensity of FITC-dextran in mouse retinas vascular leakage in the retinas of streptozotocin-induced (Fig. 2B) and found that intravitreal injection of cysteamine diabetic mice. Diabetic mice were intravitreally injected significantly inhibited this leakage n( = 6, P < 0.001). with cysteamine and then vascular leakage was investigated

Figure 2 Effect of cysteamine and ethanolamine on hyperglycemia-induced vascular leakage in diabetic mouse retinas. Streptozotocin-induced diabetic mouse eyes were intravitreally injected with 2 µL of PBS (diabetic), 500 mM cysteamine (diabetic + CYE) or 100 mM ethanolamine (diabetic + ETA). The eyes of nondiabetic mice were intravitreally injected with PBS (normal). Vascular leakage in retinas was visualized by confocal microscopy (n = 6 per group). (A) Representative fluorescent images of the retinas. The bottom row images are magnifications of the areas indicated by squares in the top row. Scale bar, 150 µM. (B) Retina permeability was quantified by measuring the fluorescence intensity of FITC-dextran in the images in A. ***P < 0.001. A full colour version of this figure is available athttp://dx.doi.org/10.1530/JOE-17-0109

We then intravitreally injected diabetic mice with We further investigated the effect of cysteamine on ethanolamine and examined vascular leakage using hyperglycemia-induced ROS generation in the retina of fluorescence angiography. Intravitreal injection of diabetic mice (Fig. 4A). We found that ROS levels were ethanolamine significantly inhibited hyperglycemia- elevated in diabetic compared to normal mouse retinas. induced extravasation of FITC-dextran in the retinas of Cysteamine prevented this hyperglycemia-induced Endocrinology diabetic mice (Fig. 2A and B, n = 6, P < 0.001). These results ROS generation. The changes in retinal ROS levels are of suggest that the primary amine of cysteamine is involved represented by line profiles displaying the distribution of in the prevention of hyperglycemia-induced vascular DHE RFIs, as shown by dotted lines crossing the mouse Journal leakage in the retinas of diabetic mice. retina. We then quantitatively analyzed ROS generation by measuring the RFIs of DHE in retinal tissues (Fig. 4B). The average ROS level in the diabetic retina was approximately Cysteamine, but not ethanolamine, inhibits VEGF-induced ROS generation in endothelial cells and hyperglycemia-induced ROS generation in the retinas of diabetic mice

In order to study the mechanism by which cysteamine prevents vascular leakage, initially, we determined the effects of cysteamine and ethanolamine on VEGF-induced generation of intracellular ROS in HUVECs (Fig. 3). We found that cysteamine inhibited VEGF-induced ROS generation in a concentration-dependent manner, with maximum effect at 50 µM (P < 0.001). The ROS scavengers N-acetylcysteine and Trolox prevented VEGF-induced ROS generation (data not shown). However, the sulfhydryl group-free ethanolamine had no effect on VEGF-induced Figure 3 Effects of cysteamine and ethanolamine on VEGF-induced ROS ROS generation at any concentration (up to 500 µM), generation in HUVECs. HUVECs were treated with the indicated demonstrating that the sulfhydryl group is essential concentrations of cysteamine or ethanolamine for 30 min and then incubated with 10 ng/mL VEGF and 10 µM H DCFDA in low-serum medium for antioxidant activity of cysteamine. Consistently, 2 for 10 min. Intracellular ROS levels were determined by confocal ethanolamine had no DPPH free radical scavenging microscopy. Results are expressed as mean ± s.d. from three independent activity (data not shown). experiments. **P < 0.01, ***P < 0.001.

