Background While chimeric antigen receptor T cell (CAR T) treatment has been efficacious in blood cancers, mirroring this success in solid tumors, like head and neck squamous cell carcinoma (HNSCC) has proven difficult due to the challenge of identifying and validating suitable target antigens. To rapidly model CAR-T treatment against HNSCC antigens, we have developed a robust, scalable model using fertilized chicken eggs to facilitate CAR-T screening against vascularized solid tumors. The fertilized chicken egg inner shell membrane, the chorioallantoic membrane (CAM), is a highly vascularized membrane which facilitates nutrient delivery and oxygen exchange for growth of the developing chick. The nutrient transfer capabilities of the CAM can be exploited to drive the formation of vascularized solid tumors which can be targeted by CAR T within one week of engraftment. In contrast to murine systems which can be costly, labor intensive, or use murine cells, this system quickly models the development of human cancers expressing human epitopes treated with human cell therapies.

Methods HER2+ (FaDu and SCC-47) and HER2- (MDA-MB-468) cell lines with luciferase reporters were engrafted onto the CAM. Tumor growth was monitored by IVIS imaging of luciferase activity of viable tumor cells. Engrafted cells grew for three days during which tumors establish and incorporate into the CAM. Established tumors were treated with second generation HER2-directed CAR T with a CD28 costimulatory endodomain or an expanded T cell control which was not transduced with CARs. Viability of tumor cells was monitored through quantification of luciferase activity from engrafted cells before and following four days of CAR-T treatment. Histology was used to visualize tumor structure on the CAM and CAR-T infiltration of tumor tissue. Staining was scored by percent positive area and intensity using QuPath software.

Results Luciferase activity of viable tumor cells was significantly reduced in CAR-T treated FaDu tumors (p<0.0001). Following tumor harvest, we observed a significant reduction in Ki67 staining for CAR-T treated tumors relative to T cell treatment alone for both SCC-47 and FaDu tumors (p=0.0121 and p=0.0176 respectively). We also observed CAR-T persistence in tissue and infiltration of tumors.

Conclusions The CAM system supports the growth of solid tumors which can be targeted by immune cell infiltrates. These data support continued development of the chick CAM model as a candidate in vivo system for rapid, scalable screening of CAR T efficacy against human solid tumors.