Differential effects of P2Y1 deletion on glial activation and survival of photoreceptors and amacrine cells in the ischemic mouse retina

Abstract

Gliosis of retinal Müller glial cells may have both beneficial and detrimental effects on neurons. To investigate the role of purinergic signaling in ischemia-induced reactive gliosis, transient retinal ischemia was evoked by elevation of the intraocular pressure in wild-type (Wt) mice and in mice deficient in the glia-specific nucleotide receptor P2Y ₁ (P2Y₁ receptor-deficient (P2Y1R-KO)). While control retinae of P2Y1R-KO mice displayed reduced cell numbers in the ganglion cell and inner nuclear layers, ischemia induced apoptotic death of cells in all retinal layers in both, Wt and P2Y1R-KO mice, but the damage especially on photoreceptors was more pronounced in retinae of P2Y1R-KO mice. In contrast, gene expression profiling and histological data suggest an increased survival of amacrine cells in the postischemic retina of P2Y1R-KO mice. Interestingly, measuring the ischemia-induced downregulation of inwardly rectifying potassium channel (Kir)-mediated K ⁺ currents as an indicator, reactive Müller cell gliosis was found to be weaker in P2Y1R-KO (current amplitude decreased by 18%) than in Wt mice (decrease by 68%). The inner retina harbors those neurons generating action potentials, which strongly rely on an intact ion homeostasis. This may explain why especially these cells appear to benefit from the preserved Kir4.1 expression in Müller cells, which should allow them to keep up their function in the context of spatial buffering of potassium. Especially under ischemic conditions, maintenance of this Müller cell function may dampen cytotoxic neuronal hyperexcitation and subsequent neuronal cell loss. In sum, we found that purinergic signaling modulates the gliotic activation pattern of Müller glia and lack of P2Y ₁ has janus-faced effects. In the end, the differential effects of a disrupted P2Y₁ signaling onto neuronal survival in the ischemic retina call the putative therapeutical use of P2Y₁ -antagonists into question.

OPEN

Citation: Cell Death and Disease (2014) 5, e1353; doi:10.1038/cddis.2014.317

& 2014 Macmillan Publishers Limited All rights reserved 2041-4889/14 http://www.nature.com/cddis

Web End =www.nature.com/cddis

T Pannicke1, I Frommherz1, B Biedermann1, L Wagner1, K Sauer1, E Ulbricht1,2, W Hartig1, U Krgel3, U Ueberham1, T Arendt1, P Illes3, A Bringmann4, A Reichenbach1 and A Grosche*,1,5

Gliosis of retinal Mller glial cells may have both benecial and detrimental effects on neurons. To investigate the role of purinergic signaling in ischemia-induced reactive gliosis, transient retinal ischemia was evoked by elevation of the intraocular pressure in wild-type (Wt) mice and in mice decient in the glia-specic nucleotide receptor P2Y1 (P2Y1 receptor-decient (P2Y1R-KO)). While control retinae of P2Y1R-KO mice displayed reduced cell numbers in the ganglion cell and inner nuclear layers, ischemia induced apoptotic death of cells in all retinal layers in both, Wt and P2Y1R-KO mice, but the damage especially on photoreceptors was more pronounced in retinae of P2Y1R-KO mice. In contrast, gene expression proling and histological data suggest an increased survival of amacrine cells in the postischemic retina of P2Y1R-KO mice. Interestingly, measuring the ischemia-induced downregulation of inwardly rectifying potassium channel (Kir)-mediated K currents as an indicator, reactive Mller cell gliosis was found to be weaker in P2Y1R-KO (current amplitude decreased by 18%) than in Wt mice (decrease by 68%). The inner retina harbors those neurons generating action potentials, which strongly rely on an intact ion homeostasis. This may explain why especially these cells appear to benet from the preserved Kir4.1 expression in Mller cells, which should allow them to keep up their function in the context of spatial buffering of potassium. Especially under ischemic conditions, maintenance of this Mller cell function may dampen cytotoxic neuronal hyperexcitation and subsequent neuronal cell loss. In sum, we found that purinergic signaling modulates the gliotic activation pattern of Mller glia and lack of P2Y1 has janus-faced effects. In the end, the differential effects of a disrupted P2Y1 signaling onto neuronal survival in the ischemic retina call the putative therapeutical use of P2Y1-antagonists into question.

Cell Death and Disease (2014) 5, e1353; doi:http://dx.doi.org/10.1038/cddis.2014.317

Web End =10.1038/cddis.2014.317 ; published online 31 July 2014

Glial cells are crucially involved in the maintenance of neuronal activity in nervous tissues.1 The homeostasis of the extra-cellular space is regulated by various glial functions including spatial K buffering, cell volume regulation and uptake of neurotransmitters.24 Activation of membrane receptors and ion channels is critically implicated in mediating the neuron-supportive glial functions. The dominant K conductance of glial cells mediates spatial K buffering and is important for the very negative membrane potential of these cells, thereby supporting electrogenic membrane transporters.5 Alterations in glial function are characteristic for pathological processes of the nervous system.6 Reactive gliosis may have benecial and detrimental effects and is considered as an attempt to maintain neuronal function, protecting the tissue from further destruction, and to initiate tissue regeneration.7,8 However, reactive gliosis may cause secondary neuronal damage as major neuron-supportive functions of glial cells get lost.6

Gliotic alterations of Mller cells, the dominant macroglia of the vertebrate retina, have been observed in various models of

Abbreviations: AC, amacrine cell; AD, adenosine; cAMP, 30-50-cyclic adenosine monophosphate; CRALBP, cellular retinaldehyde-binding protein; Cy, carbocyanine; GCL, ganglion cell layer; GFAP, glial brillary acidic protein; Glu, glutamate; HEPES, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid; HIOP, high intraocular pressure; Iba1, ionized calcium binding adaptor molecule 1; INL, inner nuclear layer; iNOS, inducible nitric oxide synthase; i.p., intraperitoneal; IPL, inner plexiform layer; Kir, inwardly rectifying potassium channel; LSM, laser scanning microscope; ONL, outer nuclear layer; PKCa, protein kinase Ca; P2Y1R, P2Y1 receptor; P2Y1R-KO, P2Y1 receptor-decient; TGF-b3, transforming growth factor-b3; Tris, tris(hydroxymethyl)aminomethane; TUNEL, terminal deoxynucleotidyl transferase-dUTP nick end labeling; Wt, wild type

Received 06.5.14; revised 02.6.14; accepted 04.6.14; Edited by A Verkhratsky

Differential effects of P2Y1 deletion on glial activation and survival of photoreceptors and amacrine cellsin the ischemic mouse retina

retinal diseases.9,10 A prominent feature of Mller cell gliosis is the downregulation of the inwardly rectifying K conductance mediated by inwardly rectifying K (Kir) channels.9 It has been demonstrated in astrocytes that downregulation or conditional knockout of Kir4.1 results in an impairment of glial glutamate (Glu) uptake.11,12 In addition, it has been suggested that autocrine/paracrine purinergic signaling may have a causative role in the development of reactive gliosis in brain and retina.13,14 Mller cells express different subtypes of P2

nucleotide receptors including P2Y1 and P2Y4.15,16 P2Y1 receptors have been demonstrated to be functionally expressed by Mller cells and microglial cells, rather than by neurons.1518

Retinal ischemia, a characteristic of various important human blinding diseases including diabetic retinopathy, results in neuronal degeneration and reactive gliosis.19,20 The reduced

K permeability of Mller cell membranes is associated with an impaired cell volume regulation under hypoosmotic stress after high intraocular pressure (HIOP)-induced ischemia.21 It has

been observed that tandem-pore domain K channels may

1Paul-Flechsig-Institut fr Hirnforschung, Universitat Leipzig, Leipzig, Germany; 2Biotechnology Center, Technische Universitat Dresden, Dresden, Germany;