Figure 4 Effects of cysteamine and ethanolamine on hyperglycemia-induced ROS generation in diabetic mouse retinas. Streptozotocin-induced diabetic mice were intravitreally injected with 2 µL of 500 mM cysteamine (diabetic + CYE) or 100 mM ethanolamine (diabetic + ETA) into one eye and an equal volume of PBS into the contralateral eye (diabetic). Nondiabetic mice were intravitreally injected with 2 µL PBS into both eyes (normal). Retinal ROS levels were visualized by confocal microscopy and quantified by fluorescence intensityn ( = 6 per group). (A) Representative images of ROS levels in retinas. Scale bar, 80 µM. Retinal ROS levels are represented by histograms of DHE fluorescence intensity (bottom), as indicated by the dotted lines in the images. RFI, relative fluorescence intensity. ONL, outer nuclear layer; INL, inner nuclear layer; GCL, ganglion cell layer. (B) Retinal ROS levels were quantified by measuring the fluorescence intensity in the retinas. ***P < 0.001. A full colour version of this figure is available athttp://dx.doi.org/10.1530/JOE-17-0109

2.8-fold higher than that in normal mice (P < 0.001, activity (data not shown). These results demonstrate that n = 6). Cysteamine reversed this hyperglycemia-induced the primary amine of cysteamine and ethanolamine is increase of ROS levels in diabetic retinas (P < 0.001, n = 6). involved in preventing VEGF-induced TGase activation in Consistent with in vitro findings, ethanolamine had no endothelial cells. Endocrinology inhibitory effect on the hyperglycemia-induced ROS We then investigated the effect of cysteamine on of generation in the retinas of diabetic mice (Fig. 4A and hyperglycemia-induced TGase activation in the diabetic B). Thus, our results suggest that the sulfhydryl group of retinas (Fig. 6A). In vivo TGase activity was highly

Journal cysteamine is required for inhibiting ROS generation, but elevated in the blood vessels and ganglion cells of the not essential for preventing vascular leakage in diabetic diabetic compared to the normal mouse retina (n = 6); this mouse retinas. hyperglycemia-induced TGase activation was suppressed by intravitreal injection of cysteamine. We quantitatively analyzed in vivo TGase activity by determining the Cysteamine and ethanolamine inhibits VEGF-induced fluorescence intensities of FITC-conjugated streptavidin TGase activation in endothelial cells and in retinal tissues (Fig. 6B). The average TGase activity hyperglycemia-induced TGase activation in the retinas in the diabetic retina was approximately 2.5-fold higher of diabetic mice than that of normal mice (P < 0.001, n = 6), and cysteamine Because the antioxidant activity of cysteamine is not prevented this increase (P < 0.001, n = 6). Ethanolamine essential for preventing vascular leakage, we examined the also prevented hyperglycemia-induced TGase activation effect of cysteamine on VEGF-induced TGase activation in diabetic retinas (Fig. 6A and B). Consistent with a in HUVECs (Fig. 5). VEGF increased in situ TGase activity previous report (Lee et al. 2016), the TGase inhibitors in a time-dependent manner, with maximal activation at cystamine and MDC inhibited hyperglycemia-induced 2 h (P < 0.001, Fig. 5A), and cysteamine inhibited VEGF- vascular leakage in the retinas of diabetic mice (data induced TGase activation in a concentration-dependent not shown), demonstrating the important role of TGase manner, with maximal effect at 500 µM (P < 0.001, Fig. 5B). in hyperglycemia-induced vascular permeability. Taken Ethanolamine also inhibited VEGF-induced TGase together, our results demonstrate that the primary activation in a concentration-dependent manner. The amine of cysteamine plays a key role in the inhibition of TGase inhibitors cystamine and monodansylcadaverine hyperglycemia-induced TGase activation and subsequent (MDC) prevented VEGF-induced elevation of TGase vascular leakage in the retinas of diabetic mice.

inhibited VEGF-induced TGase activation in HUVECs, and intravitreal injection of cysteamine into the eyes of streptozotocin-induced diabetic mice prevented vascular leakage by inhibiting hyperglycemia-induced TGase activation in diabetic mouse retinas. TGase activation is essential for VEGF-induced vascular leakage in the retinas of diabetic mice (Lee et al. 2016), and we found that cysteamine prevented this leakage by inhibiting hyperglycemia-induced TGase activation. Due to the key role of TGase in vascular leakage in the diabetic retina, it is important to investigate TGase inhibitors to identify potential drugs for prevention and treatment of diabetic retinopathy. We demonstrated the inhibitory effect of cysteamine on VEGF-induced TGase activation using an in situ TGase activity assay in HUVECs. Furthermore, using an in vivo TGase activity assay, we found that intravitreal injection of cysteamine inhibited hyperglycemia-induced TGase activation in diabetic mouse retinas. Thus, the sequential application of confocal microscopy-based in situ and in vivo TGase activity assays is a useful approach for identifying potential drugs to prevent TGase-associated diseases, including diabetic retinopathy, and for investigating physiological functions of TGase in cells and tissues.

