Regulatory T cell (Treg) therapy emerges for various indications associated with a breakdown of immune tolerance. Antigen-specific CAR-Tregs are frontrunners for transplantation and auto-immune diseases and are currently clinically evaluated. We aimed to link CAR antigen-engagement with immunosuppressive cargo release into the local microenvironment to boost efficacy and reduce side effects. We used our HLA-A*02 CAR and immunosuppressive IL-10 as model components to generate human CAR-Tregs that release IL-10 upon CAR-engagement. They were compared to CAR-Tregs with constitutive or no IL-10 expression by evaluating phenotypes, antigen-specific IL-10 release, and suppression of effector cell proliferation in vitro and performance in vivo in a humanised xeno-GvHD model. We demonstrated successful multi-construct engineering of CAR-Tregs, which released upon CAR-engagement 2.5-fold more IL-10 than CAR-Tregs lacking the corresponding antigen-specific IL-10 secretion module. Neither phenotype nor function was affected by expressing this module. In the xeno-GvHD model, we showed the beneficial effect of IL-10 release, particularly evident when compared to constitutive IL-10 expression that impaired CAR-Treg efficacy. We provide first proof-of-principle for engineering human CAR-Tregs to release an immunosuppressive cytokine upon CAR-engagement. This approach will both enhance the potency of CAR-Tregs at the intended target sites and limit their off-target effects.