3Rudolf-Boehm-Institut fr Pharmakologie und Toxikologie, Universitat Leipzig, Leipzig, Germany; 4Klinik und Poliklinik fr Augenheilkunde, Universitat Leipzig, Leipzig, Germany and 5Institut fr Humangenetik, Universitat Regensburg, Regensburg, Germany*Corresponding author: A Grosche, Institut fr Humangenetik, Universitat Regensburg, Franz-Josef-Strau-Allee 11, 93053 Regensburg, Germany. Tel: [phone]; Fax: [phone]; E-mail: mailto:[email protected]

Web End [email protected]

Role of P2Y1 signaling in ischemic mouse retina T Pannicke et al

2

fulll certain functions under conditions where Kir channels are downregulated or lacking.22,23 A malfunctional Mller cell volume regulation was also found after deletion of P2Y1 in

the mouse retina.16 It has been suggested that impaired glial K buffering and cell volume regulation may contribute to neuronal degeneration in the ischemic retina by inducing neuronal hyperexcitation and Glu-induced cell death.14 In order

to determine whether endogenous purinergic signaling is implicated in mediating and/or protecting from neuronal degeneration, we investigated the effects of HIOP-induced ischemia in the retinae of P2Y1-decient mice.

Results

Retinal morphology. In order to compare the cell numbers between retinae from Wt and P2Y1 receptor-decient (P2Y1R-KO) mice, we counted cell nuclei in the ganglion cell layer (GCL), inner nuclear layer (INL) and outer nuclear layer (ONL) in TO-PRO-3-labeled retinal slices from untreated eyes within a 100-mm-wide area close to the optic nerve head (Figure 1a). As shown in Figure 1b, the retina of

P2Y1R-KO mice displayed signicantly reduced numbers of cell nuclei in the GCL and INL as compared with the Wt control retina. Moreover, the inner plexiform layer (IPL) of untreated P2Y1R-KO mice was signicantly thinner than the IPL of Wt animals (Figure 1c).

Retinal ischemia-reperfusion results in neuronal degeneration.19,20 In order to determine ischemic retinal

degeneration, we induced transient retinal ischemia in mice for 90 min and quantied the number of cell nuclei in retinal slices 7 days after reperfusion. Transient ischemia caused signicant reductions in the numbers of cell nuclei in retinae of Wt and P2Y1R-KO mice compared with untreated controls (Figure 1b). Interestingly, the number of cell (photoreceptor) nuclei in the ONL was more reduced (Po0.05) in P2Y1R-KO mice than in

Wt retinae (Figure 1b). Although the thickness of the IPL was signicantly reduced in the postischemic retinae of Wt and P2Y1R-KO mice, the decrease in the thickness of the IPL was signicantly less pronounced in P2Y1R-KO mice (Figure 1c).

Apoptotic cell death was determined with terminal deoxynucleotidyl transferase-dUTP nick end labeling (TUNEL) labeling of retinal slices obtained 1 day after ischemia, the presumed time point of the peak apoptotic activity in the HIOP-induced rodent retinal ischemia model.24 No TUNEL-

labeled cell nuclei were found in untreated retinae, neither of Wt nor of P2Y1R-KO mice (Figure 1d). Between 40 and 50% of the cell nuclei in the INL were found to be TUNEL labeled in the postischemic retinae of Wt and P2Y1R-KO mice (Figure 1e). A similar percentage of cells was TUNEL positive in the ONL of Wt mice, whereas this number was signicantly increased in the ONL of P2Y1R-KO (Po0.01; Figure 1e).

Differential retinal gene expression in Wt and P2Y1R-KO mice. To assess information about a differential gene expression in control and postischemic retinae as a consequence of P2Y1 receptor deciency, which may explain the described differential degeneration pattern in Wt and P2Y1RKO mice, we performed a RT2 Proler PCR Array cAMP (30-50-cyclic adenosine monophosphate)/Ca2 Signalling PathwayFinder. Expression levels of a total of 84 genes

(see also Supplementary Table S1) from retinae 7 days after transient ischemia of 90-min duration were investigated. For 21 genes, no major changes in expression levels were observed neither if expression levels in postischemic retinae were compared with the respective control eyes, nor if expression levels were compared between Wt and P2Y1RKO retinae (Figure 2a). We found 14 genes with different expression in the untreated retina of P2Y1R-KO mice compared with Wt retinae; all but one were upregulated (Figure 2b). The expression levels of these genes remained unaltered after ischemia in retinae of Wt and P2Y1R-KO mice. At least ve of the upregulated genes are known to be expressed in Mller cells, including the transcription factors Junb and Stat3 (see Supplementary Table, for abbreviations of genes).25,26 A higher expression level of the enzyme, inducible nitric oxide synthase (iNOS; Nos2), was also observed (Figure 2b). Nos2 is known to be expressed in microglial and Mller cells of the ischemic retina.27,28 The

higher expression levels of Junb, Nos2, Stat3 and Tacr1 (Figure 2b) may suggest a slight chronic activation of microglia and Mller cells in the P2Y1R-KO retina.

Transient ischemia of the Wt retina induced alterations in the expression of 24 genes of which 21 were downregulated (Figure 3a). Calbindin (Calb) is expressed in amacrine and horizontal cells of the rat retina and was downregulated after retinal ischemia (Figure 3a).29 Amacrine cells (ACs) also express neuropeptide Y (Npy), vasoactive intestinal polypeptide (Vip) and tyrosine hydroxylase (Th).3032 In the postischemic

retina of P2Y1R-KO mice, only the expression of Ppp2ca, the gene of the a-isoform of the catalytic subunit of the protein phosphatase 2 (PP2A), was markedly decreased (Figure 3a).

PP2A is present in rod outer segments where it is involved in vesicle trafcking within photoreceptor synapses.3335

Finally, we found a group of 16 genes that were differentially expressed in control and ischemic retinae of Wt and P2Y1RKO mice (Figure 3b). The expression levels of these genes were higher in control retinae of P2Y1R-KO mice than in control retinae of Wt mice (Figure 3b). Among these genes are those for transcription factors (Atf3 and Egr2), enzymes (Eno2, Prkar1a, Ppp1r15a and Ptgs2), and the neuropeptides glucagon (Gcg) and preproenkephalin (Penk) (Figure 3b). Both neuropeptides are expressed in ACs.36,37 We found increased

expression of the presumable microglial voltage-gated potassium channel Kv1.5 gene (Kcna5)38 in control retinae of P2Y1R-KO mice compared with Wt retinae (Figure 3b).

Glial cells are one source of thrombospondin-1, a matricellular glycoprotein with manifold functions. It supports neurite formation, has anti-inammatory effects and prevents angiogenesis in the retina.3942 Putative downstream signaling mechanisms involve activation of transforming growth factor-b (TGF-b).43

Here we found a remarkably higher level of thrombospondin-1 gene (Thbs1) expression together with a slightly upregulated gene expression of TGF-b3 (Tgfb3) in control P2Y1R-KO retinae if compared with the Wt tissue. In agreement with recent studies,44 we found a moderate upregulation of Thbs1 in postischemic Wt retinae, but the expression level did not reach the level detected in retinae of P2Y1R-KO animals.

Effects on specic retinal neurons. The gene proling data suggest that various retinal cell types are differently

Cell Death and Disease

Role of P2Y1 signaling in ischemic mouse retina T Pannicke et al

3

Figure 1 Ischemic degeneration of the murine retina after transient retinal ischemia of 90 min. (a) Counting of cell nuclei was performed in 100-mm-wide areas close to the optic nerve head from retinae after 7 days survival marked as the white square left and shown at larger magnication right. Cell nuclei were labeled with TO-PRO-3 (blue), astrocytes were immunostained with anti-GFAP antibodies (green) and Mller cells with anti-CRALBP antibodies (red).OPL, outer plexiform layer. (b) MeanS.E.M. number of cell nuclei in different retinal layers. (c) MeanS.E.M. thickness of the IPL. (b and c) Signicant differences compared with the respective control eyes of the same genotype: ***Po0.001, **Po0.01. Signicant differences between the indicated strains: KPo0.05, KKPo0.01, KKKPo0.001. Data were obtained from 4 to 6 animals.

