Comparison of same donor and mixed donor endothelial cell:PBMC co-culture assays to detect cytokine storm reactions to biologics

DM Reed1, KE Paschalaki2, NA Mohamed1, HH Gashaw1, LK Bailey1, RD Starke2, AM Randi2, TT Hansel3, JA Mitchell1. 1Dept. of Cardiothoracic Pharmacology, National Heart and Lung Institute, Imperial College London, London, UK, 2Vascular Science, National Heart and Lung Institute, Imperial College London, London, UK, 3Imperial Clinical and Respiratory Research Unit, London, UK

There is an urgent unmet clinical need for improved human tissue bioassays that can predict cytokine storm responses to biological drugs. Gold standard cytokine storm causing molecules such as the anti CD28 antibody, TGN1412 do not activate monocultures of peripheral blood mononuclear cells (PBMCs) or endothelial cells. TGN1412 does however activate co-cultures of PBMCs and endothelial cells, although these assays are comprised of heterologous cells, and may not optimally delineate mild and severe cytokine storm inducing drugs. Endothelial cells line blood vessels and as such are inaccessible. We have previously shown that use of endothelial cells from blood progenitors (blood outgrowth endothelial cells; BOECs) co-cultured with same donor PBMCs provides an assay that detects cytokine storm and has applications in personalised medicine (1) (patent 1305318.6). Here we have compared the ability of same donor endothelial cell:PBMC assays and mixed donor endothelial cells (using BOEC):PBMC assays to delineate severe (TGN1412-like, anti-CD28 drug) and mild (Campath, anti-CD52 drug) cytokine storm causing drugs from control drugs (Herceptin, Avastin, Arzerra) (all 10μg/ml). Both ‘same donor’ and ‘mixed donor’ endothelial cells and PBMCs, but not monocultures of either cell type alone, responded avidly by releasing CXCL8 (measured by ELISA) to a TGN1412-like drug at 24h. Mixed donor assays responded to Campath with comparable CXCL8 to TGN1412-like drug whilst same donor assays responded less to Campath. Both types of assay did not respond to the control antibodies Herceptin and Arzerra whilst mixed donor, and not same donor assays responded to the control drug, Avastin. Thus, mixed donor endothelial cell:PBMC co-cultures appear to respond to non-cytokine causing drugs as well as not delineating mild and severe cytokine storm causing drugs. We have shown, for the first time, therefore, that use of same donor endothelial:PBMC co-culture is critical for the accurate detection of cytokine storm causing drugs.

Figure 1: Effect of therapeutic antibodies on CXCL8 from (A) same donor and (B) mixed donor blood outgrowth endothelial cells (BOEC):PBMC co-cultures. Cells were treated with anti-CD28, Campath, Herceptin, Avastin or Arzerra for 24h (all 10μg/ml). Data are mean ± SEM (n=8 same donor assay, n=3 non-same donor assay). Data were analysed by one-way ANOVA followed by Bonferroni test (*p<0.05).

1. Reed, D. M., et al Toxicology Letters, Interlaken, Switzerland. S164.