Extracellular matrix organization in various regions of rat brain grey matter

Previous studies revealed the concentration of extracellular matrix proteoglycans in the so-called perineuronal nets on the one hand and in certain zones of the neuropil on the other. This nonhomogeneous distribution suggested a non-random chemical and spatial heterogeneity of the extracellular space. In the present investigation, regions dominated by one of both distribution patterns, i.e. piriform and parietal cortex, reticular thalamic nucleus, medial septum/diagonal band complex and cerebellar nuclei, were selected for correlative light and electron microscopic analysis. The labelling was performed by the use of the N-acetylgalactosamine-binding plant lectin Wisteria floribunda agglutinin visualized by peroxidase staining and additionally by photoconversion of red carbocyanine fluorescence labelling for electron microscopy. The intense labelling of the neuropil of a superficial piriform region, presumably identical with sublayer Ia, was confined to a fine meshwork spreading over the extracellular space between non-myelinated axons, dendrites and glial profiles. In the reticular thalamic nucleus the neuronal cell bodies were embedded in zones of labelled neuropil. In contrast to these patterns, the labelled extracellular matrix in different cortical layers and in the other subcortical regions was concentrated in perineuronal nets as large accumulations at surface areas of the neuronal perikarya and dendrites and the attached presynaptic boutons. Astrocytic processes usually were separated from the neuronal surface by the interposed extracellular material. Despite a great variability, the width of the extracellular space containing the labelled matrix components in all perineuronal nets appeared to be considerably larger than that in the labelled zones of neuropil and the non-labelled microenvironment of other neurons. Our results support the view that differences expressed in topographical and spatial peculiarities of the extracellular matrix constituents are related to neuron-type- and system-specific functional properties.