Biochemical and Biophysical Research Communications 418 (2012) 420–425

Contents lists available at SciVerse ScienceDirect

Biochemical and Biophysical Research Communications

journal homepage: www.elsevier .com/locate /ybbrc

Hyperglycemia induces early upregulation of the calcium sensorKChIP3/DREAM/calsenilin in the rat retina

Erika Chavira-Suárez a, Alejandro Sandoval b, Heberto Quintero a, Prisca Bustamante a, Ricardo Felix c,Mónica Lamas a,⇑a Departamento de Farmacobiología, Centro de Investigación y de Estudios Avanzados del IPN (Cinvestav-IPN), Mexico Distrito Federal, Mexicob Carrera de Medicina, Facultad de Estudios Superiores Iztacala, Universidad Nacional Autónoma de México (UNAM), Tlalnepantla, Mexicoc Departamento de Biología Celular, Cinvestav-IPN, México Distrito Federal, Mexico

a r t i c l e i n f o

Article history:Received 6 January 2012Available online 18 January 2012

Keywords:KChIP3/DREAM/calsenilinRetinaHyperglycemiaKV4 channelsDiabetic retinopathy

0006-291X/$ - see front matter � 2012 Elsevier Inc. Adoi:10.1016/j.bbrc.2012.01.048

⇑ Corresponding author. Address: Departamento deIPN Sede Sur, Calzada de los Tenorios 235, Col. GraFederal, CP 14330, Mexico. Fax: +52 55 54832863.

E-mail address: mlamas@cinvestav.mx (M. Lamas)

a b s t r a c t

Hyperglycemia alters the tight control of intracellular calcium dynamics in retinal cells and may lead to thedevelopment of diabetic retinopathy. The potassium channel interacting protein 3 (KChIP3) also known asDREAM (Downstream Regulatory Element Antagonist Modulator) or calsenilin (KChIP3/DREAM/calseni-lin), a member of the neuronal calcium sensor protein family, is expressed in Müller glial cells and upreg-ulated under high glucose experimental culture conditions. Here, we analyzed the expression and functionof KChIP3 in the retina of streptozotocin induced diabetic Long Evans rats by immunofluorescence confocalmicroscopy, western blot, co-immunoprecipitation, whole cell patch clamp recording on isolated cells andKChIP3 gene silencing by RNA interference. Three weeks after streptozotocin application, KChIP3 wasincreased throughout the different retinal layers and this process was not linked to augmented apoptosis.KChIP3 co-immunoprecipitated with voltage gated K+ channels of the KV4.2–4.3 subtype in retinal extractsfrom control and hyperglycemic rats. Electrophysiological analysis showed that control cells did notexpress A type (KV4-mediated) K+ currents but most of the cells from streptozotocin treated retinas dis-played macroscopic currents with an inactivating component sensitive to 4-AP, suggesting the persistenceof the A type currents at early times after treatment. siRNA analysis in Müller cells cultures grown underhigh glucose experimental conditions corroborated that, when the expression of KChIP3 is 50% reduced,the number of cells expressing A type currents decreases significantly. Together these data suggest analtered expression and function of KChIP3 after streptozotocin induced hyperglycemia that might helpexplain some pathological alterations in early diabetic retinopathy.

� 2012 Elsevier Inc. All rights reserved.

1. Introduction

Glucose homeostasis plays a fundamental role in retinal func-tion and it has been well established that hyperglycemia triggersbiochemical abnormalities that may lead to diabetic retinopathy[1–3]. Increasing evidence suggests that hyperglycemia inducesalterations of the retinal intracellular calcium (Ca2+) dynamics,including changes in the expression and function of several mem-brane ion channels and Ca2+ binding proteins, which have beenassociated to distinct pathological processes [4–7].

The neuronal calcium sensor (NCS) protein family includesCa2+-binding proteins that are enriched or expressed only in thenervous system and neuroendocrine cells and have specializedfunctions [8]. Two members of the family, the recoverin and the

ll rights reserved.

