ABSTRACT
INTRODUCTION: In 2006 the anti-CD28 superagonistic IgG4 TGN1412, having passed pre-clinical safety screens, caused a severe 'cytokine storm' in 6 healthy volunteers. Others have shown that for TGN1412 to induce an inflammatory signal in human peripheral blood mononuclear cells (PBMCs) or in human diluted blood, endothelial cells or bound monoclonal antibody (mAb) is required as part of a bioassay complex. These types of protocols rely on different donor cells and therefore have limitations as bioassays for pre-clinical testing. METHODS: We performed studies using human PBMC/endothelial cell co-cultures, whole blood/endothelial cell co-cultures and human whole blood alone. We bracketed responses of a CD28 superagonist antibody with mAbs against CD52 (alemtuzumab, MabCampath-1H) or epidermal growth factor receptor (cetuximab, Erbitux) and with the immunostimulant lipopolysaccharide. We detected cytokine responses at the level of protein release (using ELISAs and Luminex assays) and gene induction (using real-time PCR arrays). RESULTS: Here we confirm that IL-8 release was induced in a mixed endothelial cell-PBMC system by the anti-CD28 mAb. We go on to show that an alemtuzumab and an anti-CD28 mAb, but not cetuximab induced the release of a range of cytokines including IL-8, IL-6, IFNγ, IL-2 and IL10 after 24h and induced cytokine gene induction after 1h. Co-cultures of whole blood and HUVECS showed larger variability but no superiority over whole blood alone at a range of time points (0.5-48h). DISCUSSION: We suggest that, whilst limitations exist, human blood-based in vitro assays may prove useful in assessing the potential of mAbs and other biotherapeutics to cause release of cytokines in humans.
