BACKGROUND: CTLA4-blocking antibodies induce tumor regression in a subset of patients with melanoma. Analysis of immune parameters in peripheral blood may help define how responses are mediated. METHODS: Peripheral blood from HLA-A*0201-positive patients with advanced melanoma receiving tremelimumab (formerly CP-675,206) at 10 mg/kg monthly was repeatedly sampled during the first 4 cycles. Samples were analyzed by 1) tetramer and ELISPOT assays for reactivity to CMV, EBV, MART1, gp100, and tyrosinase; 2) activation HLA-DR and memory CD45RO markers on CD4+/CD8+ cells; and 3) real-time quantitative PCR of mRNA for FoxP3 transcription factor, preferentially expressed by T regulatory cells. The primary endpoint was difference in MART1-specific T cells by tetramer assay. Immunological data were explored for significant trends using clustering analysis. RESULTS: Three of 12 patients eligible for immune monitoring had tumor regression lasting > 2 years without relapse. There was no significant change in percent of MART1-specific T cells by tetramer assay. Additionally, there was no generalized trend toward postdosing changes in other antigen-specific CD8+ cell populations, FoxP3 transcripts, or overall changes in surface expression of T-cell activation or memory markers. Unsupervised hierarchical clustering based on immune monitoring data segregated patients randomly. However, clustering according to T-cell activation or memory markers separated patients with clinical response and most patients with inflammatory toxicity into a common subgroup. CONCLUSION: Administration of CTLA4-blocking antibody tremelimumab to patients with advanced melanoma results in a subset of patients with long-lived tumor responses. T-cell activation and memory markers served as the only readout of the pharmacodynamic effects of this antibody in peripheral blood. CLINICAL TRIAL REGISTRATION NUMBER: NCT00086489.