Farmacobiología, Cinvestav-njas Coapa, Mexico, Distrito

.

guanylate cyclase (GC)-activating proteins (GCAP), are specificallyexpressed in the retina where they play fundamental roles in pho-totransduction [8]. Another NCS family member, KChIP3, alsoknown as DREAM or calsenilin, a multifunctional protein that actsin different cellular compartments to modulate cell survival andfunction [9–13], is also expressed in the retina where it was origi-nally associated with the regulation of rhythmically expressed cir-cadian genes in ganglion cells [14,15]. Recently we described thatKChIP3 is upregulated in cultures of retinal glial Müller cells keptunder high glucose (diabetic like) conditions and that these cellshave alterations in the inactivation of the macroscopic outwardK+ current associated to a specific protein–protein interaction be-tween KChIP3 and the voltage-gated K+ channels of the KV4.2/4.3subtype [16]. To further characterize the role of KChIP3 in the dia-betic retina, here we analyzed KChIP3 expression in control rat ret-inas and after chemical induction of experimental diabetes bystreptozotocin injection in neonatal animals. We found that3 weeks after toxin application KChIP3 was induced through theretinal layers and our functional analysis showed changes in the

E. Chavira-Suárez et al. / Biochemical and Biophysical Research Communications 418 (2012) 420–425 421

properties of the macroscopic inactivating K+ currents in thehyperglycemic retina Müller cells. In the light of these results wediscuss the possible function that KChIP3 might have at early timesin a rat model of diabetic retinopathy.

2. Materials and methods

All experiments were conducted on laboratory animals handledfollowing the guidelines of the internal animal care committee(CICUAL Cinvestav). Long Evans rats used in the experiments wereobtained from Harlan Sprague Dawley, maintained on a 12–12light–dark cycle with access to standard rat chow and water adlib and were euthanized using an overdose (80 mg/kg) of pentobar-bithal (Anestesal, Pfizer).

2.1. Induction of diabetes

Diabetes was induced in 5 day old Long Evans rats following amethod previously described [17]. Briefly, animals (10 ± 2 g) weretreated with a single i.p. injection of 80 mg/kg streptozotocin(Sigma Aldrich) freshly dissolved in sterile MilliQ H2O (pH 5) at aconcentration of 0.016 mg/ll. Control rats were given a single i.p.injection of H2O. Blood glucose tests were performed throughoutthe experiment using CU-Chek Active glucose test strips (RocheDiagnostics). Blood samples were weekly collected from the tailof rats in fed state at 9:00 a.m. Diabetes was confirmed by hyper-glycemia (glucose >200 mg/dl) for 2 consecutive weeks.

2.2. Immunostaining

Freshly dissected eyes from at least three different animals perexperiment were fixed in 4% (w/v) paraformaldehyde for 12 h at4 �C in PBS (137 mM NaCl, 2.7 mM KCl, 4.3 mM Na2PO4, 1.47 mMKH2PO4) and cryoprotected by immersion in 30% (w/v) sucrose inPBS. The tissue was embedded in Optimal Cutting Temperature(OCT) compound (Tissue Tek; Sakura Finetek) and stored at�20 �C. Twelve to twenty micrometers of sections were cut on acryostat, mounted onto Superfrost glass slides (Cole–Parmer) andstored at �20 �C. Control and experimental sections were mountedon the same slides and stained together. After thawing, the sliceswere incubated with 3% bovine serum albumin (BSA), 2% donkeyserum (Sigma Aldrich) or 2% goat serum (GIBCO BRL). Incubationswere performed with goat anti DREAM (1:50) (Santa Cruz Biotech-nology) or mouse anti CRALBP (1:100) (Affinity BioReagents). Sec-ondary incubations were performed with donkey anti-goat AlexaFluor 488 or anti mouse (1:200) (Invitrogen). The slides werecounterstained with 40,6-diamidino-2 phenylindole (DAPI) (VectorLaboratories). The images were processed in a fluorescence micro-scope (Axiovert 40 CFL, Zeiss) at 20� and 40� or in a confocalmicroscope (TCS-SP5-Tandem-MO-DI6000, Leica) with 2 photonsTI SAPH laser at 63 � 1.4 na. Confocal settings were kept constantwhen comparing retinal sections from control and hyperglycemicanimals. Figures show single optical sections of 0.35 lm thickness.Brightness and contrast of images were optimized with LeicaMicrosystems LAS AF Version 1.8.0.