Endocrinology Figure 5 We demonstrated that cysteamine is a bifunctional of Effects of cysteamine and ethanolamine on VEGF-induced TGase aminothiol with ROS scavenging and TGase-inhibiting activation in HUVECs. HUVECs were treated with the indicated activities. Here, we found that the antioxidant activity concentrations of amine compounds for 30 min and incubated with of cysteamine was attributed by the sulfhydryl group. Journal 10 ng/mL VEGF for 2 h. In situ TGase activities were determined by confocal microscopy. (A) Time-course changes in VEGF-induced TGase Ethanolamine, a sulfhydryl group-free cysteamine activation. (B) Concentration-dependent inhibition of VEGF-induced analog, had no DPPH scavenging activity and no effect TGase activation by amine compounds. Results are expressed as mean ± s.d. from three independent experiments. *P < 0.05, **P < 0.01, on VEGF-induced ROS generation, while cysteamine had ***P < 0.001. a ROS scavenging effect in vitro and in situ. The primary amine group of cysteamine played an essential role in Discussion the inhibition of VEGF-induced TGase activation in It has been suggested that cysteamine has beneficial endothelial cells and diabetic retinas. Intracellular ROS effects against diseases such as cystinosis, Huntington’s generation and/or TGase activation have been implicated disease and Parkinson’s disease. The various functions in the disease pathogenesis of diabetic complications of cysteamine include: depletion of lysosomal cysteine including diabetic retinopathy and cardiovascular in cystinosis, antioxidative effects through increasing diseases, as well as neurodegenerative diseases including intracellular glutathionine levels and regulating enzyme Huntington’s disease and Parkinson’s disease (Park et al. activity (e.g., caspase 3 and transglutaminases) (Shin et al. 2010, Besouw et al. 2013, Bhatt et al. 2013, 2016, Lee et al. 2011, Wilmer et al. 2011, Besouw et al. 2013). In addition, Table 1 Clinical data from age-matched non-diabetic control cysteamine affects the gene expression of brain-derived and two-week diabetic mice. neurotrophic factor and heat shock proteins, among others (Besouw et al. 2013). Here, we describe a new Group Control Diabetes function of cysteamine for preventing hyperglycemia- Number of mice (n) 6 6 induced vascular leakage in the retinas of diabetic mice. Non-fasting blood sugar (mmol/L) 8.1 ± 0.9 29.3 ± 3.7 Retinal VEGF (pg/mg) 161.5 ± 9.8 250.9 ± 16.2** Hyperglycemia significantly elevated the levels of VEGF in

the retinas of diabetic mice (P < 0.01; Table 1). Cysteamine Data are means ± s.d. **P < 0.01 vs control mice.

Figure 6 Effects of cysteamine and ethanolamine on hyperglycemia-induced TGase activation in diabetic mouse retinas. Streptozotocin-induced diabetic mice were intravitreally injected with 2 µL of 500 mM cysteamine (diabetic + CYE) or 100 mM ethanolamine (diabetic + ETA) into one eye and an equal volume of PBS into the contralateral eye (diabetic). Nondiabetic mice were intravitreally injected with 2 µL PBS into both eyes (normal). TGase activity in the retinas were visualized by confocal microscopy and quantitatively determined (n = 6 per group). (A) Representative images of TGase activity in the retinas. The bottom row images are magnifications of the areas indicated by squares in the top row. Scale bar, 100 µM. (B) In vivo TGase activity was quantified by measuring the fluorescence intensity in retinas. ***P < 0.001. A full colour version of this figure is available athttp://dx.doi.org/10.1530/JOE-17-0109