From each animal, 46 central slices were analyzed leading to n 1733. (d) Apoptotic cell death determined with TUNEL staining 1 day after transient retinal ischemia of

90 min. Staining of all cell nuclei (blue in the left parts of the images) and of TUNEL-labeled nuclei (red in the left parts and white in the right parts of the images) in retinae from Wt and P2Y1R-KO mice. Bars, 50 mm. (e) MeanS.E.M. number of TUNEL-labeled cell nuclei in the INL and ONL. KKPo0.01. Apoptotic activity was determined in 1318 central retinal slices derived from 4 to 5 animals

Cell Death and Disease

Role of P2Y1 signaling in ischemic mouse retina T Pannicke et al

4

Figure 2 Gene expression proling in retinae of Wt and P2Y1R-KO mice. The tissues were isolated 7 days after transient retinal ischemia of 90 min and from untreated control eyes from three mice of each genotype. (a) Genes that did not show any differences in expression levels according to the analysis using the RT2 Proler PCR Array Data Analysis software. (b) Genes are depicted that are differentially expressed in control retinae of Wt and P2Y1R-KO mice. Genes in red were signicantly upregulated, genes in blue were signicantly downregulated compared with the respective reference group

affected by ischemia and absence of P2Y1R. Retinal slices were immunolabeled for the marker proteins calretinin (ganglion cells, ACs),29,45 calbindin (amacrine and horizontal cells), protein kinase Ca (PKCa; bipolar cells) and cellular retinaldehyde-binding protein (CRALBP; Mller cells) to quantify the respective cell types. Calretinin immunostaining revealed positive cell bodies in the GCL and INL and three positive bands in the IPL (Figure 4a). Whereas this staining was markedly reduced after ischemia in the Wt retina in both the GCL and INL, the effect was less prominent in the P2Y1R-KO retina (Figure 4a). The total number of calretinin-positive cells was signicantly higher in the postischemic P2Y1R-KO retina compared with values found in Wt mice in both layers (Figure 4c). Of note, we also found major differences in the thickness of the IPL in ischemic retinae of Wt and P2Y1R-KO mice (Figure 1c). As the dendritic tree of retinal ganglion cells and ACs, which seem to be less affected in P2Y1R-KO mice, should considerably contribute to IPL thickness, we

determined the overall uorescence signal of calretinin immunoreactivity in the IPL. In line with the abovementioned assumption and with the observed superior survival of calretinin-positive cells per se, the level of calretinin immunoreactivity in the IPL (Figure 4d) was signicantly higher in ischemic retinae of P2Y1R-KO animals than in ischemic retinae of Wt animals.

Calbindin immunoreactivity was observed in ACs (located in the inner part of the INL) and horizontal cells (located in the outer part of the INL) with a tendency of a minor loss of calbindin-labeled cells in the postischemic P2Y1R-KO retina as compared with the retinae of Wt mice (Figures 4b and e). We found no apparent differences in the PKCa immunolabeling of bipolar cells (Figure 4b) and in the CRALBP immunostaining of Mller cells (Figure 4a) between the retinae of both mouse strains.

Mller cell gliosis. In the normal retina, the intermediate lament protein, glial brillary acidic protein (GFAP), is

Cell Death and Disease

Role of P2Y1 signaling in ischemic mouse retina T Pannicke et al

5

Figure 3 Ischemia-induced alterations in gene expression in retinae of Wt and P2Y1R-KO mice 7 days after 90 min of transient ischemia. Relative expression levels for two groups of genes are shown (a) genes that were differentially expressed in the ischemic retinae of Wt animals only and were stable after transient ischemia in retinae of P2Y1R-KO animals, and (b) a smaller subset of genes that displays a rather complicated regulation pattern. Genes in red were signicantly upregulated, genes in blue signicantly downregulated compared with the respective reference group

predominantly expressed by retinal astrocytes rather than by Mller cells; upregulation of GFAP is an early marker of Mller cell gliosis.46 We found that Mller cells

were devoid of GFAP immunoreactivity in the untreated retinae of Wt and P2Y1R-KO mice (Figure 5a). In contrast, Mller cells in the postischemic retinae of Wt and

Cell Death and Disease

Role of P2Y1 signaling in ischemic mouse retina T Pannicke et al

6

Figure 4 Effects of transient ischemia on retinal neurons 7 days after 90 min of transient ischemia. (a) Retinal slices were immunostained for calretinin (green) and CRALBP (red). Cell nuclei were labeled with TO-PRO-3 (blue). Ischemia resulted in a marked reduction of the calretinin staining of distinct cell bodies and a disappearance of the calretinin-labeled bands in the IPL. These effects were more pronounced in the retina of Wt mice than in the P2Y1R-KO retina. (b) Calbindin (red) and PKCa (green)

immunoreactivities in retinal slices. Cell nuclei were labeled with TO-PRO-3 (blue). Postischemic retinae displayed reduced levels of calbindin immunoreactivity in the INL compared with untreated control retinae. The level of PKCa immunoreactivity was only slightly different between control and postischemic retinae. (c) MeanS.E.M. number of calretinin-positive cells counted in the INL and the GCL. (d) MeanS.E.M. relative intensity of calretinin immunoreactivity in the IPL, reecting the calretinin content of AC dendrites. The data were obtained in a 460-mm-wide area. (e) MeanS.E.M. number of calbindin-positive amacrine and horizontal cells. (ce) Data were obtained by analysis of 1014 central retinal slices derived from 3 to 4 animals. Signicant difference to values from the respective untreated control: *Po0.05, **Po0.01, ***Po0.001;

KPo0.05, KKPo0.01, KKKPo0.001. Scale bars, 20 mm. OPL, outer plexiform layer

Cell Death and Disease

Role of P2Y1 signaling in ischemic mouse retina T Pannicke et al

7

P2Y1R-KO mice were immunolabeled for GFAP over their entire length.

In line with ndings from Hirrlinger et al.,47 we found that Mller cells of Wt animals, isolated 7 days after transient retinal ischemia for 60 min, displayed a reduction of the Kir currents by approximately 25% as compared with cells from untreated retinae (Table 1). Prolongation of ischemia up to 90 min aggravated this effect resulting in a decrease of the Kir currents in Wt Mller cells by almost 70%. Ischemia also induced an increase in the membrane capacitance of Wt Mller cells (Table 1), suggesting a hypertrophy of the cells.

As P2Y1R signaling was implicated to be involved in gliosis induction,48 we investigated the dependence of the ischemic reduction of the Kir currents on purinergic signaling. Surprisingly, cells from untreated P2Y1R-KO eyes displayed a small but signicant (Po0.05) reduction of the Kir current amplitude by about 10% if compared with cells from untreated Wt retinae (Table 1). Transient ischemia of 60 or 90 min resulted in a Kir current amplitude reduced by 1020% of the control current in P2Y1R-KO mice (Table 1). This reduction was signicantly less pronounced than in cells from Wt animals (Figure 5b; Table 1). The data indicate that deletion of P2Y1 attenuates the effect of retinal ischemia-reperfusion on the Kir currents of Mller cells.

We found a slight but signicant (Po0.01) increase in the membrane capacitance of isolated Mller cells from untreated

P2Y1R-KO mice as compared with cells from Wt animals suggesting an increase in the size of Mller cells of P2Y1RKO animals (Table 1). The membrane capacitance was explicitly increased only in some but not all cases of ischemia, whereas ischemia always induced a decrease of inward current densities (Table 1).