2.3. Terminal transferase dUTP nick end labeling (TUNEL) assay

Apoptotic cell death was analyzed with in situ Cell DeathDetection Kit (Roche Diagnostics) following the manufacturer’sinstructions. As negative and positive controls the retinas wereincubated respectively with no antibody or pre-treated with200 u/ml DNase I (Sigma Aldrich) for 10 min and then processed.

2.4. Cell culture

Enucleated eyes from Long Evans rats were obtained as previ-ously described [15]. The technique employed is routinely usedas a source of Müller cell enriched cultures and has been previouslydemonstrated to enrich Müller cells to the range of >95% free fromneuronal, fibroblastic or astrocytic contamination [18–20]. Forelectrophysiology and for experiments using small interferingRNAs (siRNAs), the cells obtained from control or diabetic ratswere kept in culture in medium supplemented with 25 mM glu-cose for 5 days prior to the experiments.

2.5. Western blot

Total protein extracts from whole retinas were homogenized in2� Laemmli Buffer, separated on 10–12% SDS–polyacrylamide gelsand transferred to PDVF membranes (Immobilon-P, Millipore). Themembranes were blocked as previously described [16] and incu-bated with mouse anti calsenilin/DREAM (Upstate) or mouse antiactin clone C4 (Millipore). After that, the membranes were incu-bated with HRP anti mouse (1:20,000) (Jackson ImmunoResearch).Immunoreactivity was detected as previously described [16]. Theimmunoreactive bands from non-saturated films were scannedand quantified by densitometry with LabWorks 4.5 software (UVP).

2.6. Co-immunoprecipitation

Six retinas were pooled for each condition, washed with coldPBS and centrifuged. The pellets were lysated with lysis buffer(50 mM TRIS–HCl, pH 7.4; 1% NP-40; 0.25% sodium deoxycholate;150 mM NaCl, 1 mM EDTA, 1 mM PMSF; 1 lg/ml each aprotinin,leupeptin, pepstatin; 1 mM NaF, 1 mM Na3VO4) for 1 h on ice.The supernatants were incubated in protein A-Sepharose (GEHealthcare Ltd.) with rabbit anti KV4.2/4.3 (Santa Cruz Biotechnol-ogy), mouse anti calsenilin/DREAM (Upstate) or an antibodyagainst an unrelated protein, Sox9, (Santa Cruz Biotechnology). Im-mune complexes were resuspended in 2� Laemmli buffer and sub-jected to SDS PAGE and Western Blot.

2.7. Electrophysiology

Cells were subject to the whole-cell variant of the patch-clamptechnique as described elsewhere [21,16]. Experiments were car-ried out at the holding potential (Vh) of �80 mV. Currents were re-corded with an Axopatch 200B amplifier, filtered at 2 kHz anddigitized at 5.71 kHz using a DigiData 1320 interface. Data wereanalyzed using the pCLAMP and SigmaPlot software. Linear capaci-tative currents were minimized analogically. The remaining linearcomponents were subtracted using a P/4 protocol. Membranecapacitance (Cm) was determined as described previously [21].Macroscopic K+ currents were acquired in response to 200 ms testpulses of variable amplitude. Recording solutions consisted of (inmM): 140 NaCl, 3 KCl, 5 CaCl2, 10 HEPES, 5 glucose (external solu-tion) and 130 K-Asp, 8 KCl, 0.2 MgCl2, 10 EGTA, 5 Na-ATP (internalsolution). All data are presented as mean ± SEM. Statistical differ-ences were significant at the p < 0.05 value, (unpaired t-test) withSystat Software version 3.11.

