To investigate the effect of targeting immunotherapy for leukemia cells by using cytotoxic T lymphocyte (CTL) specifically against WT (Wilm's tumor) 1-derived peptide. A 9-mer WT1 peptide (CMTWNQMNL) containing HLA-A*0201-binding anchor motifs was synthesized. Dendritic cells (DCs) generated from the peripheral blood mononuclear cells of an HLA-A*0201-positive healthy donor were cultured and divided into 2 groups: experimental group to be loaded with Wilms' tumor 1 (WT1) peptide so as to elicit CTL's specifically for WT1 peptide, restricted by HLA-A*0201, and control group. Six days later rhTNF-alpha was added for 3 days more to promote the maturation of DCs. Before loading of WT1 peptide and 2 days after the addition of rhTNF-alpha direct immunofluorescence labeling method was used with PE or FITC labeled mono-antibodies to detect the expression of the surface antigens: CD83, CD1alpha, CD80, CD86, CD14, and HLA-DR. DCs suspended and attached to wall were collected and then divided into 2 groups: pure T cell group (group D) to be cultured in culture medium without IL-2, and IL-2 + T cell group (Group C) to be cultured in 1640 culture medium with IL-2. Eight days later the T cells of Group C were co-cultured with the DCs of the experimental group (WTI peptide + DC + IL-2 + T cells, Group A) or the DCs of the control group (DC + IL-2 + T cells, Group B). Five days later the killing activity was detected. The CTLs of Groups A, B, and C at logarithmic growth phase were mixed with leukemic cells of the lines: NB4/WT1D, NB4WT1A, NB4 (all HLA-A*0201 +, WT1 +), U937 (HLA-A*0201 +, WTl -), and K562 (HLA-A*0201 -, WTI +), and mononuclear cells of the bone marrow of leukemic patients at different effector cell-target cell of 20:1 and 10:1. MTT method was used to examine the killing rate of CTL to the target cells. (1) The killing rates of Group A cells to NB4/WT1 D, NB4WTA, and NB4, leukemic cells were 60.4% +/- 3.1%, 56.4% +/- 5.7%, and 55.0% +/- 3.7%, all significantly higher than those of the Group B cells (10.9% +/- 1.6%, 11.1% +/- 2.7%, and 11.9% +/- 2.5%), and those of Group C cells (9.1% +/- 1.0%, 9.2% +/- 1.7%, and 9.4% +/- 1.8%) (all P < 0.01). There were no significant differences in the killing rates to U937 and K562 leukemic cells among the 3 groups. (2) When the effect-target ratio was 20: 2, the killing activity of the CTLs of Group A to the HLA-A*0201 +, WT1 + NB4/WT1 D, NB4/WT1A and NB4 leukemic cells was significantly higher than those to the HLA-A*0201 +, WT1 - U937 cells and the HLA-A*0201 -, WT1 + K562 cells (both P < 0.001), however, not significantly different from that to the U937 and K562 cells. (3) There were no significant differences in the killing activity of Group A cells to NB4/WT1D, NB4/WT1A, and NB4 cells (P = 0. 065, P = 0.621). (4) When the effect-target ratio was 10:2, the killing rates of Group A cells to the NB4/ WT1D, NB4/WT1A, and NB4 cells were 45.9% +/- 3.9%, 43.9% +/- 3.7%, and 44.1% +/- 3.2% respectively, all significantly lower than those when the effect-target ratio was 20:1 (all P < 0.01). CTLs specific for WT1 and restricted by HLA-A*0201 exert specific lysis upon leukemia cell lines and primary leukemia cells, but not upon normal hematopoietic cells, which provides a rationale for developing a strategy of WT1 peptide-based adoptive T-cell therapy and vaccination for human leukemia and solid tumors.