The predominant Kir channel subtype expressed by Mller cells is Kir4.1.49,50 It is concentrated in cell membranes contacting inner limiting membrane and in those surrounding blood vessels. Kir4.1 immunolabeling was similar in retinal slices from untreated Wt and P2Y1R-KO eyes (Figure 5c). Kir4.1 immunoreactivity was more evenly distributed and partially downregulated in the postischemic Wt and P2Y1R-KO retina.

Microglia activation. In addition to Mller cells, microglial cells were shown to express P2Y1.17,51 To assure that observed effects of P2Y1R deciency were largely due to an altered gliotic activation of Mller glia rather than being mediated by a changed activation pattern of microglia, we characterized the latter on the basis of cell numbers and morphological parameters (Supplementary Figure S1) in retinal tissues isolated 1 day after HIOP-induced retinal ischemia for 90 min. We only found minor differences in the characteristics of microglia in the control and postischemic retina of Wt and P2Y1R-KO mice (see Supplementary Information), which do not explain the nding of an altered degeneration pattern in postischemic Wt and P2Y1R-KO retinae.

Discussion

Effects of P2Y1 deciency in the control retina. Analysis of the untreated retina revealed signicantly less cells in the GCL and INL of P2Y1R-KO compared with Wt mice (Figure 1b). As P2Y1 stimulates the proliferation of retinal progenitors,52 our data may indicate that P2Y1-mediated

signaling has a role in the development of the murine retina. Alternatively, and although we did not nd TUNEL-positive cells in the control P2Y1R-KO retinae (Figure 1d), slow degenerative processes in the adult P2Y1R-KO retina leading to the reduced cell numbers cannot be excluded. Interestingly, cellular hypertrophy and decreased Kir current amplitudes (Table 1), but no elevation in GFAP expression, suggest a low-level Mller cell reactivity in the P2Y1R-KO retina. Gene expression proling revealed enhanced expression levels of Nos2, Stat3, Junb and Thbs1 in untreated P2Y1R-KO retinae (Figures 2b and 3b) whose upregulation has been associated with Mller cell gliosis.25,26,28 For example, Mller cells

produce thrombospondin-1 under ischemic conditions.44

Thrombospondin-1 activates latent TGF-b.43 Accordingly, we also detected an enhanced expression of TGF-b3 and of its downstream target inhibin b-A (Inhba)53 in P2Y1R-KO control retinae (Figures 2b and 3b). Upregulation of S100a8/S100a9, as found in P2Y1R-KO retinae (Figure 3b), has been associated with microglia activation.54,55 However, we found only slight differences in the microglial morphology between P2Y1R-KO and Wt retinae (Supplementary Figure S1). Possibly, thrombospondin-1 counterbalances microglia

Figure 5 Ischemia-induced Mller cell gliosis. (a) GFAP immunoreactivity (green) in slices of retinae isolated 7 days after transient retinal ischemia of 90 min. Labeling of cell nuclei (blue) revealed reduced cell numbers in all nuclear layers. (b) Examples of whole-cell K currents recorded in Mller cells isolated from untreated control retinae (left for each strain) and from retinae that were obtained 7 days after transient retinal ischemia of 90 min (right for each strain). Outward currents are depicted upwardly, inward currents are depicted downwardly. Currents were evoked by 20-mV incremental steps from a holding potential of 80 mV. (c)

Distribution of Kir4.1 immunoreactivity in slices of control retina and of retinae obtained 7 days after transient retinal ischemia of 90 min. Filled arrows, Kir4.1 labeling of the inner limiting membrane. Unlled arrow, perivascular Kir4.1 labeling. Scale bars, 20 mm

Cell Death and Disease

Role of P2Y1 signaling in ischemic mouse retina

T Pannicke et al

8

Table 1 Electrophysiological properties of murine Mller cells

Wt P2Y1R-KO

Untreated HIOP 60 min HIOP 90 min Untreated HIOP 60 min HIOP 90 min

Inward current amplitude (pA) 3082606 22371069*** 989680*** 2753761J 2238803** 2358975*,JJJ Relative inward currents (%) 10019 7535*** 3222*** 10020 9048 8237**,JJJ

Membrane potential (mV) 864 853 7817** 854 855 865J

Membrane capacitance (pF) 4614 679*** 5315 5613JJ 7016*** 6410**,JJ Current density (pA/pF) 7425 3415*** 2912*** 5321JJJ 3617*** 3715***

numbers 9 mice 6 mice 3 mice 11 mice 6 mice 5 mice 42 cells 31 cells 19 cells 61 cells 34 cells 29 cells

The cells were isolated 7 days after HIOP-induced retinal ischemia. For the calculation of the relative inward currents, the mean current amplitude recorded from cells of the untreated control eye was set to 100% for each mouse, and the relative value for each cell from the contralateral postischemic eye was calculated. *Po0.05, **Po0.01, ***Po0.001 compared with the respective value of the untreated eye from animals of the same strain. JPo0.05, JJPo0.01, JJJPo0.001 compared with the respective value from Wt mice. Data are given as meanS.D.

activation in P2Y1R-KO animals.37 It remains to be determined whether the low-level glial activation in the P2Y1R-KO retina reects degenerative processes or alterations of glial function without signicant effects on retinal integrity.

Ischemic retinal degeneration. Signicantly less photoreceptors survived in P2Y1R-KO mice than in Wt retinae (Figure 1b). The TUNEL assay, revealing more apoptotic cells in the ONL of P2Y1R-KO than in Wt mice, conrmed this nding (Figure 1e). Interestingly, the IPL in postischemic P2Y1R-KO retinae was thicker than in Wt mice, despite similar cell numbers in the GCL and INL (Figure 1c), probably indicating a better preservation of neurites in the IPL of P2Y1R-KO mice. Gene expression proling points to ACs as candidates displaying increased survival in the P2Y1-

decient postischemic retina. We found that 12 genes that are primarily expressed in ACs2932 were downregulated in the postischemic Wt but stably expressed in the postischemic P2Y1R-KO retinae (Figure 3a). These data were conrmed by immunolabelings showing higher levels of AC (calretinin, calbindin), but also ganglion cell (calretinin) markers in postischemic P2Y1R-KO retinae indicating a better survival of these cell types (Figure 4). Both cell types are highly susceptible to ischemic damage.27,56 We found

that only 15% of putative ACs survived in postischemic Wt retinae, whereas the mean overall reduction of retinal cells was approximately 50% (Figures 1b and 4c).

Effects of P2Y1 deciency on ischemia-induced Mller cell gliosis. Transient retinal ischemia resulted in a reduction of the Kir current amplitude, which was signicantly weaker in P2Y1R-KO Mller cells than in Wt cells (Table 1). Therefore, we concluded that P2Y1 signaling is involved in the regulation of the Kir channel expression during gliotic Mller cell activation. This is in agreement with previous studies showing that P2Y1 signaling has a role in gliosis induction in the brain.13,48

Using gene expression proling, we found that several genes rather specically expressed in Mller cells (e.g., Adrb1, Calcrl, Chga, Mif and TGFb3) were differently regulated by ischemia in P2Y1R-KO retinae compared with

Wt retinae (Figure 3). A differential Mller cell degeneration as a reason for the downregulation of Mller cell-specic genes in Wt retinae appears unlikely, because the overall

appearance of Mller cells was similar in the postischemic Wt and P2Y1R-KO retinae. Instead, we assume that this downregulation is depending on P2Y1-mediated signaling.