2.8. Small interfering RNAs (siRNAs)

To silence KChIP3 gene expression, �2 � 106 Müller cells cul-tured under high glucose (25 mM) conditions were transfectedwith 30 nM KChIP3 specific siRNA (Sense: GGACGGUUAC-AUCACCAAAtt, antisense: UUUGGUGAUGUAACCGUCCtt. Cat. #4390771, ID # s134769; Ambion), an unspecific siRNA ‘‘Scramble’’(Cat. # AM4611; Ambion) or mock transfected, using 40 ll of the

Fig. 1. Streptozotocin-induced changes in neonatal Long Evans rats. Quantification of blood glucose (A) and weight (B), in control (open circles) and streptozotocin treated(filled circles) rats. Values are mean ± SD (n = 12). ⁄p < 0.033. (C and D) Representative epifluorescence photomicrographs of paraformaldehyde fixed rat retinas from DNase1treated, control or hyperglycemic rats 3 weeks or 3 months after streptozotocin injection (HG) double stained with DAPI and TUNEL(C) or from control or hyperglycemic (HG)rats 3 weeks after injection, stained with a GFAP antibody (red) and DAPI (blue) (D). IPL: inner plexiform layer; INL: inner nuclear layer; OPL: outer plexiform layer; ONL:outer nuclear layer; NFL: nerve fiber layer. Scale bar = 50 lm. Arrows indicate TUNEL immunopositive cells. (For interpretation of the references to color in this figure legend,the reader is referred to the web version of this article.)

Fig. 2. Hyperglycemia increases KChIP3 expression in rat retinas. (A) Representative confocal microscopy images of tissue sections from a control or a hyperglycemic rat (HG)retina 3 weeks after streptozotocin injection stained with KChIP3 (green), CRALBP (red) antibodies and DAPI (blue). GCL: ganglion cell layer; IPL: inner plexiform layer; INL:inner nuclear layer; OPL: outer plexiform layer; ONL: outer nuclear layer; PR: photoreceptor region; PE: pigmented epithelium. Scale bar = 50 lm. (B) Western blots anddensitometric analysis of KChIP3 and actin protein levels in control and hyperglycemic (HG) rat single retinas. The graph represents the level of KChIP3 signal relative to actinin each retina. (C) Co-immunoprecipitation (IP) from six pooled control (C) and six pooled hyperglycemic (HG) retinas using KChIP3 or KV4.2/4.3 antibodies, followed byimmunobloting (WB) with KV4.2/4.3 or KChIP3 antibodies respectively. As a control, an antibody against SOX9 was used in the immunoprecipitation assays. HCh denotesHeavy chain (immunoglobulin loading control). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

422 E. Chavira-Suárez et al. / Biochemical and Biophysical Research Communications 418 (2012) 420–425

transfection agent siPORT NeoFx (Ambion). Twenty-four hoursafter transfection the medium was replaced with fresh OPTIMEMsupplemented with glucose (25 mM). Forty-eight hours aftertransfection the cells were collected for analysis of gene expres-sion. Examination of mock transfected cells indicated that thetransfection process did not cause cytotoxicity.

3. Results

3.1. High glucose increases KChIP3 expression in the retina

To characterize the effect of hyperglycemia on KChIP3 retinalexpression we employed an experimental model of chemical

Fig. 3. K+ currents (IK) in the early hyperglycemic retina. (A) Representative wholecell K+ currents (IK) recorded from control Müller cells (upper trace) and cellsisolated from hyperglycemic rat retinas 3 weeks after streptozotocin injection(middle trace). Current traces in the lower panel correspond to recordings fromstreptozotocin treated cells before and after application of 4-AP (5 mM). Calibrationbar: 50 pA and 50 ms. (B) Comparison of the percentage of cells expressing thetransient and/or the long lasting components of IK. (C) Average Idensity-V relation-ships for Müller cells isolated from control and streptozotocin treated rats.

E. Chavira-Suárez et al. / Biochemical and Biophysical Research Communications 418 (2012) 420–425 423

induction of diabetes by streptozotocin injection in neonatal rats.Three weeks after injection of streptozotocin, the treated animalsshowed a significant increase in plasma glucose levels from107 ± 24 in the control to 294 ± 96 mg/dl in the hyperglycemiccondition (HG), associated with a �20% decrease in body weight(Fig. 1A and B). To corroborate that streptozotocin was triggeringretinal damage we evaluated the appearance of apoptotic TUNELpositive cells 3 weeks and 3 months after toxin injection (Fig. 1C)and the expression of the intermediate filament protein glial fibril-lary acid protein (GFAP), an early marker for retinal injury [22–24](Fig. 1D). Our results show no TUNEL immunopositive cells at earlytimes (3 weeks) in any condition. In contrast, at later times(3 months) a number of apoptotic cells were observed in the strep-tozotocin treated animals. However, 3 weeks after streptozotocininjection, upregulation of GFAP expression was already detectablein the Nerve Fiber Layer (NFL) of the retina from treated animals(Fig. 1D).

