HTLV-1 is perhaps the most oncogenic virus in humans. Five percent to ten percent of individuals infected by HTLV-1 develop a fatal T-cell leukemia (ATL) after 2-3 decades of latency. Conventional chemotherapy only creates escape mutants in ATL. Novel treatments including retroviral drugs, Arsenic Trioxide/IFNa combination, or Allogeneic stem cell transplant have increased the 5-year survival, but there is still need to establish a cure. We chose an immunotherapy approach because all leukemic cells express some HTLV-1 viral proteins allowing to distinguish normal and HTLV-1 infected cells by the host immune system. In addition, anti-HTLV-1 immunity is suppressed in ATL patients. We aim to target HBZ (an antisense protein of HTLV-1) for this purpose because; 1) the frequency of T cells recognizing HBZ correlates well with low HTLV-1 proviral loads in contrast to T cells against Tax-1 or Env. 2) Unlike Tax-1, HBZ is constantly expressed by all stages of ATL cells. The challenge is that HBZ is an extremely weak immunogen. We identified a few epitopes of HBZ that stably find to class I HLA molecules. We tested them using an in vivo model of ATL in humanized mice. Immunization by HBZ showed little therapeutic effects in these mice. DC immunization showed prolonged survival, but did not protect mice from ATL. Thus, we grew anti-HBZ CD8 T cells by immunizing normal T cells ex vivo, enriched them by HBZ-tetramer sorting. T-cell therapy involving anti-HBZ CD8 T cells successfully protected host mice, prompting us to rely on an “engineered T cell” strategy as novel treatment of acute/lymphomatous ATL cases.