Objectives:(1) To determine the safety and feasibility of repetitive reinfusions of activated autologous CD8 cells followed by low-dose continuous interleukin (IL)-2 infusion in patients with AIDS. (2) To study the relationships between clinical responses, surface marker phenotypic distributions and cytokine expression patterns of both cultured CD8 cells and lymphocytes in the peripheral blood compartment.Design:Six adult patients with Centers for Disease Control and Prevention group IV HIV-1 disease ranging from mild to severe, were studied. All patients were receiving zidovudine prior to and during the study period, and had initial CD4 and CD8 cell counts >50 and 200×106/l, respectively.Methods:Autologous CD8 T cells (108-1010) were reinfused five times afterex vivoculture and stimulation with phytohemagglutinin and recombinant (r) IL-2. The fifth such infusion was followed by 5 days of rlL-2 infusion. Phenotypes and cytokine expression patterns of the expanded cells were determined as well as serum levels of immune mediators throughout the study.Results:Patients showed stable CD4 and CD8 cell counts, p24 antigenemia, and minimal toxicity over the 24-week protocol study. Clinical improvement was observed in lymphadenopathy (six out of six), oral hairy leukoplakia (three out of four), and Kaposi's sarcoma (KS; two out of two) in the patients studied. In vivo induction of detectable levels of bioactive acid-stable interferon (IFN)-a, but not of other cytokines studied, upon activated CD8 cell reinfusion was associated consistently with improvement of oral hairy leukoplakia. However, partial regression of KS was observed after the CD8 cell infusion cycles and without IFN-α induction. In one of the two patients studied, KS regression was associated with decreased IL-1 α serum levels. In the other patient, who had failed previous IFN-α therapy, KS regression was observed after a decline in reinfused CD8 cell-associated gene expression of tumor necrosis factor (TNF)-β. Both IL-1 α and TNF-β are growth factors for KS cells.Conclusions:These observations demonstrate the feasibility and safety ofex vivoCD8 cell activation, expansion, and reinfusion, and rlL-2 infusion in AIDS patients. The findings in this Phase I trial suggest potential clinical efficacy and encourage Phase II trials. The correlations obtained between clinical and immunological states could contribute to an understanding of the relationship between CD8 T-cell function and HIV-1-associated disease progression.