We then analyzed retinal KChIP3 expression by immunohisto-chemistry and confocal microscopy. Three weeks after toxin appli-cation, KChIP3 was slightly induced through the ONL, OPL and INLretinal layers and strongly induced in the IPL/GCL and photorecep-tor (PR) layers (Fig. 2A). Co-immunolabeling with the cellularretinaldehyde binding protein (CRALBP), a molecular marker thatrecognizes Müller cells, astrocytes and the retinal pigmentedepithelium, indicated that KChIP3 expression in the retina is notrestricted to Müller cells (Fig. 2A). We corroborated and quantifiedthe induction of KChIP3 expression by analyzing total protein ex-tracts from control and diabetic rat retinas by Western blot using

a specific antibody that recognizes a unique 30 KD protein band(Fig. 2B).

3.2. Müller cell KV4 channel function in the early hyperglycemic ratretina

These and our previous results showing that high glucoseexperimental conditions led to the modulation of A type K+ cur-rents in cultured Müller cells [16], prompted us to analyze Müllercell KV4 channel function in the early hyperglycemic rat retina. Asemi-quantitative assay of 6 pooled retinas with KChIP3 orKV4.2/4.3 antibodies led to co-immunoprecipitation of KV4.2/4.3in the first case and KChIP3 in the reverse procedure, both in con-trol and streptozotocin conditions (Fig. 2C). The specificity of thereaction was confirmed with a non-related antibody.

We compared the functional properties of KV4 channel currents(IK) recorded in Müller cells isolated from retinas of control anddiabetic rats 3 weeks after the injection of streptozotocin (Fig. 3).IK was elicited by depolarization from a Vh of�80 mV to test poten-tials ranging from �60 to +40 mV. In the control condition most ofthe cells (�83%) did not display A type K+ currents (Fig. 3A and B).In contrast, after 3 weeks of increased blood glucose, IK was rapidlyactivated in �88% of the cells on depolarization to +40 mV, andconsisted of a transient component that resembles the A type cur-rent and a steady state component similar to the non- or slowlyinactivating delayed rectifier current (Fig. 3A, medium trace). Extra-cellular application of 5 mM 4-AP blocked the transient componentof IK (Fig. 3A, lower traces), presumably due to the inhibitory effectof the drug on KV4 channels. We next compared the level of chan-nel expression by analyzing IK density over a range from �60 to+40 mV. Current density was calculated as the ratio of current ateach voltage step divided by the cell capacitance. We found asignificant increase in IK density in cells from the streptozotocintreated animals at voltages above +10 mV (Fig. 3C).

Last, we analyzed the consequences of exposing the cells to aspecific KChIP3 siRNA by transfection of Müller cells culturedunder high glucose conditions. This procedure led, 48 h after trans-fection, to an approximately 50% reduction in the expression ofKChIP3 that was specific, as transfection with an unrelated siRNA(scramble) was ineffective in modifying KChIP3 levels (Fig. 4A).KV4.2/4.3 channel expression was also significantly reduced inKChIP3 siRNA treated cells (Fig. 4B). Likewise, KV4 channel currents(IK) recorded in Müller cells under high glucose conditions indi-cated that >60% of the cells displayed the rapidly inactivating Atype currents (Fig. 4C and D). However, when KChIP3 expressionis reduced by the presence of KChIP3 siRNA, a smaller number ofthe recorded cells (<10%) exhibited the fast component of IK

(Fig. 4C and D), although IK density over a wide range of potentialsranging from �60 to +40 mV in cells cultured in high glucose con-ditions in the presence of a KChIP3 siRNA or the scrambled siRNAwas similar (Fig. 4E).

