We have used in situ hybridization with a digoxigenin-labelled probe for tenascin-C mRNA and immunocytochemistry with antibodies against tenascin-C, glial fibrillary acidic protein, OX-42 and the 200 kDa neurofilament protein to study the expression, distribution and cellular relationships of tenascin-C mRNA and protein in the developing (postnatal) and adult spinal cord of rat, and the effects thereon of dorsal root, ventral root and sciatic nerve injuries. The most interesting finding was that on postnatal day 7 (P7), P14 and in the adult, but not on P0 or P3, a group of neurons in the lumbar ventral horn expressed the tenascin-C mRNA gene. They represented about 5% of ventral horn neurons in the adult and were among the smaller such neurons. Since 40-60% of such cells were lost at P13 following sciatic nerve crush on P0, some were almost certainly motor neurons. In addition, we found that at P0 and P3, mRNA-containing glial cells were widespread in grey and white matter but sparse in the developing dorsal columns; tenascin-C immunofluorescence showed a similar distribution. By P7 there were fewer mRNA-containing cells in the ventral horns and in the area of the dorsal columns containing the developing corticospinal tract where immunofluorescence was also weak. At P14 there were no glial-like mRNA-containing cells in the grey matter; such cells were confined to the periphery of the lateral and ventral white columns but were present throughout the dorsal columns where tenascin-C immunofluorescence was also strong. No glial-like mRNA-containing cells were present in the adult lumbar spinal cord and tenascin-C immunofluorescence was confirmed to irregular patches in the ventral horn, especially around immunonegative cell bodies of small neurons, a zone around the central canal, and a thin zone adjacent to the glia limitans. Thus the expression of tenascin-C is differentially developmentally regulated in the grey matter and in different parts of the white matter. Three days after injury of dorsal roots L4-6, many cells containing tenascin-C mRNA, some identified as glial fibrillary acidic protein-positive astrocytes, were present in the ipsilateral dorsal column, but were rare after longer survivals. Immunoreactivity, however, was elevated in the ipsilateral dorsal column at 3 days, remained high for several months and disappeared at 6.5 months. Dorsal root injury had no effect on tenascin-C mRNA or protein in the grey matter. Sciatic nerve or ventral root injury had no effect on these molecules in any part of the spinal cord.