Plasma levels of soluble interleukin-2 receptor (sIL-2R), soluble CD8 (sCD8) and soluble intercellular adhesion molecule 1 (sICAM-1) were determined by ELISA assays in about 100 patients with hairy cell leukemia (HCL), acute myelomonocytic leukemia (AMMoL), acute myelocytic leukemia (AML), chronic lymphocytic leukemia (CLL), prolymphocytic leukemia (PLL), acute lymphoblastic leukemia (ALL), adult T-cell leukemia (ATL), and mycosis fungoides (MF). Additionally, cultured AML, ALL, and CLL cells grown with and without 12-0-tetra-decanoyl-phorbol-13-acetate (TPA) were tested for IL-2R (CD25) expression by indirect immunofluorescence. Supernatants of these cultures were also tested for sIL-2R by ELISA. Elevated sIL-2R levels were found in HCL patients at initial diagnosis and relapse, in AMMoL, in AML, in the accelerated and non-accelerated phases of B-CLL, in PLL, in non-T/non-B ALL, in B-ALL in mixed lineage ALL, in T-CLL, in T-ALL, and in active MF. Reduced levels of sIL-2R were encountered in HCL patients in remission, in pre-T-ALL, and in MF patients in remission. Also, in non-accelerated CLL sIL-2R levels were less elevated than in later stages of the disease. In T-CLL, sIL-2R was only slightly elevated. Thus, we believe sIL-2R could prove to be a useful marker of disease stage, subtype, and prognosis in several hematologic malignancies. The cultures with and without TPA suggested that the undetermined source of sIL-2R in HCL, ALL and AML could indeed be the malignant cells but perhaps not so in the case of B-CLL. Plasma sCD8 was found to be below normal control levels in HCL, and lowest in relapsing cases. In addition, sCD8 levels were below normal in pre-T-ALL, and in MF. Levels in the non-accelerated phase of B-CLL approximated those of controls. Elevated levels of sCD8 were observed in AML, AMMoL, accelerated stage B-CLL, PLL, non-T/non-B ALL, B-ALL, mixed lineage ALL, T-ALL, T-CLL, and ATL. Thus, in a few instances, sCD8 may also correlate with disease subtype, as well as stage. Although sICAM-1 levels were elevated in all leukemias, its levels in CLL did not appear to be related to disease activity. Whether this is true or not for other leukemias would require additional work on sICAM-1 levels and its relationship to disease activity and prognosis.