4. Discussion

We recently reported the upregulation of KChIP3/DREAM/calse-nilin expression in Müller cells kept in culture under high glucoseexperimental conditions [16]. Being a pleiotropic protein that con-tributes to integrate intracellular signaling cascades, we speculatedthat KChIP3 could have an impact in the progression of diabetic ret-inopathy. Here, using a previously described method that promotesa significant increase in plasma glucose levels, slight decrease inweight gain and sustained insulin resistance that resembles the dia-betes mellitus development in humans [17,25,26], we searched forearly retinal alterations after the onset of hyperglycemia. Threeweeks after STZ injection we observed upregulation of the retinal

Fig. 4. siRNA-mediated suppression of KChIP3. (A and B) Western blots and densitometric analysis of KChIP3 or KV4.2/4.3 protein levels in high glucose cultured Müller cellstransfected with non specific siRNA (Scramble), KChIP3 siRNA, or mock transfected. The graph represents KChIP3 signal relative to actin in KChIP3 siRNA and scrambletransfected cells. ⁄p < 0.05. (C) Representative macroscopic K+ currents recorded from cultured Müller cells in HG conditions transfected with scrambled (Scr) or KChIP3 (KCh)siRNAs. (D) Comparison of the percentage of cells expressing the transient and/or the long lasting components of IK. (E) Average Idensity-V relationships for Müller cellsrecorded in the control condition and after KChIP3 knockdown.

424 E. Chavira-Suárez et al. / Biochemical and Biophysical Research Communications 418 (2012) 420–425

expression of GFAP, a well known marker of reactive gliosis, indi-cating that reactive changes in the retina have started to developat this early time. Interestingly, we show that the expression ofKChIP3 is increased at early times in this model and that thisincrease is not linked to augmented retinal apoptosis but it is asso-ciated to the persistence of the A type K+ currents in Müller glialcells from streptozotocin treated animals.

KV4.2–4.3 channels, the primary contributors of transient out-ward voltage gated K+ currents (known also as A currents) in thenervous system [27], have been previously described as targetsof KChIP proteins in astrocytes [28] and Müller cells [16]. Theeffects of the interaction between KChIP3 and KV4 channels includean increase in the current density [12,27–30] and specific changesin current kinetics [27]. Corroborating previous reports thatestablished a developmental downregulation of this current indifferentiated rat Müller cells [31], we found that Müller cells fromnon-treated animals did not express A type currents. In contrast,our data showed that >85% of the cells derived from streptozotocintreated retinas display macroscopic currents with an inactivatingcomponent sensitive to 4-AP. These results and the functionalinvolvement of KChIP3 in the modulation of A type K+ currentsin Müller cells were further corroborated using the alternativeapproach of silencing KChIP3 gene expression using specificsiRNAs. As expected, disruption of KChIP3 expression significantlyreduced the number of cells expressing A type K+ currents.

Taking in account that the hyperglycemia mediated induction ofexpression of KChIP3 is not restricted to Müller cells but rather

evident through all retinal layers we suggest that KChIP3 may af-fect independent physiological processes in different retinal celltypes that include, among others, the modulation of cellular mem-brane excitability through functionally interacting with the voltagegated K+ channels of the KV4 class [29]. Importantly, an upregula-tion of inactivating K+ currents has been also shown under otherpathological conditions of the retina [32–34], suggesting that thisprocess could be linked to the development of retinal abnormali-ties in the course of various diseases. The current notion of diabeticretinopathy implies that non vascular abnormalities in neuronsand glial cells, tightly associated to metabolic abnormalities causedby hyperglycemia, not only precede the development of the vascu-lar changes that characterize the disease advanced stages but alsomay constitute the molecular mechanisms that trigger the exacer-bation of the pathology [4–6,35–37]. When considered as anearly hyperglycemia mediated retinal abnormality, the character-ization of KChIP3 as a relevant molecular player in diabeticretinopathy could have a positive impact in diagnosis andtreatment.

Acknowledgments

We thank I.J. Galván-Mendoza from the Confocal and Multipho-ton Microscopy Unit, Cinvestav, and A. Huerta for assistance. Sup-ported by grants from Conacyt to ML and RF. EC-S, HQ and PBreceived two doctoral fellowships from Conacyt and an undergrad-uate fellowship form FARMED Conacyt respectively.

E. Chavira-Suárez et al. / Biochemical and Biophysical Research Communications 418 (2012) 420–425 425

References

[1] C.J. Chiu, A. Taylor, Dietary hyperglycemia, glycemic index and metabolicretinal diseases, Prog. Retin. Eye Res. 30 (2011) 18–53.

