Diabetic retinopathy remains a leading cause of blindness despite recent advance in therapies. Traditionally, this complication of diabetes was viewed predominantly as a microvascular disease but research has pointed to alterations in ganglion cells, glia, microglia, and photoreceptors as well, often occurring without obvious vascular damage. In neural tissue, the microvasculature and neural tissue form an intimate relationship with the neural tissue providing signaling cues for the vessels to form a distinct barrier that helps to maintain the proper neuronal environment for synaptic signaling. This relationship has been termed the neurovascular unit (NVU). Research is now focused on understanding the cellular and molecular basis of the neurovascular unit and how diabetes alters the normal cellular communications and disrupts the cellular environment contributing to loss of vision in diabetes.

Retinal microglia were selectively and sequentially labelled in different layers of the retina of postnatal rats following a single intravenous injection of the fluorescent dye, rhodamine isothiocyanate (RhIc). The fluorescent cells were doubly immunostained with OX-42 and ED-1 antibodies that recognise complement type 3 (CR3) receptors and macrophage antigen, respectively. RhIc was first detected in the retinal blood vessels 5 min after injection. At 1 h, a variable number of microglia in the inner layers of the retina, namely, the nerve fibre and ganglion cell layers appeared to emit weak fluorescence. Labelled microglial cells in the inner nuclear and outer plexiform layers were not detected until 1 and 2 d had elapsed following RhIc injection. The number of labelled retinal microglia was progressively increased with time, peaking at 4 d after RhIc injection. The frequency of RhIc labelled cells also increased with age, with the largest number of cells occurring in 7-d-old rats but declined thereafter. In 11 d or older rats, RhIc was confined to the retinal blood vessels. It is concluded that when injected into the circulation, RhIc could readily gain access into the retina tissues due to an inefficient blood-retina barrier in early postnatal stages. It became impeded with maturation of the blood-retina barrier, which was established between 11 and 13 d of age. RhIc that inundated the retinal tissues was thoroughly sequestered by the resident microglial cells. It is therefore suggested that the latter could play a protective role against serum-derived substances that may be deleterious to the developing retina.

There is evidence that the endothelial cell (EC) glycocalyx is a significant determinant of vascular permeability, acting as a charge-size filter to permeant molecules. We have therefore examined its oligosaccharide composition in 3 classes of microvessel with differing permeabilities. EC in rat brain, retina and myocardium were labelled with a panel of lectins and subjected to a semiquantitative analysis. Surprisingly, no substantial differences were evident for any lectin labelling between the 3 microvessel types despite their marked morphophysiological diversity. In particular, all showed substantial sialic acid expression, with Maackia amurensis (MAA) labelling sialic acid in an α2–3 linkage to β-galactose and Sambucus nigra (SNA) recognising sialic acid in an α2–6 linkage to β-galactose. Arachis hypogaea (PNA) binding after neuraminidase digestion indicated the presence of Gal β1–3GalNAc attached to terminal sialic acid. The results therefore show that the sequences NeuNAc α2–3Gal β1–3GalNAc and NeuNAc α2–6Gal β1–3GalNAc are strongly expressed in the 3 microvessel types irrespective of their permeability properties. This homogeneity suggests that these lectin ligands may be involved in a common set of EC functions, e.g. cell[ratio ]cell and cell[ratio ]matrix interactions. However, we cannot rule out the possibility that glycocalyx differences may exist between vessels in the paracellular cleft which may alter its filtration properties.