Taurine deficiency damages retinal neurones: cone photoreceptors and retinal ganglion cells

Abstract

In 1970s, taurine deficiency was reported to induce photoreceptor degeneration in cats and rats. Recently, we found that taurine deficiency contributes to the retinal toxicity of vigabatrin, an antiepileptic drug. However, in this toxicity, retinal ganglion cells were degenerating in parallel to cone photoreceptors. The aim of this study was to re-assess a classic mouse model of taurine deficiency following a treatment with guanidoethane sulfonate (GES), a taurine transporter inhibitor to determine whether retinal ganglion cells are also affected. GES treatment induced a significant reduction in the taurine plasma levels and a lower weight increase. At the functional level, photopic electroretinograms were reduced indicating a dysfunction in the cone pathway. A change in the autofluorescence appearance of the eye fundus was explained on histological sections by an increased autofluorescence of the retinal pigment epithelium. Although the general morphology of the retina was not affected, cell damages were indicated by the general increase in glial fibrillary acidic protein expression. When cell quantification was achieved on retinal sections, the number of outer/inner segments of cone photoreceptors was reduced (20 %) as the number of retinal ganglion cells (19 %). An abnormal synaptic plasticity of rod bipolar cell dendrites was also observed in GES-treated mice. These results indicate that taurine deficiency can not only lead to photoreceptor degeneration but also to retinal ganglion cell loss. Cone photoreceptors and retinal ganglion cells appear as the most sensitive cells to taurine deficiency. These results may explain the recent therapeutic interest of taurine in retinal degenerative pathologies.

Fig. 1

Taurine deficiency in GES-treated mice. Plasma taurine levels were measured in control animals and in GES mice after 1 and 2 months of treatment (n = 8, SEM, P < 0.001, triple asterisks denote Mann–Whitney test) (a). Taurine concentration in retinal tissue was also reduced after 1 month of treatment (n = 8, SEM, P = 0.046, asterisk denotes Mann–Whitney test) (b)

Fig. 2

The weight development is reduced in GES-treated mice. No difference in the mean weight is noted at the beginning of the study (week 0), between the treated group and the control group (SEM, n = 8, P > 0.05, Student’s t test). After 2 months of treatment, the mean weight of the GES-treated mice is significantly reduced (SEM, n = 8, P = 0,024, asterisk denotes Student’s t test)

Fig. 3

Retinal cell function in GES-treated mice. Electroretinogram (ERG) response of a 10 cds m⁻² scotopic flash light stimulus recorded in a control animal, and in GES mouse (a). Response of a 25 cds m⁻² photopic flash light stimulus recorded in a control animal, and in GES mouse (b). Photopic ERG response to a 15-Hz flickers light stimulus in a control animal, and in GES mouse (c). Oscillary Potentials (Ops) isolated in a control, and a GES mouse (d). Quantification of scotopic and photopic ERG amplitudes in controls (n = 15), and GES-treated animals showing the significant decrease in standard flash and flickers photopic ERG amplitudes. (SEM, n = 8, P < 0.05, asterisk denotes Student’s t test) (e)

Fig. 4

Taurine deficiency modifies retinal autofluorescence (AF) but does not alter vascular permeability in GES-treated adult mice. Vascular network seen in late phase of the angiography (a, c). No difference is detected between GES-treated mice (a) and controls (c). AF images of the fundus (b, d) revealed the presence of sparse numerous round autofluorescent dots in both controls and GES mice. The intensity and the number of these dots were lower in GES-treated mice (b) than in controls (d). AF level was analysed on eye sections stained with DAPI in GES-treated mice (e–g) and in controls (h–j): scattered autofluorescent bodies (arrows) were present at the level of the retinal pigment epithelium (RPE) in both groups but a general increase in the RPE fluorescence was noted in GES-treated mice, such that the fluorescent bodies offered less contrast than in controls