Impact of purinergic signaling on glial support of neuronal survival. The low-level activation of Mller and microglial cells in the untreated retina of P2Y1R-KO mice may provide a preconditioning to pathological events like ischemia and, thus, may be protective for inner retinal neurons such as ACs (Figure 4). Preconditioning, for example, by brief ischemia, has neuroprotective effects in the retina57,58 and typically involves an upregulation of enzymes involved in nitric oxide formation, such as iNOS (Nos2).59 We found a higher expression level of Nos2 in the retina of P2Y1R-KO mice as compared with Wt retina (Figure 2b), however, we did not study the functional role of iNOS in more detail. Therefore, it remains to be determined in future whether the glial activation in the retina of P2Y1R-KO mice provides preconditioning.

In contrast to this baseline glial activation, some aspects of reactive Mller cell gliosis, including the decrease in Kir currents (Table 1), are alleviated in the retina of P2Y1R-KO mice if compared with the Wt. This suggests that P2Y1 signaling is involved in the regulation of distinct aspects of Mller cell gliosis, particularly, the expression level of Kir4.1 channels. The rather high potassium conductance of Mller cells and the unaltered or increased glial expression of neurotrophic and anti-inammatory factors such as TGF-b3 (Tgfb3), macrophage migration inhibitory factor (Mif) and thrombospondin-1 (Thbs1) (Figure 3) presumably account for the improved survival of some neuronal subtypes in the inner retina of P2Y1R-KO animals (Figure 6). The maintained high potassium conductance of Mller cells in the postischemic P2Y1R-KO retina may stabilize the ion and volume home-ostasis, and, thereby, may prevent a cytotoxic hyperexcitation of the already energy-deprived neurons of the inner retina being highly sensitive to hyperexcitation associated with ischemia (Figure 6).27,56,60

A different scenario was observed in the outer retina. An increased photoreceptor cell loss in the postischemic P2Y1RKO compared with the Wt retinae (Figure 1) suggests that P2Y1-mediated signaling has a protective effect on photo-receptors in ischemia. Mller cells support photoreceptors by delivery of lactate and neurotrophic factors.61,62 Selective

Cell Death and Disease

Role of P2Y1 signaling in ischemic mouse retina T Pannicke et al

9

ablation of Mller cells results in photoreceptor loss.63,64 It

remains to be claried in future studies, which of these putatively involved mechanisms are affected in the postischemic P2Y1R-KO retina. Further mechanisms, for example, the regulation of the extracellular space volume,65 may

contribute to the glial support of photoreceptor survival (Figure 6). One major alteration in Mller cell physiology in P2Y1R-KO mice is the impaired capability to maintain cellular volume control under hypoosmotic conditions.16 The latter

involves a signaling cascade induced by an activation of metabotropic Glu and P2Y1 receptors, and glial release of adenosine 50-triphosphate and adenosine (AD).14 In the outer plexiform layer, this signaling cascade may be continuously activated by the constant release of Glu from photoreceptor terminals.66 Interruption of this feedback loop in P2Y1R-KO retinae may contribute to photoreceptor degeneration because of the impaired Glu-induced release of the neuroprotectant AD from Mller cells (Figure 6).67

Anti-purinergic agents were suggested to serve as therapeutics for the treatment of ischemic disorders in neural tissues. The present results of differential effects of P2Y1 deciency on the survival of neuronal subtypes in the ischemic retina call the use of such agents into question. This is consistent with conicting results regarding a potential protective effect of P2Y signaling in ischemia published by others. P2Y1 antagonists applied in a rat model of cerebral ischemia reduced the infarct volume and improved motor function

recovery.68 In contrast, neuroprotective effects of P2Y1-mediated signaling in brain astrocytes under various pathological conditions have been reported.6971 Further research

is required to determine and to separate the benecial and detrimental effects of P2Y1-mediated signaling in the ischemic brain and retina.

Materials and MethodsMaterials. Papain was from Roche (Mannheim, Germany). All other substances used were from Sigma-Aldrich (Taufkirchen, Germany), unless stated otherwise. The following primary antibodies were used: rabbit anti-Kir4.1 (1 : 200; Sigma-Aldrich), mouse anti-GFAP (1 : 200; G-A-5 clone, Sigma-Aldrich), rabbit anti-ionized calcium binding adaptor molecule 1 (Iba1; 1 : 500, Wako, Neuss, Germany), goat anti-calretinin (1 : 500, Swant, Marly, Switzerland), mouse anti-calbindin (1 : 400, Swant), rabbit anti-PKCa (1 : 300, Santa Cruz Biotechnology,

Heidelberg, Germany), rabbit anti-CRALBP (1 : 300, Santa Cruz Biotechnology) and mouse anti-glutamine synthetase (1 : 1000, Merck Millipore, Darmstadt, Germany). The following secondary antibodies were used: Cy5-conjugated donkey anti-goat, Cy3-conjugated donkey anti-rabbit, Cy2-conjugated donkey anti-mouse, Cy3-conjugated goat anti-rabbit and Cy2-conjugated goat anti-mouse. All secondary antibodies were applied in a 1 : 200 dilution and were obtained from Dianova (Hamburg, Germany). The apoptosis rate was detected using the in situ cell death detection kit, tetramethylrhodamine red (Roche).

Animals. All experiments were done in accordance with the European Communities Council Directive 86/609/EEC, and were approved by the local authorities. Animals were maintained with free access to water and food in an air-conditioned room on a 12-h lightdark cycle. Adult (26 months old) mice decient in the nucleotide receptor P2Y1 (P2Y1R-KO; 129Sv background) were characterized as described.72 Briey, for genotyping of the P2Y1R-KO mice and wild-type (Wt) littermate controls, the genomic region of the P2Y1 receptor was characterized. PCR analysis was made with the following primers: knockout, common, 50-GCAGTGTTTGGGGTCAGAAT-30, neo, 50-GGGGAACTTCCTGACT

AGGG-30; Wt, common, 50-GCAGTGTTTGGGGTCAGAAT-30, Wt, 50-AACATACG CTGCAAGGCTCT-30. Age- and weight-matched littermate Wt controls were used.

Retinal ischemia. Transient retinal ischemia was induced in one eye by the HIOP method. The other eye remained untreated as internal control. Anesthesia was induced with ketamine (100 mg/kg body weight, intraperitoneal (i.p.); Ratiopharm, Ulm, Germany), xylazine (5 mg/kg, i.p.; Bayer Vital, Leverkusen, Germany) and atropine sulfate (100 mg/kg, i.p.; Braun, Melsungen; Germany). The anterior chamber of the test eye of anesthetized mice was cannulated from the pars plana with a 30-gauge infusion needle, connected to a saline bottle. The intraocular pressure was increased to 160 mm Hg for 60 or 90 min by elevating the bottle. After removing the needle, the animals survived for 1 or 7 days and, subsequently, were killed with carbon dioxide.

Preparation of isolated Mller cells. Isolated retinae were incubated in papain (0.2 mg/ml)-containing Ca2-/Mg2-free phosphate-buffered saline, pH7.4, for 30 min at 37 1C, followed by several washing steps with saline. After short incubation in saline supplemented with deoxyribocnuclease I (200 U/ml), the tissue pieces were triturated by a 1-ml pipette tip, to obtain isolated retinal cells. The cells were stored at 4 1C in serum-free minimum essential medium until use within 4 h after cell isolation. Mller cells were identied in the cell suspensions according to their characteristic morphology.

