COX-2/PGE2 driven, photoimmunotherapy-induced edema is reduced by prophylactic NSAIDs in mouse tumor models

ASP-1929 photoimmunotherapy (PIT) is an investigational therapy that combines binding of ASP-1929, an anti-EGFR antibody conjugated to IRDye®700DX (IR700), with red-light illumination for selective cell killing. ASP-1929 PIT has shown preliminary clinical activity for head and neck cancer (HNC) and although it has a manageable safety profile, edema is a commonly reported adverse event, with severe laryngeal edema noted in some patients. Thus, a critical need exists to reduce ASP-1929 PIT-associated edema without compromising therapeutic efficacy. Here, we elucidate the underlying mechanism of PIT-induced edema in vitro and in vivo, expand upon our previous finding for the use of prophylactic COX-2 inhibition to reduce edema, and translate our findings to a clinically relevant HNC xenograft model. Briefly, CT26-EphA2 or LL2-EphA2 cells were incubated with anti-EphA2-IR700 ± light and evaluated for changes in COX-2 expression using real-time PCR and Western blot. Results show an increase in COX-2 post-PIT which was unaffected by co-incubating with COX antagonists, robenacoxib (selective COX-2) or flurbiprofen (pan COX). Considering prostaglandin E2 (PGE2) is downstream of COX-2 and a known effector of vascular permeabilization, we evaluated PGE2 secretion in cell culture media by ELISA and observed an induction in PGE2 levels immediately post-PIT, which was completely blocked with either COX inhibitor. To evaluate the effect of pan COX vs selective COX-2 inhibition on edema reduction post-PIT, mice bearing CT26-EphA2 or LL2-EphA2 tumors were administered anti-EphA2-IR700 ± light. In agreement with our previous data, edema peaked 6h post-PIT and was significantly reduced when mice were pre-treated with either COX inhibitor (p≤0.0001). To confirm the role of COX-2, we evaluated tumor lysates and found a significant increase in COX-2 expression post-PIT, with no effect on COX-1, and no reduction in the presence of either COX inhibitor. In alignment with our in vitro data, increased secretion of PGE2 was confirmed by evaluating the PGE2 metabolites in circulation post-PIT and this effect was rescued in the presence of COX inhibition. Lastly, we developed a HNC xenograft model using human hypopharyngeal cancer cells and evaluated the effect of COX inhibition on edema formation post-PIT. As in the syngeneic models, edema peaked at 6h post-PIT and was rescued with either COX inhibitor (p≤0.0001). Importantly, COX inhibition did not compromise anti-tumor efficacy. In conclusion, this is the first report showing increased COX-2/PGE2 post-PIT, linking PGE2 secretion with edema generation, and suggests COX-2 inhibition to combat ASP-1929 PIT-induced edema for improved patient safety. The translation of this result to the clinical setting is under investigation at treatment centers throughout Japan (jRCTs031250140).