Fig. 5

Autofluorescence in the retinal pigment epithelium (RPE). Autofluorescence at the RPE level was measured in four distinct areas on retinal sections along the dorso-ventral axis: dorso-peripheral (DP), dorso-central (DC), ventro-central (VC) and ventro-peripheral (VP). Dapi-labelled nuclei are shown in blue while grey levels correspond to the autoflorescence measured under a 556-nm excitation light. Scale bar represents 1 mm (a). Representative pictures of autofluorescence observed at the RPE level in one control and one GES-treated mouse in each area. Dapi-labelled nuclei correspond to choroid cells (C) and RPE cells (R). The corresponding autofluorescence measurement is represented on the graph at the right. Note the increase in autofluorescence observed in GES-treated animals at the RPE level. Horizontal axis corresponds to arbitrary units of autofluorescence and vertical axis corresponds to distance from RPE (μm). Scale bar represents 20 μm (b). Mean maximal values of autofluorescence at the RPE level according to retina area and treatment. Significant increase of autofluorescence in DP, DC and VP areas were found in GES mice (SEM, n = 8, P < 0.05, asterisk denotes Mann–Whitney test) (c)

Fig. 6

Taurine deficiency induces abnormal glial reactivity. The eyes sections were stained with DAPI and immunolabelled with anti-GFAP antibodies (a, b): glial reactivity was found in GES-treated mice as GFAP-positive processes extended vertically throughout the retina (a, c). The GFAP labelling was normally limited to the retinal ganglion cell layer and OPL in the control mice (b, d). Muller cells processes (red) extended to the outer limiting membrane, in contact with disorganized photoreceptors segments [stained in green with peanut lectin (PNA)] in GES-treated mice (c) as compared with controls (d)

Fig. 7

Taurine deficiency induces cone photoreceptors loss in GES-treated mice. Retinal sections of control animal (b) and GES-treated mice (a) were stained with peanut lectin (a, b), Inner and outer segments of cone photoreceptors are absent in many points of the retina in GES mice, leaving optically empty spaces throughout the segments photoreceptors line (a). These spaces are not detected on control eyes sections (b). Even if present, the inner/outer segments of cone photoreceptors seem broken and not well lined up in the treated mice (a) whereas they are aligned in controls (b). Cone photoreceptors count revealed a decrease number of cells in the GES-treated mice (SEM, n = 8, P < 0.05, asterisk denotes Student’s t test). Scale bars represent 25 μm (ONL outer nuclear layer, OPL outer plexiform layer, INL inner nuclear layer) (c)

Fig. 8

Taurine deficiency induces abnormal bipolar cell plasticity. Discrete neuronal plasticity was indicated by a few extensions of Goα immunopositive bipolar cell dendrites into the ONL in the treated group (a), those extensions were absent in control animals (b). ON rod bipolar cells were stained in green with PKCα antibody and both rod and cone ON bipolar cells were stained in red with Goα antibody (c, d): co-immunolabelled dendrites extending into the ONL were observed in GES-treated mice (c) and not in controls (d) indicating a synaptic plasticity in rod bipolar cells. Scale bars represent 25 μm (ONL outer nuclear layer, OPL outer plexiform layer, INL inner nuclear layer, IPL inner plexiform layer, RGCL retinal ganglion cell layer)

Fig. 9

Taurine deficiency induces retinal ganglion cells loss in GES-treated adult mice. Retinal sections of GES-treated mice (a) and control mice (b) were immunolabelled with antibodies directed to Brn-3A (a, b). Ganglion cells count showed a decrease in ganglion cell density in GES-treated mice as compared with controls (SEM, n = 8, P < 0.05, asterisk denotes Student’s t test) (c). Scale bars represent 25 μm (ONL outer nuclear layer, OPL outer plexiform layer, INL inner nuclear layer, IPL inner plexiform layer, RGCL retinal ganglion cell layer)

Fig. 10

Amacrine cells were not affected in GES mice. The density of amacrine cells did not differ between treated and untreated mice. In both groups (GES, a) (controls, b), amacrine cells were stained in red with anti-calretinin antibodies while GABA imunoreactive cells were stained in green with anti-GABA antibodies. GABA imunoreactive amacrine cells were co-labelled and appeared in yellow. Cell count did not show any difference in the density of both types of amacrine cells between the two groups of animals (c). Scale bars represent 50 μm (ONL outer nuclear layer, OPL outer plexiform layer, INL inner nuclear layer, IPL inner plexiform layer, RGCL retinal ganglion cell layer)