Whole-cell patch-clamp records of isolated Mller cells. The whole-cell currents of freshly isolated Mller cells were recorded at room temperature using the Axopatch 200A amplier (Axon Instruments, Foster City, CA, USA) and the ISO-2 computer program (MFK, Niedernhausen, Germany). The signals were low-pass ltered at 1, 2 or 6 kHz (eight-pole Bessel lter) and digitized at 5, 10 or 30 kHz, respectively, using a 12-bit A/D converter. Patch pipettes were pulled from borosilicate glass (Science Products, Hofheim, Germany) and had resistances between 4 and 6 MO when lled with a solution containing (mM) 10 NaCl, 130 KCl, 1 CaCl2, 2 MgCl2, 10 ethylene glycol tetraacetic acid and 10 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), adjusted to pH 7.1 with tris(hydroxymethyl)aminomethane (Tris). The recording chamber was continuously perfused with extracellular solution that

Figure 6 Artists view on the putative janus-faced effect of P2Y1 deciency in the postischemic retina. Left: schematic drawing of a Mller cell (light green) and its associated neurons (dark blue). Right: downregulation of Kir4.1 in Wt mice prevents Mller cells from fullling their important function to maintain the retinal potassium homeostasis by transcellular clearance of the latter from the extracellular space into bigger uid lled compartments such as the blood system. This may lead to cytotoxic neuronal hyperexcitation especially of cells generating action potentials such as retinal ganglion cells and subtypes of ACs (indicated by red membrane) in the inner retina. In contrast, photoreceptors in the outer retina appear to benet from functional P2Y1R, probably due to P2Y1-mediated release of neuroprotective substances like

AD or other trophic factors. In P2Y1R-KO mice the stable Kir4.1 expression appears to allow a better survival of neurons in the inner retina, while the lack of P2Y1R

considerably disrupts the essential support of photoreceptors by Mller glia. BV, blood vessel; mGluR, metabotropic Glu receptor; NT, nucleoside transporter; OPL, outer plexiform layer; PRC, photoreceptor cell; RGC, retinal ganglion cell

Cell Death and Disease

Role of P2Y1 signaling in ischemic mouse retina T Pannicke et al

10

contained (mM) 135 NaCl, 3 KCl, 2 CaCl2, 1 MgCl2, 1 Na2HPO4, 10 HEPES and 11 glucose, adjusted to pH 7.4 with Tris. To evoke potassium currents, de- and hyperpolarizing voltage steps of 250 ms duration, with increments of 10 mV, were applied from a holding potential of 80 mV. The amplitude of the steady-state Kir

currents was measured at the end of the 250-ms voltage step from 80 to 140 mV. The membrane capacitance of the cells was measured by the integral of the uncompensated capacitive artifact (ltered at 6 kHz) evoked by a 10-mV voltage step in the presence of extracellular BaCl2 (1 mM). Current densities were calculated by dividing inward current amplitudes evoked by 60 mV hyperpolarization by the membrane capacitance. The resting membrane potential was measured in the current-clamp mode.

Histological staining and immunohistochemistry. Isolated retinae were xed in 4% paraformaldehyde for 2 h. After several washing steps in buffered saline, the tissues were embedded in saline containing 3% agarose (w/v), and 70-mm thick slices were cut by using a vibratome. The slices were incubated in 5%

normal goat serum plus 0.3% Triton X-100 plus 1.0% dimethyl sulfoxide in saline for 2 h at room temperature and, subsequently, in the primary antibodies overnight at 4 1C. After washing in 1% bovine serum albumin, the secondary antibodies were applied for 2 h at room temperature. Cell nuclei were labeled with TO-PRO-3 (1 : 1000; Life Technologies, Carlsbad, CA, USA). Control slices were stained without the primary antibody; no unspecic labeling was observed following incubation with the secondary antibody alone (data not shown). Images were taken with the LSM 510 Meta (Zeiss, Oberkochen, Germany). The number of cell nuclei in different retinal layers was counted in a 100-mm-wide area close to the optic nerve head (optical slice thickness, 1.5 mm; Figure 1a).

Iba1-immunolabeling revealed microglial cells in paraformaldehyde-xed retinal whole-mount preparations. Z-stacks created at the laser scanning microscope (LSM) were used to quantify microglia properties using the software of the LSM. First, the number of cells was counted in different retinal layers in an area of 325 325 mm. Second, the area occupied by individual cells was

determined by recording the area conned by the most distal ends of microglial processes. Third, the number of branching points per cell was detected. A branching point was dened as a spot on a microglial process where this process is divided into two parts of 42 mm length. Finally, the total area of the soma was calculated.

The TUNEL assay was performed according to the manufacturers protocol (Roche). To assure good accessibility for the components of the assay to the tissue, we subjected free oating retinal slices to a brief microwave treatment in citrate buffer (pH 6.0, 0.1 M). After washing with buffered saline and tissue permeabilization for 10 min in 4% Triton X-100, the slices were incubated for 90 min in the labeling solution. Afterward, the slices were counterstained with TO-PRO-3 and mounted onto slides.

Real-time PCR. Total RNA was prepared from control and postischemic (after 7 days survival) retinae of Wt and P2Y1R-KO mice (n 3 each) with

RNAeasy Mini Kit (Qiagen, Hilden, Germany) according to the manufacturers instructions. The quality of the RNA was controlled by agarose gel electrophoresis. A NanoDrop spectrophotometer was used to measure RNA concentration. The A260/A280 ratio of the optical density was 41.9 in all samples, indicating highly puried RNA. cDNA was synthesized with 300 ng RNA for each sample using the RT2 First Strand Kit (Qiagen). The investigation of gene expression was carried out using the RT2 Proler PCR Array cAMP/Ca2 Signalling PathwayFinder (Qiagen; Supplementary Table S1). Real-time PCR was performed following the manufacturers instruction using a Rotor-Gene 6000 cycler (Corbett Research, Sydney, Australia) and RT2 SYBR Green/ROX qPCR Master Mix (Qiagen). Changes in gene expression levels were calculated according to the

DDCt method using the web based RT2 Proler PCR Array Data Analysis software (http://pcrdataanalysis.sabiosciences.com/pcr/

Web End =http://pcrdataanalysis.sabiosciences.com/pcr/ arrayanalysis.php).

Statistics. Data are expressed as meanS.E.M. or S.D. (patch-clamp data). Statistical analysis was made using Prism (Graphpad Software, San Diego, CA, USA); signicance was determined by the non-parametric MannWhitney U-test.

Conict of InterestThe authors declare no conict of interest.

Acknowledgements. This study was supported by Deutsche Forschungsgemeinschaft (RE 849/16-1 to AR, FOR 748 to AG and AR, GRK 1097 to AR, SPP 1172 to AR, PA 615/2-1 to TP), the EU FP 7 Program EduGlia 237956 to AR and PRO RETINA-Stiftung to AG.

Author contributionsTP, IF, BB and AG conceived, carried out experiments and analyzed data. LW, KS, EU and WH carried out experiments. UK, UU, TA and PI gave valuable technical support or provided non-commercially available materials. AB, AR, TP and AG were involved in data interpretation and wrote the paper. All authors were involved in writing the paper and had nal approval of the submitted and published versions.

1. Parpura V, Heneka MT, Montana V, Oliet SHR, Schousboe A, Haydon PG et al. Glial cells in (patho)physiology. J Neurochem 2012; 121: 427.

2. Kofuji P, Newman EA. Potassium buffering in the central nervous system. Neuroscience 2004; 129: 10451056.

3. Marcaggi P, Attwell D. Role of glial amino acid transporters in synaptic transmission and brain energetics. Glia 2004; 47: 217225.

4. Simard M, Nedergaard M. The neurobiology of glia in the context of water and ion homeostasis. Neuroscience 2004; 129: 877896.

5. Bringmann A, Grosche A, Pannicke T, Reichenbach A. GABA and glutamate uptake and metabolism in retinal glial (Mller) cells. Front Cell Endocrinol 2013; 4: 7285.

6. Sofroniew MV. Reactive astrocytes in neural repair and protection. Neuroscientist 2005; 11: 400407.

7. Sofroniew MV. Molecular dissection of reactive astrogliosis and glial scar formation. Trends Neurosci 2005; 32: 638647.

8. Robel S, Berninger B, Gtz M. The stem cell potential of glia: lessons from reactive gliosis. Nat Rev Neurosci 2011; 12: 88104.

9. Bringmann A, Pannicke T, Grosche J, Francke M, Wiedemann P, Skatchkov SN et al. Mller cells in the healthy and diseased retina. Prog Retin Eye Res 2006; 25: 397424.

