Evidence for the co-localization of another connexin with connexin-43 at astrocytic gap junctions in rat brain.

Gap junctions between astrocytes as well as between astrocytes and oligodendrocytes in rat brain were immunohistochemically labelled with a monoclonal and an affinity-purified polyclonal antibody generated against connexin-26. By light microscopy, the immunolabelling patterns obtained were, with a few exceptions, remarkably similar to previously described distribution patterns of the gap junctional protein connexin-43, which is expressed by astrocytes and is localized at astrocytic gap junctions. By electron microscopy, immunoreactivity with these two anti-connexin-26 antibodies was restricted to astrocytes; inter-astrocytic gap junctional membranes were symmetrically labelled, heterologous oligo-astrocytic junctional membranes were asymmetrically labelled only on the astrocyte side and oligo-oligodendrocyte junctions were unlabelled. Two additional anti-connexin-26 antibodies that were found to produce punctate labelling in leptomeninges and liver failed to do so in brain parenchyma, consistent with reports indicating the absence of authentic connexin-26 in this tissue. Antibodies that labelled astrocytic gap junctions exhibited no cross-reaction with connexin-43 or connexin-32, as demonstrated by western blotting, but recognized liver connexin-26 as well as several brain proteins, including an approximately 32000 mol. wt protein that did not correspond to connexin-32 and a 26000 mol. wt protein that co-migrated with liver connexin-26. These results suggest that connexin-26, or more likely a protein having sequence homology with connexin-26, is targeted to astrocytic gap junctions and raise the possibility of the existence of connexins that may be co-expressed with connexin-43 in most, but perhaps not all, astrocytes.