Human myelin basic protein (MBP) was inserted into phosphatidyl-serine liposomes and Hartley guinea pigs were treated with 1 or 2 injections of MBP-liposomes mixture by the intracardial route before an encephalitogenic challenge with MBP in CFA. Other groups of guinea pigs were pretreated with equivalent amounts of MBP in saline or of MBP in IFA. Of 24 untreated animals, 22 developed EAE and died or had to be sacrificed within 15 days after challenge. Only 4 of 29 guinea pigs treated with either 75 microgram or 112 microgram MBP-liposomes developed clinical signs of the disease. In the group pretreated with MBP in saline, 11 of 15 animals died, whereas in the MBP-IFA group, only 4 out of 10 died. Histologic modifications were also decreased in the MBP-liposomes and MBP-IFA groups, but in many instances, clinical normal guinea pigs showed disseminated perivascular infiltrates in the brain and spinal cord. Delayed hypersensitivity reactions to MBP were positive in all but 1 animal tested. Early T-rosette levels followed essentially the clinical course of the disease: they were drastically decreased 7 days after challenge in the untreated and MBP-saline-treated groups, but remained essentially normal in the MBP-liposomes group throughout the experiment. Lymphocyte transformation tests carried out in parallel with soluble MBP and with MBP-liposomes indicated that animals in certain groups responded preferentially to 1 form or the other of the antigen. Compared with prevailing procedures, the single i.v. injection of a relatively small amount of liposome-associated MBP appears to represent a promising approach for the antigen-specific suppression of EAE.