Selecting T-Cell Subsets for Adoptive T-Cell Therapy to Optimize Potency and Persistence

Adoptive T-cell therapy with tumor-reactive T cells is emerging as a highly effective strategy for eliminating even the most advanced chemotherapy refractory malignancies. Endogenous T cells specific for tumor-associated antigens can sometimes be isolated and expanded from the patient’s blood or tumor infiltrate, or more expeditiously can be engineered by gene transfer to express a T-cell receptor specific for a tumor associated MHC/peptide complex or a synthetic chimeric antigen receptor (CAR) specific for a tumor associated cell surface molecule. The remarkable regression of advanced acute lymphocytic leukemia and lymphoma in patients treated with T cells engineered to express CD19-specific CARs illustrates the potential for this approach to transform clinical care. Therapeutic activity is variable in individual patients, however, and this appears to correlate with the ability of transferred, tumor-reactive T cells to persist and proliferate in vivo, and to retain effector function. These attributes may reflect both the qualities of the T cells that are isolated or engineered for therapy, and the local tumor microenvironment that may contain regulatory T cells; cells that express ligands that engage inhibitor receptors on effector T cells or cytokines that inhibit effector T-cell proliferation. The CD4⁺ and CD8⁺ T cell pools in normal individuals contain a variety of naïve, memory, and regulatory T-cell subsets that differ in epigenetic, transcriptional, and functional properties. Because most clinical protocols have used polyclonal peripheral blood mononuclear cells as recipients for CAR gene transfer, the composition of T-cell products that are being administered is highly variable, particularly when the T cells are obtained from cancer patients that have received prior cytotoxic chemotherapy that can skew the phenotypic composition of the peripheral T-cell pool. As a consequence, transferring tumor-targeting receptors into polyclonal unselected cell populations provides poor control over the cellular composition of the final T-cell product, which may in part explain the marked differences in efficacy and toxicity that have been observed in the clinic, and may complicate regulatory approval of these novel therapies. Methods to derive T cells from distinct naïve and memory T-cell subsets have been developed, enabling the rapid production of therapeutic T cells of uniform composition. The results of preclinical studies that illustrate the improved potency of defined T-cell products that are engineered with tumor-specific CARs, and the clinical implementation of this approach in B-cell malignancies will be presented.