[2] S.A. Madsen-Bouterse, R.A. Kowluru, Oxidative stress and diabetic retinopathy:pathophysiological mechanisms and treatment perspectives, Rev. Endocr.Metab. Disord. 9 (2008) 315–327.

[3] Diabetes Control and Complications Trial Research Group, The effect ofintensive treatment of diabetes on the development of long-termcomplications in insulin-dependent diabetes mellitus, N. Engl. J. Med. 329(1993) 977–986.

[4] A.R. Santiago, S.C. Rosa, P.F. Santos, et al., Elevated glucose changes theexpression of ionotropic glutamate receptor subunits and impairs calciumhomeostasis in retinal neural cells, Invest. Ophthalmol. Vis. Sci. 47 (2006)4130–4137.

[5] G. Costa, T. Pereira, A.M. Neto, et al., High glucose changes extracellularadenosine triphosphate levels in rat retinal cultures, J. Neurosci. Res. 87 (6)(2009) 1375–1380.

[6] T. de O. Pereira, G.N. da Costa, A.R. Santiago, et al., High glucose enhancesintracellular Ca2+ responses triggered by purinergic stimulation in retinalneurons and microglia, Brain Res. 1316 (2010) 129–138.

[7] Y.K. Ng, X.X. Zeng, E.A. Ling, Expression of glutamate receptors and calcium-binding proteins in the retina of streptozotocin-induced diabetic rats, BrainRes. 1018 (2004) 66–72.

[8] R.D. Burgoyne, J.L. Weiss, The neuronal calcium sensor family of Ca2+-bindingproteins, Biochem. J. 353 (2001) 1–12.

[9] C. Lilliehook, S. Chan, E.K. Choi, et al., Calsenilin enhances apoptosis by alteringendoplasmic reticulum calcium signaling, Mol. Cell Neurosci. 19 (4) (2002)552–559.

[10] L. Fedrizzi, D. Lim, E. Carafoli, et al., Interplay of the Ca2+-binding proteinDREAM with presenilin in neuronal Ca2+ signaling, J. Biol. Chem. 283 (41)(2008) 27494–27503.

[11] A.M. Carrion, W.A. Link, F. Ledo, et al., DREAM is a Ca2+-regulatedtranscriptional repressor, Nature 398 (1999) 80–84.

[12] A. Ruiz-Gomez, B. Mellstrom, D. Tornero, et al., G protein-coupled receptorkinase 2-mediated phosphorylation of downstream regulatory elementantagonist modulator regulates membrane trafficking of Kv4.2 potassiumchannel, J. Biol. Chem. 282 (2007) 1205–1215.

[13] Y. Zhang, P. Su, P. Liang, et al., The DREAM protein negatively regulates theNMDA receptor through interaction with the NR1 subunit, J. Neurosci. 30 (22)(2010) 7575–7586.

[14] W.A. Link, F. Ledo, B. Torres, et al., Day–night changes in downstreamregulatory element antagonist modulator/potassium channel interactingprotein activity contribute to circadian gene expression in pineal gland, J.Neurosci. 24 (2004) 5346–5355.

[15] E. Chavira-Suarez, M. Ramirez, M. Lamas, D-Serine/N-methyl-D-aspartatereceptor signaling decreases DNA-binding activity of the transcriptionalrepressor DREAM in Muller glia from the retina, Neurosci. Lett. 432 (2008)121–126.

[16] E. Chavira-Suárez, A. Sandoval, R. Felix, et al., Expression and high glucose-mediated regulation of K(+) channel interacting protein 3 (KChIP3) and K(V)4channels in retinal Müller glial cells, Biochem. Biophys. Res. Commun. 404 (2)(2011) 678–683.

[17] J. Takada, M.A. Machado, S.B. Peres, et al., Neonatal streptozotocin-induceddiabetes mellitus: model of insulin resistance associated with loss of adiposemass, Metabolism 56 (2007) 977–984.

[18] D. Hicks, Y. Courtois, The growth and behaviour of rat retinal Müller cellsin vitro. 1. An improved method for isolation and culture, Exp. Eye Res. 51(1990) 119–129.

