On the organization of astrocytic gap junctions in rat brain as suggested by LM and EM immunohistochemistry of connexin43 expression

Abstract Gap junctions and the intercellular communication syncytium they form between glial cells are thought to play a critical role in glial maintenance of appropriate metabolic environments in neural tissues. We have previously suggested (Yamamoto et al., Brain Res. 508 :313–319, '90) that the vast majority of astrocytes in rat brain express connexin43, one of several recently identified gap junction proteins. Here, we confirm ultrastructurally that astrocytes in a number of brain regions of rat are immunolabelled with an antibody against connexin43 and that neurons and oligodendrocytes are devoid of labelling. The distribution of connexin43 immunoreactivity throughout the brain is presented at the light microscope (LM) level. By LM, immunoreactive structures consisted primarily of round or elongated puncta ranging from 0.3 μm to 4 μm in length and of annular profiles ranging from 1 to 10 μm in diameter. Immunolabelled fibrous processes were only occasionally seen and no labelling was observed in astrocytic cell bodies. Long, linear arrays of puncta were rare in gray matter but were common in white matter where they were arranged parallel to myelinated fibers. Puncta organized in a honeycomb pattern were seen near the cerebral cortical surface and frequently around blood vessels. Regional immunoreaction density, which was a reflection of either the concentration or staining intensity of immunoreactive elements, was remarkably heterogeneous; dramatic differences in labelling intensity frequently delineated anatomical boundaries between adjacent nuclei. Abrupt as well as graded fluctuations of immunoreaction intensity were also observed within nuclear structures. By electron microscopy (EM), gap junctions of fibrous and protoplasmic astrocytes were intensely stained and labelled organelles were often observed intracellularly in areas near gap junctions. These junctions and the spread of immunoreaction product to perijunctional organelles in their vicinity were considered to correspond to puncta seen by LM. Labelling within astrocytic cell bodies was seen in only a few instances. In some brain areas, astrocytic processes commonly gave rise to immunoreactive lamellae that partially ensheathed neuronal cell bodies, axon terminals, dendrites, and synaptic glomeruli. Such lamellae were considered to correspond to immunoreactive annular profiles seen by LM. Perivascular endfoot processes of astrocytes displayed intense staining of their gap junctions and portions of their apposing membranes. To the extent that regional differences and ultrastructural localization of connexin43 immunoreactivity reveal global distributions of astrocytic gap junctions composed of connexin43, we suggest (1) that the absence of widespread cytoplasmic labelling within astrocytes may be indicative of local connexin43 synthesis near gap junctions, (2) that small scale heterogeneities within nuclei may reflect relationships between the organizational patterns of glial gap junctions and neural systems, and (3) that enormous variations in demands for gap junctional coupling within astrocytic “functional syncytia” of different brain regions may lead to the establishment of largely separate syncytial compartments.