Artificial T-cell receptors (TCR) are generated by connecting an antigen recognizing ectodomain to a signal transducing endodomain. Most frequently the variable chains of Immunoglobulin molecules expressed as a single chain (ScFv) are utilized as ectodomains and the intracellular portion of CD3-ζ is used as endodomain. When expressed by primary T-cells these molecules can redirect the cellular immune response to almost any surface target molecule for which a monoclonal antibody can be made. However, clinical studies with these chimeric T-cells have been disappointing, with no clear clinical benefit, and only minimal in vivo persistence of infused T-cells. Transmitted CD3-ζ signal is only sufficient to activate cell-killing and Inteferon-γ release but fails to induce IL-2 release or proliferation. Full T cell activation requires co-stimulatory signals that are rarely provided by the tumor cells and therefore may need to be incorporated in the endodomain of the artificial TCR. Indeed, inclusion of a CD28 signaling component resulted in IL-2 release and limited proliferation, but T cell activation appears still incomplete. OX40 is a TNFR family molecule expressed by activated T-cells. It transmits a potent and prolonged activation signal and has been found to be an important molecule for maintaining a prolonged immunological response e.g. in chronic inflammation. We held the hypothesis that an artificial TCR providing 3 signals - CD3-ζ, CD28 and OX40 in cis would result in more potent activation and more prolonged proliferation. We generated and compared a number of constructs based on GD2 recognizing scFv 14g2a: 14g2a-ζ, 14g2a-CD28-ζ, 14g2a-OX40-ζ, 14g2a-CD28-OX40-ζ. We first co-immunoprecipitated TRAF-2 with OX40 containing constructs. This demonstrated that the OX40 binding site was unaffected by fusion with other proteins. Incorporating 3 signals - CD3ζ, CD28 and OX40 in cis from a single endodomain of an artificial TCR recruited a 10 fold higher level of NFkB quantified by Luciferase-reporter than two signals (14g2a.CD28-ζ) and over 50 fold higher than a single signal (14g2a.ζ). T-cells transduced with all of these constructs were capable of lysing GD2+ neuroblastoma cells. Only limited expansion (1.6 fold, range 0.9–3) was induced upon stimulation with tumor cells in T cells transduced with 14g2a.OX40.ζ. Adding a CD28 domain resulted in a 5.2 fold (range: 1.6–7.2) expansion within 7 days but this proliferation could not be maintained. In contrast, 14g2a.CD28.OX40.ζ transduced T cells expanded 10.7 fold (range: 4–17) within 7 days and continued to proliferate with weekly stimulations with tumor cells, even in the absence of exogenous IL-2. This increased proliferation of 14g2a.CD28.OX40.ζ transduced T cells was accompanied by a >10-fold increase in IL-2 and 5-fold increase in TNF-a secretion as compared to the 14g2a.CD28-ζ construct. Sustained proliferation was accompanied by persisting function - T-cells transduced with 14g2a.CD28-OX40-ζ were still capable of killing GD2+ targets after 35 days of culture. These improved functional characteristics should favor the overall utility of chimeric T-cells.

Author notes

Corresponding author

Sign in via your Institution