Aims/hypothesis. Neurodegenerative changes in the diabetic retina occurring before diabetic retinopathy could be inevitable by the altered energy (glucose) metabolism, in the sense that dynamic image-processing activity of the retinal neurons is exclusively dependent on glucose. We therefore investigated the morphological changes in the neural retina, including neuronal cell death, of a streptozotocin-induced model of diabetes. Methods. Streptozotocin was intravenously injected. Rats were maintained hyperglycaemic without insulin treatment for 1 week and 4, 8, 12, and 24 weeks, respectively. Diabetic retinas were processed for histology, electron microscopy, and immunohistochemistry using the TUNEL method. Results. A slight reduction in the thickness of the inner retina was observed throughout the diabetic retinas and a remarkable reduction was seen in the outer nuclear layer 24 weeks after the onset of diabetes.The post-synaptic processes of horizontal cells in the deep invaginations of the photoreceptors showed degeneration changes from 1 week onwards. A few necrotic ganglion cells were observed after 4 weeks. At 12 weeks, some amacrine cells and a few horizontal cells showed necrotic features. Three to seven cellular layers in the outer nuclear layer and nerve terminals, rolled by the fine processes of the Müller cells near the somata of the degenerated ganglion cells, were apparent at 24 weeks. Apoptosis appeared in a few photoreceptor cells at 4 weeks, and the number of apoptotic photoreceptors increased thereafter. Conclusion/interpretation. These findings suggest that the visual loss associated with diabetic retinopathy could be attributed to an early phase of substantial photoreceptor loss, in addition to later microangiopathy. [Diabetologia (2003[Diabetologia ( ) 46:1260[Diabetologia ( -1268

AII-amacrine cells were characterized from Golgi-stained sections and wholemounts of the macaque monkey retina. Similar to other mammalian retinae, they are narrow-field, bistratified amacrine cells with lobular appendages in the outer half of the inner plexiform layer (IPL) and a bushy, smoother dendritic tree in the inner half. AII cells of the monkey retina were stained immunocytochemically with antibodies against the calcium-binding protein calretinin. Their retinal mosaic was elaborated, and their density distribution across the retina was measured. Convergence within the rod pathway was calculated. Electron microscopy of calretinin-immunolabelled sections was used to study the synaptic connections of the AII cells. They receive a major input from rod bipolar cells, and their output is largely onto cone bipolar cells. Thus, the rod pathway of the primate retina follows the general mammalian scheme as it is known from the cat, the rabbit, and the rat retina. The spatial sampling properties of macaque AII-amacrine cells are discussed and related to human scotopic visual acuity.

Two immunocytochemical markers were used to label the rod pathway of the rat retina. Rod bipolar cells were stained with antibodies against protein kinase C and AII-amacrine cells with antibodies against parvalbumin. The synaptic circuitry of rod bipolars in the inner plexiform layer (IPL) was studied. Rod bipolar cells make approximately 15 ribbon synapses (dyads) in the IPL. Both postsynaptic members of the dyads are amacrine cells; one is usually the process of an AII-amacrine cell and the other one frequently provides a reciprocal synapse. No direct output from rod bipolar cells into ganglion cells was found. AII-amacrine cells make chemical output synapses with cone bipolar cells and ganglion cells in sublamina a of the IPL. They make gap junctions with cone bipolar cells and other AII-amacrine cells in sublamina b of the IPL. The rod pathway of the rat retina is practically identical to that of the cat and of the rabbit retina. It is very likely that this circuitry is a general feature of mammalian retinal organization.

Parallel pathways for visual information processing start at the first synapse of the retina, at the cone pedicle. At least eight different types of bipolar cells receive direct synaptic input from an individual cone. The present study explores whether enough synaptic sites are available at the cone pedicle to supply all these bipolar cells. Monkey retinae were optimally fixed for electron microscopy. Serial horizontal sections were cut through the cone pedicle layer in a piece close to the fovea (eccentricity: 0.75 mm) and in a peripheral piece (eccentricity: 5-6 mm). The ribbon synapses (triads) at the cone pedicle base were analysed. The average number of synaptic ribbons per cone pedicle increased from 21.4 +/- 1.6 (n = 26) in central retina to 41.8 +/- 3 (n = 14) in peripheral retina. Five central and five peripheral pedicles were reconstructed and the invaginating bipolar cell dendrites forming the central elements of the triads were characterized. Close to the fovea an average of 18 invaginating bipolar cell dendrites was found, in peripheral retina the average was 90. Pedicles of foveal cones have one invaginating central process per ribbon, pedicles of peripheral cones have two. It is possible that midget bipolar cell dendrites occupy the majority of triads in the fovea, while in peripheral retina both midget and diffuse bipolar cell dendrites share the triads.