Abstract
The failure of toxicity studies in non-human primates to predict the cytokine release syndrome during a first-in-man study of the CD28-specific monoclonal antibody TGN1412 has remained unexplained so far. In this issue of the BJP, work from the NIBSC first identifies the effector memory subset of human T-lymphocytes as the most likely source of the pro-inflammatory cytokines released during the study, and goes on to show that in cynomolgus monkeys, this subset lacks CD28, the target molecule of TGN1412. We discuss the implications for the TGN1412 catastrophe and for preclinical evaluation of biologicals in animal models in general.