10. Reichenbach A, Bringmann A. Mller Cells in the Healthy and Diseased Retina. Springer: New York, 2010.

11. Djukic B, Casper KB, Philpot BD, Chin L-S, McCarthy KD. Conditional knock-out of Kir4.1 leads to glial membrane depolarization, inhibition of potassium and glutamate uptake, and enhanced short-term synaptic potentiation. J Neurosci 2007; 27: 1135411365.

12. Kucheryavykh YV, Kucheryavykh LY, Nichols CG, Maldonado HM, Baksi K, Reichenbach A et al. Downregulation of Kir4.1 inward rectifying potassium channel subunit by RNAi impairs potassium transfer and glutamate uptake by cultured cortical astrocytes. Glia 2007; 55: 274281.

13. Franke H, Krgel U, Schmidt R, Grosche J, Reichenbach A, Illes P. P2 receptor-types involved in astrogliosis in vivo. Br J Pharmacol 2001; 134: 11801189.

14. Wurm A, Pannicke T, Iandiev I, Francke M, Hollborn M, Wiedemann P et al. Purinergic signalling involved in Mller cell function in the mammalian retina. Prog Retin Eye Res 2011; 30: 324342.

15. Wurm A, Erdmann I, Bringmann A, Reichenbach A, Pannicke T. Expression and function of P2Y receptors on Mller cells of the postnatal rat retina. Glia 2009; 57: 16801690.16. Wurm A, Lipp S, Pannicke T, Linnertz R, Krgel U, Schulz A et al. Endogenous purinergic signalling is required for osmotic volume regulation of retinal glial cells. J Neurochem 2010; 112: 12611272.

17. De Simone R, Niturad CE, De Nuccio C, Ajmone-Cat MA, Visentin S, Minghetti L. TGF-b and LPS modulate ADP-induced migration of microglial cells through P2Y1 and P2Y12 receptor expression. J Neurochem 2010; 115: 450459.

18. Orellana JA, Montero TD, von Bernhardi R. Astrocytes inhibit nitric oxide-dependent Ca(2 ) dynamics in activated microglia: involvement of ATP released via pannexin 1

channels. Glia 2013; 61: 20232037.19. Osborne NN, Block F, Sontag KH. Reduction of ocular blood ow results in glial brillary acidic protein (GFAP) expression in rat retinal Mller cells. Vis Neurosci 1991; 7: 637639.

20. Osborne NN, Casson RJ, Wood JP, Chidlow G, Graham M, Melena J. Retinal ischemia: mechanisms of damage and potential therapeutic strategies. Prog Retin Eye Res 2004; 23: 91147.

21. Pannicke T, Iandiev I, Uckermann O, Biedermann B, Kutzera F, Wiedemann P et al. A potassium channel-linked mechanism of glial cell swelling in the postischemic retina. Mol Cell Neurosci 2004; 26: 493502.

22. Skatchkov SN, Eaton MJ, Shuba YM, Kucheryavykh YV, Derst C, Veh RW et al. Tandem-pore domain potassium channels are functionally expressed in retinal (Mller) glial cells. Glia 2006; 53: 266276.

23. Zayas-Santiago A, Agte S, Rivera Y, Benedikt J, Ulbricht E, Karl A et al. Unidirectional photoreceptor-to-Mller glia coupling and unique K channel expression in caiman retina. PLoS One 2014; 9: e97155.

24. Kuroiwa S, Katai N, Shibuki H, Kurokawa T, Umihira J, Nikaido T et al. Expression of cell cycle-related genes in dying cells in retinal ischemic injury. Invest Ophthalmol Vis Sci 1998; 39: 610617.

Cell Death and Disease

Role of P2Y1 signaling in ischemic mouse retina T Pannicke et al

11

25. Ohki K, Yoshida K, Yamakawa A, Harada T, Matsuda H, Imaki J. Jun-B gene expression in rat retinal cells following retinal injury. Curr Eye Res 1995; 14: 10211024.

26. Nelson CM, Ackermann KM, OHayer P, Bailey TJ, Gorsuch RA, Hyde DR. Tumor necrosis factor-alpha is produced by dying retinal neurons and is required for Mller glia proliferation during zebrash retinal regeneration. J Neurosci 2013; 33: 65246539.

27. Neufeld AH, Kawai S-I, Das S, Vora S, Gachie E, Connor JR et al. Loss of retinal ganglion cells following retinal ischemia: the role if inducible nitric oxide synthase. Exp Eye Res 2002; 75: 521528.

28. Kobayashi M, Kuroiwa T, Shimokawa R, Okeda R, Tokoro T. Nitric oxide synthase expression in ischemic rat retinas. Jpn J Ophthalmol 2000; 44: 235244.

29. Lee J-H, Shin JM, Shin Y-J, Chun M-H, Oh S-J. Immunochemical changes of calbindin, calretinin and SMI32 in ischemic retinas induced by increase of intraocular pressure and by middle cerebral artery occlusion. Anat Cell Biol 2011; 44: 2534.

30. Casini G, Brecha NC. Postnatal development of tyrosine hydroxylase immunoreactive amacrine cells in the rabbit retina: I. Morphological characterization. J Comp Neurol 1992; 326: 283301.

31. Casini G, Brecha NC. Colocalization of vasoactive intestinal polypeptide and GABA immunoreactivities in a population of wide-eld amacrine cells in the rabbit retina. Vis Neurosci 1992; 8: 373378.

32. Oh SJ, DAngelo I, Lee EJ, Chun MH, Brecha NC. Distribution and synaptic connectivity of neuropeptide Y-immunoreactive amacrine cells in the rat retina. J Comp Neurol 2002; 446: 219234.

33. Selke D, Anton H, Klumpp S. Serine/threonine protein phosphatases type 1, 2A and 2C in vertebrate retinae. Acta Anat (Basel) 1998; 1621: 151156.

34. Guatimosim C, Hull C, Von Gersdorff H, Prado MA. Okadaic acid disrupts synaptic vesicle trafcking in a ribbon-type synapse. J Neurochem 2002; 82: 10471057.

35. Haeseleer F, Sokal I, Gregory FD, Lee A. Protein phosphatase 2A dephosphorylates CaBP4 and regulates CaBP4 function. Invest Ophthalmol Vis Sci 2013; 54: 12141226.

36. Fischer AJ, Skorupa D, Schonberg DL, Walton NA. Characterization of glucagon-expressing neurons in the chicken retina. J Comp Neurol 2006; 496: 479494.

37. Huang J, Lin Y, Han R, Chen J, Wang YY, Wanf W et al. Spatial and temporal distribution patterns of enkephalinergic neurons in adult and developing retinas of the preproenkephalin-green uorescent protein transgenic mouse. Cells Tissues Organs 2012; 195: 563574.

38. Jou I, Pyo H, Chung S, Jung SY, Gwag BJ, Joe EH. Expression of Kv1.5 K channels in activated microglia in vivo. Glia 1998; 24: 408414.

39. Eroglu C, Allen NJ, Susman MW, ORourke NA, Park CY, Ozkan E et al. Gabapentin receptor alpha2delta-1 is a neuronal thrombospondin receptor responsible for excitatory CNS synaptogenesis. Cell 2009; 139: 380392.

40. Ng TF, Turpie B, Masli S. Thrombospondin-1-mediated regulation of microglia activation after retinal injury. Invest Ophthalmol Vis Sci 2009; 50: 54725478.

41. Wang Y, Wang S, Sheibani N. Enhanced proangiogenic signalling in thrombospondin-1-decient retinal endothelial cells. Microvasc Res 2006; 71: 143151.

42. Yu K, Ge J, Summers JB, Li F, Liu X, Ma P et al. TSP-1 secreted by bone marrow stromal cells contributes to retinal ganglion cell neurite outgrowth and survival. PLoS One 2008; 3: e2470.

