Polyclonal CD8+CD45RClow/- Tregs are potent regulatory cells able to control solid organ transplantation rejection and even induce tolerance. However, donor major histocompatibility complex (MHC)-specific Tregs are more potent than polyclonal Tregs in suppressing T-cell responses and preventing acute as well as chronic rejection in rodent models. The difficulty of identifying disease-relevant antigens able to stimulate Tregs has reduced the possibility of obtaining antigen-specific Tregs. To bypass this requirement and gain the advantage of antigen specificity, and thus improve the therapeutic potential of CD8+ Tregs, we stably introduced a chimeric antigen receptor (CAR) derived from a HLA-A*02 antigen-specific antibody (A2-CAR) in human CD8+ Tregs and developed a clinically compatible protocol of transduction and expansion. We demonstrated that A2-CAR CD8+ Tregs were not phenotypically altered by the process, were specifically activated, and did not exhibit cytotoxic activity toward HLA-A*02+ kidney endothelial cells (ECs). We showed that A2-CAR CD8+ Tregs were more potent suppressors of immune responses induced by HLA-A*02 mismatch than control-CAR CD8+ Tregs, both in vitro and in vivo, in models of human skin graft rejection and graft-versus-host disease (GVHD) in NOD.Cg-PrkdcscidIl2rgtm1Wjl/SzJ (NSG) mice. We showed that integrity of human skin graft was preserved with A2-CAR CD8+ Tregs at least 100 days in vivo after administration, and that interaction between the A2-CAR CD8+ Tregs and HLA-A*02+ kidney ECs resulted in a fine-tuned and protolerogenic activation of the ECs without cytotoxicity. Together, our results demonstrated the relevance of the CAR engineering approach to develop antigen-specific CAR-CD8+ Tregs for clinical trials in transplantation, and potentially in other diseases.