2016). Thus, cysteamine may be used as a multifunctional TGase inhibitors, including cysteamine and cystamine drug for preventing diseases involving ROS generation and TGase 2-specific siRNA, may have the potential for and/or TGase activation. topical ocular application to prevent diabetic retinopathy Amine compounds are potential competitive TGase or other TGase2-associated ocular diseases. inhibitors by competing with natural amine substrates In conclusion, we found that cysteamine prevents Endocrinology in the transamidation reaction. We showed that vascular leakage by inhibiting hyperglycemia-induced of cysteamine and ethanolamine inhibited TGase activation TGase activation rather than ROS generation in the retinas in endothelial cells. Polyamines, including cystamine, of diabetic mice. The primary amine group of cysteamine Journal putrecine, spermine and spermidine, are effective substrates was involved in VEGF-induced endothelial permeability for TGase and can be used as competitive TGase inhibitors and vascular leakage. The sulfhydryl group of cysteamine (Lentini et al. 2004, Lee et al. 2016); the monoamine MDC was important for inhibiting ROS generation, but not is a well-known TGase inhibitor (Lee et al. 2016, Bhatt et al. essential for TGase activation and vascular leakage. 2013). The transamidation of monoamines, including Thus, cysteamine has a potential for treating diabetic serotonin, dopamine and norepinephrine, is mediated by retinopathy and TGase-associated diseases. TGase (Hummerich et al. 2012). Thus, amine compounds, including monoamine and polyamines, might have potential as drugs for treating TGase-associated diseases. Declaration of interest The authors declare that there is no conflict of interest that could be In this study, intravitreal injection of cysteamine perceived as prejudicing the impartiality of the research reported. prevented hyperglycemia-induced TGase activation and vascular leakage in the retinas of diabetic mice.

Low-molecular-weight inhibitors prevent TGase Funding transamidating activity, and the most widely used TGase This work was supported by the National Research Foundation of Korea inhibitor is cystamine, the dimer form of cysteamine (2015R1A4A1038666 and 2016R1A2A1A05004975). (Basso & Ratan 2013). Previous clinical and preclinical trials using cystamine for treatment of neurodegenerative Author contribution statement diseases and cystic fibrosis suggest that cystamine could Y-J L and S-H J designed and performed experiments, analyzed data and be used for prevention of diabetic vascular leakage. wrote the manuscript. J Y H contributed to samples and analyzed data. Consistent with a previous report (Lee et al. 2016), we S J, E-T H, W-S P, S-H H and Y-M K designed experiments and analyzed data. K-S H conceptualized the study, designed experiments, analyzed and demonstrated that intravitreal injection of cysteamine interpreted the data and wrote the manuscript. All authors have approved prevents vascular leakage in diabetic mouse retinas. Thus, the final version of the manuscript.