[19] A.V. Das, K.B. Mallya, X. Zhao, et al., Neural stem cell properties of Muller gliain the mammalian retina: regulation by Notch and Wnt signaling, Dev. Biol.299 (2006) 283–302.

[20] M. Lamas, I. Lee-Rivera, M. Ramirez, et al., D-serine regulates CREBphosphorylation induced by NMDA receptor activation in Muller glia fromthe retina, Neurosci. Lett. 427 (2007) 55–60.

[21] G. Avila, A. Sandoval, R. Felix, Intramembrane charge movement associatedwith endogenous K+ channel activity in HEK-293 cells, Cell. Mol. Neurobiol. 24(2004) 317–330.

[22] A. Bignami, D. Dahl, The radial glia of Muller in the rat retina and theirresponse to injury. An immunofluorescence study with antibodies to the glialfibrillary acidic (GFA) protein, Exp. Eye Res. 28 (1979) 63–69.

[23] A.J. Eisenfeld, A.H. Bunt-Milam, P.V. Sarthy, Muller cell expression of glialfibrillary acidic protein after genetic and experimental photoreceptordegeneration in the rat retina, Invest. Ophthalmol. Vis. Sci. 25 (1984) 1321–1328.

[24] A. Bringmann, T. Pannicke, J. Grosche, et al., Muller cells in the healthy anddiseased retina, Prog. Retin. Eye Res. 25 (2006) 397–424.

[25] H.E. Wey, M.T. Subbiah, Streptozotocin-induced diabetes in the neonatal rat:effect on plasma lipids and aortic prostaglandin synthesis in adult life,Biochem. Med. 31 (2) (1984) 174–184.

[26] O. Blondel, B. Portha, Early appearance of in vivo insulin resistance in adultstreptozotocin-injected rats, Diabet. Metab. 15 (6) (1989) 382–387.

[27] S.G. Birnbaum, A.W. Varga, L.L. Yuan, et al., Structure and function of Kv4-family transient potassium channels, Physiol. Rev. 84 (2004) 803–833.

[28] T.Y. Nakamura, D.J. Pountney, A. Ozaita, et al., A role for frequenin, a Ca2+-binding protein, as a regulator of Kv4 K+-currents, Proc. Natl. Acad. Sci. USA 98(2001) 12808–12813.

[29] W.F. An, M.R. Bowlby, M. Betty, et al., Modulation of A-type potassiumchannels by a family of calcium sensors, Nature 403 (2000) 553–556.

[30] R. Shibata, H. Misonou, C.R. Campomanes, et al., A fundamental role for KChIPsin determining the molecular properties and trafficking of Kv4.2 potassiumchannels, J. Biol. Chem. 278 (2003) 36445–36554.

[31] A. Wurm, T. Pannicke, I. Iandiev, et al., The developmental expression of K+

channels in retinal glial cell is associated with a decrease of osmotic cellswelling, Glia 54 (2006) 411–423.

[32] T. Pannicke, I. Iandiev, A. Wurm, et al., Diabetes alters osmotic swellingcharacteristics and membrane conductance of glial cell in rat retina, Diabetes55 (2006) 633–639.

[33] T. Pannicke, O. Uckermann, I. Iandiev, et al., Altered membrane physiology inMüller glial cells after transient ischemia of the rat retina, Glia 50 (2005) 1–11.

[34] M. Rehak, M. Hollborn, I. Iandiev, et al., Retinal gene expression and Müller cellresponses after branch retinal vein occlusion in the rat, Invest. Ophthalmol.Vis. Sci. 50 (2009) 2359–2367.

[35] H. Sakai, Y. Tani, E. Shirasawa, et al., Development of electroretinographicalterations in streptozotocin-induced diabetes in rats, Ophthalmic. Res. 27(1995) 57–63.

[36] A.J. Barber, E. Lieth, S.A. Khin, et al., Neural apoptosis in the retina duringexperimental and human diabetes. Early onset and effect of insulin, J. Clin.Invest. 102 (4) (1998) 783–791.

[37] E. Lieth, A. Barber, B. Xu, et al., Glial reactivity and impaired glutamatemetabolism in short-term experimental diabetic retinopathy, Diabetes 47(1998) 815–820.