In an earlier report, we demonstrated that subcutaneous injection of formalin in the rat hindpaw evokes a characteristic pattern of expression of the fos protein product of the c-fos protooncogene in spinal cord neurons, and that systemic morphine reversed the fos-like immunoreactivity in a dose-dependent, naloxone-reversible manner. The present study compared the effects of intracerebroventricular administration of the mu-selective opioid ligand [D-Ala2, NMe-Phe4, Gly-ol5] enkephalin, on the pain behavior and spinal cord fos-like immunoreactivity produced by subcutaneous formalin. Formalin injection produced a biphasic pain behavioral response which lasted about 1 h. There was a significant correlation between the formalin pain score and overall fos-like immunoreactivity in the lumbar enlargement. The greatest numbers of labeled cells and most intense fos-like immunoreactivity were found in laminae I, IIo and V of the L4-5 segments, ipsilateral to the formalin-injected paw. Considerable staining was also found in the ipsilateral ventral horn laminae VII and VIII. [D-Ala2, NMe-Phe4, Gly-ol5]enkephalin produced a dose-related, naloxone-reversible inhibition of both the formalin-evoked pain behavior and fos expression in the cord. The behavioral response to formalin, however, could be completely blocked without eliminating the expression of fos in spinal neurons. Moreover, subpopulations of neurons were differentially regulated. Thus, 100% inhibition of pain behavior was produced at a dose of [D-Ala2, NMe-Phe4, Gly-ol5]enkephalin which reduced fos-like immunoreactivity in the superficial laminae by only 64% and in the neck and ventral cord by 85%. Furthermore, the dose of [D-Ala2, NMe-Phe4, Gly-ol5]enkephalin which produced approximately 50% inhibition of fos-like immunoreactivity in the neck and ventral regions of the spinal cord was without effect in the superficial dorsal horn. Since the potencies for inhibition of pain behavior and fos-like immunoreactivity in the neck and ventral horn were comparable, these data suggest that the activity of neurons in these regions is directly related to the pain behavior produced by nociceptive inputs. Finally, we found that bilateral, midthoracic lesions of the dorsal part of the lateral funiculus blocked both the antinociception and fos suppression produced by intracerebroventricular [D-Ala2, NMe-Phe4, Gly-ol5]enkephalin. These results are consistent with the hypothesis that the analgesic action of supraspinally administered opiates results from an increase in descending inhibitory controls that regulate the firing of subpopulations of spinal cord nociresponsive neurons.