Immunotherapy for cancer aims to generate cytotoxic cells that are capable of eradicating tumour cells. It has been well demonstrated that helper, non-cytotoxic CD4(+) T cells are important for the induction and maintenance of anti-tumour immunity exerted by cytotoxic CD8(+) T cells. In contrast, the existence of direct anti-tumour, effector cytotoxic CD4(+) T cells remains elusive, mainly due to the paucity of reliable experimental data, especially in human B-cell non-Hodgkin lymphomas. This study developed an appropriate, autologous follicular B-cell non-Hodgkin follicular lymphoma model, including the in vitro establishment of a malignant, human leucocyte antigen class I (HLA-I) deficient B-cell line, and the generation of three autologous anti-tumour cytotoxic CD4(+) T-cell clones originating from the peripheral blood of the same patient. These three clones were considered as tumour specific, because they were capable of killing the malignant, HLA-I-deficient B-cell line through a classical HLA-II restricted perforin-mediated pathway, but did not lyse the Epstein-Barr virus-infected autologous normal B lymphocytes. All three CD4(+)clones were T-cell receptor Vbeta17-Dbeta1-Jbeta1.2 and exhibited an identical complementarity-determining region 3, suggesting the immunodominance of a single peptide antigen presented by tumour cells. Such lymphoma models would provide a useful tool for in vivo expansion and the adoptive transfer of selected CD4(+) cytotoxic cells in immunotherapeutic strategies.