Identification of MHC-restricted antigens and progress in the induction and control of adaptive cytotoxic immune responses have led to renewed interest in immunotherapy as a treatment for severe pathologies such as cancer and autoimmune diseases. Reliable procedures for detecting and monitoring T cell responses induced by the treatment throughout a clinical trial are needed in order to design rational protocols with increased efficiency. We have attempted to develop such a procedure by combining T cell sorting using HLA-peptide complexes multimerized on magnetic beads together with the quantitative Immunoscope approach. Once a recruited patient has been typed for HLA and target antigens, relevant HLA--peptide multimers can be selected and used for sorting specific peripheral T cells prior to any treatment and at the peak of the expected response to treatment. Clonotypic primers specific for the TCR rearrangements of the specific T cell clones can then be designed and used for measuring the frequency of their TCR transcripts by quantitative PCR on blood samples or T cell subsets throughout the trial. In reconstruction experiments as well as in samples from one rheumatoid arthritis patient, we were readily able to detect and follow several T cell clones with a frequency as low as 10(-5) among CD8+ T cells. The main advantages of this procedure over other currently available assays are that it does not require any assumptions on the functional status of the specific T cells and it permits the monitoring of individual T cell clones whose phenotypic shift can thus be evaluated.