Our previous study using bromodeoxyuridine (BrdU) has shown that T cells in lesions of experimental autoimmune encephalomyelitis (EAE) in the rat central nervous system (CNS) lose their proliferating capability immediately after infiltration into the CNS. To characterize the nature of this phenomenon in more detail, we have isolated T cells from EAE lesions and examined their surface phenotype and response to encephalitogenic antigen, myelin basic protein (MBP). By flow cytometry (FCM) analysis, it was revealed that compared with peripheral blood lymphocytes, up-regulation of interleukin-2 (IL-2) receptors (0.06% --> 3.73%) and the lymphocyte function-associated antigen-I (LFA-1) molecules (0.76% --> 17.6%) on spinal cord T cells (SCT) was observed. In spite of the latter finding suggesting that SCT are activated, SCT recovered from rats with full-blown EAE responded very poorly to MBP. The addition of thymocytes or thymocytes plus astrocytes did not alter the low responsiveness of SCT. More importantly, astrocytes strongly suppressed the response of lymph node T cells to MBP. Using MBP-specific T-line cells, it was revealed that T-cell suppression might be induced by incomplete presentation of MBP and release of suppressive humoral factors by astrocytes. Since the response of SCT was still poor when assayed after three and 12 rounds of stimulation with the antigen and propagation with IL-2, this phenomenon is long lasting. These findings are consistent with the findings obtained by the BrdU study that infiltrating T cells into the CNS do not proliferate vigorously. Taken together, the poor response of infiltrating T cells to MBP would be induced by co-existing cells such as astrocytes although the T cells are in an active form as judged by their surface phenotype. The present study suggests that activation of non-haematopoietic parenchymal cells in each organ by infiltrating T cells and subsequent inactivation of the T cells are important healing processes for organ-specific autoimmune diseases.