References Hummerich R, Thumfart JO, Findeisen P, Bartsch D & Schloss P 2012 Transglutaminase-mediated transamidation of serotonin, dopamine Bacq ZM, Dechamps G, Fischer P, Herve A, Le Bihan H, Lecomte J, and noradrenaline to fibronectin: evidence for a general mechanism Pirotte M & Rayet P 1953 Protection against x-rays and therapy of monoaminylation. FEBS Letters 586 3421–3428. (doi:10.1016/j. of radiation sickness with beta-mercaptoethylamine. Science 117 febslet.2012.07.062) 633–636. (doi:10.1126/science.117.3049.633) Kaiser-Kupfer MI, Fujikawa L, Kuwabara T, Jain S & Gahl WA 1987 Basso M & Ratan RR 2013 Transglutaminase is a therapeutic target for Removal of corneal crystals by topical cysteamine in nephropathic oxidative stress, excitotoxicity and stroke: a new epigenetic kid on the cystinosis. New England Journal of Medicine 316 775–779. CNS block. Journal of Cerebral Blood Flow and Metabolism 33 809–818. (doi:10.1056/NEJM198703263161304) (doi:10.1038/jcbfm.2013.53) Lee YJ, Jung SH, Kim SH, Kim MS, Lee S, Hwang J, Kim SY, Kim YM & Ha Besouw M, Masereeuw R, van den Heuvel L & Levtchenko E 2013 KS 2016 Essential role of transglutaminase 2 in vascular endothelial Cysteamine: an old drug with new potential. Drug Discovery Today 18 growth factor-induced vascular leakage in the retina of diabetic mice. 785–792. (doi:10.1016/j.drudis.2013.02.003) Diabetes 65 2414–2428. (doi:10.2337/db15-1594) Bhatt MP, Lim YC, Hwang J, Na S, Kim YM & Ha KS 2013 C-peptide Lentini A, Abbruzzese A, Caraglia M, Marra M & Beninati S 2004 Protein- prevents hyperglycemia-induced endothelial apoptosis through polyamine conjugation by transglutaminase in cancer cell differentiation: inhibition of reactive oxygen species-mediated transglutaminase 2 review article. Amino Acids 26 331–337. (doi:10.1007/s00726-004-0079-3) activation. Diabetes 62 243–253. (doi:10.2337/db12-0293) Li R, Ren M, Chen N, Luo M, Zhang Z & Wu J 2012 Vitronectin increases Bhatt MP, Lim YC & Ha KS 2014 C-peptide replacement therapy as an vascular permeability by promoting VE-cadherin internalization at cell emerging strategy for preventing diabetic vasculopathy. Cardiovascular junctions. PLoS ONE 7 e37195. (doi:10.1371/journal.pone.0037195) Research 104 234–244. (doi:10.1093/cvr/cvu211) Lim YC, Bhatt MP, Kwon MH, Park D, Lee S, Choe J, Hwang J, Kim YM & Bhatt MP, Lee YJ, Jung SH, Kim YH, Hwang JY, Han ET, Park WS, Hong Ha KS 2014 Prevention of VEGF-mediated microvascular permeability SH, Kim YM & Ha KS 2016 C-peptide protects against hyperglycemic by C-peptide in diabetic mice. Cardiovascular Research 101 155–164. memory and vascular endothelial cell apoptosis. Journal of (doi:10.1093/cvr/cvt238) Endocrinology 231 97–108. (doi:10.1530/JOE-16-0349) Park D, Choi SS & Ha KS 2010 Transglutaminase 2: a multi-functional Buckley PF, Miller BJ, Pillai A & Kirkpatrick B 2014 Cysteamine, a pro-BDNF protein in multiple subcellular compartments. Amino Acids 39 drug, as an adjunctive treatment for schizophrenia. Schizophrenia 619–631. (doi:10.1007/s00726-010-0500-z) Research 158 268–269. (doi:10.1016/j.schres.2014.06.030) Pinto JT, Van Raamsdonk JM, Leavitt BR, Hayden MR, Jeitner TM, Thaler Caldwell RB, Bartoli M, Behzadian MA, El-Remessy AE, Al-Shabrawey M, Platt DH HT, Krasnikov BF & Cooper AJL 2005 Treatment of YAC128 mice & Caldwell RW 2003 Vascular endothelial growth factor and diabetic and their wild-type littermates with cystamine does not lead to its retinopathy: pathophysiological mechanisms and treatment perspectives. accumulation in plasma or brain: implications for the treatment Diabetes/Metabolism Research and Reviews 19 442–455. (doi:10.1002/dmrr.415) of Huntington disease. Journal of Neurochemistry 94 1087–1101. Cisbani G, Drouin-Ouellet J, Gibrat C, Saint-Pierre M, Lagace M, Badrinarayanan (doi:10.1111/j.1471-4159.2005.03255.x) S, Lavallee-Bourget MH, Charest J, Chabrat A, Boivin L, et al. 2015 Shannon KM & Fraint A 2015 Therapeutic advances in Huntington’s Endocrinology Cystamine/cysteamine rescues the dopaminergic system and shows disease. Movement Disorders 30 1539–1546. (doi:10.1002/mds.26331) of neurorestorative properties in an animal model of Parkinson’s disease. Shin YJ, Hyon JY, Kim S, Koh JW, Kwon SI & Wee WR 2011 Cysteamine Neurobiology Disease 82 430–444. (doi:10.1016/j.nbd.2015.07.012) suppresses human peripheral blood mononuclear cells – human De Stefano D, Villella VR, Esposito S, Tosco A, Sepe A, De Gregorio F, corneal endothelial cell reaction via reactive oxygen species Journal Salvadori L, Grassia R, Leone CA, De Rosa G, et al. 2014 Restoration of reduction. Molecular Vision 17 3371–3378. CFTR function in patients with cystic fibrosis carrying the F508del-CFTR Spindler V, Schlegel N & Waschke J 2010 Role of GTPases in control of mutation. Autophagy 10 2053–2074. (doi:10.4161/15548627.2014.973737) microvascular permeability. Cardiovascular Research 87 243–253. Dohil R, Schmeltzer S, Cabrera BL, Wang T, Durelle J, Duke KB, Schwimmer (doi:10.1093/cvr/cvq086) JB & Lavine JE 2011 Enteric-coated cysteamine for the treatment of Stewart MW 2016 Treatment of diabetic retinopathy: recent advances paediatric non-alcoholic fatty liver disease. Alimentary Pharmacology and and unresolved challenges. World Journal of Diabetes 7 333–341. Therapeutics 33 1036–1044. (doi:10.1111/j.1365-2036.2011.04626.x) (doi:10.4239/wjd.v7.i16.333) Fong DS, Aiello L, Gardner TW, King GL, Blankenship G, Cavallerano JD, Tolentino MS, Tolentino AJ & Tolentino MJ 2016 Current and Ferris FL 3rd & Klein R 2004 Retinopathy in diabetes. Diabetes Care 27 investigational drugs for the treatment of diabetic retinopathy. Expert (Supplement 1) S84–S87. (doi:10.2337/diacare.27.2007.S84) Opinion on Investigational Drugs 25 1011–1022. (doi:10.1080/1354378 Gibrat C & Cicchetti F 2011 Potential of cystamine and cysteamine 4.2016.1201062) in the treatment of neurodegenerative diseases. Progress in Tsilou ET, Thompson D, Lindblad AS, Reed GF, Rubin B, Gahl W, Neuro-Psychopharmacology and Biological Psychiatry 35 380–389. Thoene J, Del Monte M, Schneider JA, Granet DB, et al. 2003 (doi:10.1016/j.pnpbp.2010.11.023) A multicentre randomised double masked clinical trial of a new Gupta N, Mansoor S, Sharma A, Sapkal A, Sheth J, Falatoonzadeh P, formulation of topical cysteamine for the treatment of corneal cystine Kuppermann B & Kenney M 2013 Diabetic retinopathy and VEGF. Open crystals in cystinosis. British Journal of Ophthalmology 87 28–31. Ophthalmology Journal 7 4–10. (doi:10.2174/1874364101307010004) (doi:10.1136/bjo.87.1.28) Hammes HP 2005 Pericytes and the pathogenesis of diabetic retinopathy. Wang S, Park JK & Duh EJ 2012 Novel targets against retinal angiogenesis Hormone and Metabolic Research 37 (Supplement 1) 39–43. in diabetic retinopathy. Current Diabetes Reports 12 355–363. (doi:10.1055/s-2005-861361) (doi:10.1007/s11892-012-0289-0) He M, Pan H, Chang RC, So KF, Brecha NC & Pu M 2014 Activation of Wilmer MJ, Kluijtmans LA, van der Velden TJ, Willems PH, Scheffer the Nrf2/HO-1 antioxidant pathway contributes to the protective PG, Masereeuw R, Monnens LA, van den Heuvel LP & Levtchenko effects of Lycium barbarum polysaccharides in the rodent retina EN 2011 Cysteamine restores glutathione redox status in cultured after ischemia-reperfusion-induced damage. PLoS ONE 9 e84800. cystinotic proximal tubular epithelial cells. Biochimica et Biophysica (doi:10.1371/journal.pone.0084800) Acta 1812 643–651. (doi:10.1016/j.bbadis.2011.02.010)

Received in final form 25 July 2017 Accepted 27 July 2017 Accepted preprint published online 27 July 2017