Evaluating antitumoral lymphocytes for adoptive cell therapy : oxidative stress, neoantigen processing and co-stimulation

Abstract: Rapid advancement in the field of immunotherapy has changed the treatment landscape of cancer, especially with the approval of immune checkpoint inhibitors for a wide range of cancer types. However, a large proportion of patients does not benefit from this treatment. Adoptive cell therapy (ACT) describes the infusion of immune cells, such as large numbers of tumor-reactive cytotoxic lymphocytes that can control and eliminate the tumor. Remarkable clinical responses have been achieved with the infusion of CD19- targeting CAR T cells in hematological cancers or tumor infiltrating lymphocytes (TIL) in melanoma. However, limitations in identifying suitable antigens for CAR T cells and the immunosuppressive features of the tumor microenvironment contribute to insufficient responses to ACT in most solid tumors. The goal of the studies in this thesis was to evaluate and optimize the antitumoral efficacy of cytotoxic lymphocytes to elicit their full potential and thereby improve cell products used for ACT. Reactive oxygen species (ROS) are radicals that impair the effector functions of cytotoxic lymphocytes in the tumor microenvironment. In paper I, we hypothesized that activating the Nrf2 antioxidant pathway in these cells could overcome this limitation. Auranofin and two other compounds induced Nrf2 target gene expression, leading to reduced intracellular ROS accumulation and improved effector functions during oxidative stress in NK cells, CAR T cells, and TIL. We propose that auranofin treatment as a last step of clinical expansion protocols could improve antitumoral activity of the infused lymphocytes and thus clinical outcomes. In paper II, we aimed to understand the impact of culturing cancer cells as spheroids, on tumor cell immunogenicity. Spheroids resemble the tumor microenvironment and are therefore an advantageous cell culture model to study immune-cancer cell interactions. We found that TIL, either unsorted or neoantigen-specific, recognized autologous tumor cells cultured as spheroids more efficiently than those cultured in a monolayer. This was dependent on TCR-MHC class I interaction and immunoproteasome activity, and thus due to altered antigen processing and presentation. In Paper III, we studied the functional and phenotypic characteristics of T cells expressing mesothelin directed, second generation CARs with either CD28 (M28z CAR) or 4-1BB (MBBz CAR) co-stimulatory domain. We found that M28z, compared to MBBz, CAR T cells showed higher infiltration and killing of ovarian cancer cell spheroids with varying mesothelin expression. Further, we showed that trogocytosis of mesothelin leads to fratricide killing and antigen loss and therefore presents a limitation for CAR T cell therapy. These findings contribute to a better understanding of the influence of immunosuppressive mediators, co-stimulation, and antigen processing on the antitumoral functions of cytotoxic lymphocytes and can thereby support the development of new and improved cell therapies.