Abstract

The field of cancer immunotherapy is fast evolving and is now applied for the treatment of several malignancies. However, immune checkpoint inhibitors and adoptive T cell therapies have shown minimal efficacy against breast cancer, highlighting the need for more effective therapies. Here, we assessed the potential of αβT cells engineered to express a defined γδT cell receptor (TEGs), which sense tumor-specific metabolic changes, to target breast tumors. By co-culturing TEGs and human breast organoids we showed that TEGs can kill tumor- or metastasis-derived organoids, but not organoids derived from healthy tissue. Recognition by TEGs was dependent on the breast tumor donor (with 8 out of 14 donors targeted) but did not relate to the breast cancer subtype or mutational load. We next developed a 3D live-cell imaging platform that allows direct analysis of T-cell-mediated killing dynamics over time without the need for genetic manipulations. Based on various behavioral TEG parameters (cell death, speed, displacement and cell-to-cell interactions), we performed a multivariate time-series clustering analysis and identified a population that was enriched for TEGs cultured with organoids that are susceptible for killing. This cluster showed intermediate movement and was the only population able to establish long-term interactions with organoids. In contrast, a static and dying population was enriched for TEGs that were cultured with organoids resistant to killing. In depth analysis of the long-term interacting cluster indicated that the minority of cells had limited killing capacity, while most cells were able to perform serial killing, further demonstrating high behavioral heterogeneity within the T cell population. These studies open up new therapeutic avenues for breast cancer and opportunities to improve T cell-based immunotherapies.