43. Murphy-Ullrich JE, Poczatek M. Activation of latent TGF-beta by thrombospondin-1: mechanisms and physiology. Cytokine Growth Factor Rev 2000; 11: 5969.

44. Suzuma K, Takagi H, Otani A, Oh H, Honda Y. Expression of thrombospondin-1 in ischemia-induced retinal neovascularization. Am J Pathol 1999; 154: 343354.

45. Lee ES, Lee JY, Jeon CJ. Types and density of calretinin-containing retinal ganglion cells in mouse. Neurosci Res 2010; 66: 141150.

46. Lewis GP, Fisher SK. Up-regulation of glial fribrillary acidic protein in response to retinal injury: its potential role in glial remodeling and a comparison to vimentin expression. Int Rev Cytol 2003; 230: 263290.

47. Hirrlinger PG, Ulbricht E, Iandiev I, Reichenbach A, Pannicke T. Alterations in protein expression and membrane properties during Mller cell gliosis in a murine model of transient retinal ischemia. Neurosci Lett 2010; 472: 7378.

48. Franke H, Sauer C, Rudolph C, Krgel U, Hengstler JG, Illes P. P2 receptor-mediated stimulation of the PI3-K/Akt-pathway in vivo. Glia 2009; 57: 10311045.

49. Ishii M, Horio Y, Tada Y, Hibino H, Inanobe A, Ito M et al. Expression and clustered distribution of an inwardly rectifying potassium channel, KAB-2/Kir4.1, on mammalian retinal

Mller cell membrane: their regulation by insulin and laminin signals. J Neurosci 1997; 17: 77257735.50. Kofuji P, Ceelen P, Zahs KR, Surbeck LW, Lester HA, Newman EA. Genetic inactivation of an inwardly rectifying potassium channel (Kir4.1 subunit) in mice: phenotypic impact in retina. J Neurosci 2000; 20: 57335740.

51. Boucsein C, Zacharias R, Farber K, Pavlovic S, Hanisch U-K, Kettenmann H. Purinergic receptors on microglial cells: functional expression in acute brain slices and modulation of microglial activation in vitro. Eur J Neurosci 2003; 17: 22672276.

52. Sholl-Franco A, Fragel-Madeira L, Macama Ada C, Linden R, Ventura AL. ATP controls cell cycle and induces proliferation in the mouse developing retina. Int J Dev Neurosci 2010; 28: 6373.

53. Labb E, Lock L, Letamendia A, Gorska AE, Gryfe R, Gallinger S. Transcriptional cooperation between the transforming growth factor-beta and Wnt pathways in mammary and intestinal tumorigenesis. Cancer Res 2007; 67: 7584.

54. Kummer MP, Vogl T, Axt D, Griep A, Vieira-Saecker A, Jessen F et al. Mrp14 deciency ameliorates amyloid b burden by increasing microglial phagocytosis and modulation of amyloid precursor protein processing. J Neurosci 2012; 32: 1782417829.

55. Ziegler G, Prinz V, Albrecht MW, Harhausen D, Khojasteh U, Nacken W et al. Mrp-8 and -14 mediate CNS injury in focal cerebral ischemia. Biochim Biophys Acta 2009; 1792: 11981204.

56. Michalski D, Hartig W, Krgel K, Edwards RH, Bddener M, Bhme L et al. Region-specic expression of vesicular glutamate and GABA transporters under various ischaemic conditions in mouse forebrain and retina. Neuroscience 2013; 231: 328344.

57. Caprara C, Grimm C. From oxygen to erythropoietin: relevance of hypoxia for retinal development, health and disease. Prog Retin Eye Res 2012; 31: 89119.

58. Roth S, Li B, Rosenbaum PS, Gupta H, Goldstein IM, Maxwell KM et al. Preconditioning provides complete protection against retinal ischemic injury in rats. Invest Ophthalmol Vis Sci 1998; 39: 777785.

59. Sakamoto K, Yonoki Y, Kubota Y, Kuwagata M, Saito M, Nakahara T et al. Inducible nitric oxide synthase inhibitors abolished histological protection by late ischemic preconditioning in rat retina. Exp Eye Res 2006; 82: 512518.

60. Naruoka T, Nakahara T, Tsuda Y, Kurauchi Y, Mori A, Sakamoto K et al. ISO-1, a macrophage migration inhibitory factor antagonist, prevents N-methyl-D-aspartate-induced retinal damage. Eur J Pharmacol 2013; 718: 138144.

61. Poitry-Yamate CL, Poitry S, Tsacopolous M. Lactate released by Mller glial cells is metabolized by photoreceptors from mammalian retina. J Neurosci 1995; 15: 51795191.

62. Bringmann A, Iandiev I, Pannicke T, Wurm A, Hollborn M, Wiedemann P et al. Cellular signalling and factors involved in Mller cell gliosis: neuroprotective and detrimental effects. Prog Retin Eye Res 2009; 28: 423451.

63. Byrne LC, Khalid F, Lee T, Zin EA, Greenberg KP, Visel M et al. AAV-mediated, optogenetic ablation of Mller glia leads to structural and functional changes in the mouse retina. PLoS One 2013; 8: e76075.

64. Shen W, Zhu L, Lee SR, Chung SH, Gillies MC. Involvement of NT3 and P75NTR in photoreceptor degeneration following selective Mller cell ablation. J Neuroinammation 2013; 10: 137.

65. Wurm A, Iandiev I, Uhlmann S, Wiedemann P, Reichenbach A, Bringmann A et al. Effects of ischemia-reperfusion on physiological properties of Mller glial cells in the porcine retina. Invest Ophthalmol Vis Sci 2011; 52: 33603367.

66. Hagins WA, Penn RD, Yoshikami S. Dark current and photocurrent in retinal rods. Biophys J 1970; 10: 380412.

67. Housley GD, Bringmann A, Reichenbach A. Purinergic signalling in special senses. Trends Neurosci 2009; 32: 128141.

68. Kuboyama K, Harada H, Tozaki-Saitoh H, Tsuda M, Ushijima K, Inoue K. Astrocytic P2Y(1) receptor is involved in the regulation of cytokine/chemokine transcription and cerebral damage in a rat model of cerebral ischemia. J Cereb Blood Flow Metab 2011; 31: 19301941.

69. Fujita T, Tozaki-Saitoh H, Inoue K. P2Y1 receptor signalling enhances neuroprotection by astrocytes against oxidative stress via IL-6 release in hippocampal cultures. Glia 2009; 57: 244257.

70. Noguchi Y, Shinozaki Y, Fujishita K, Shibata K, Imura Y, Morizawa Y et al. Astrocytes protect neurons against methylmercury via ATP/P2Y1 receptor-mediated pathways in astrocytes. PloS One 2013; 8: e57898.

71. Zheng W, Talley Watts L, Holstein DM, Wewer J, Lechleiter JD. P2Y1R-initiated, IP3R-dependent stimulation of astrocyte mitochondrial metabolism reduces and partially reverses ischemic neuronal damage in mouse. J Cereb Blood Flow Metab 2013; 33: 600611.

72. Fabre JE, Nguyen M, Latour A, Keifer JA, Audoly LP, Coffman TM et al. Decreased platelet aggregation, increased bleeding time and resistance to thromboembolism in P2Y1-decient mice. Nat Med 1999; 5: 11991202.

Cell Death and Disease is an open-access journal published by Nature Publishing Group. This work is licensed under a Creative Commons Attribution 3.0 Unported License.The images or other third party material in this article are included in the articles Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/3.0/

Web End =http://creativecommons.org/licenses/by/3.0/

Supplementary Information accompanies this paper on Cell Death and Disease website (http://www.nature.com/cddis

Web End =http://www.nature.com/cddis)

Cell Death and Disease