Clinical adoptive cellular immunotherapy of malignancy and viral infection should transfer T cells that expand in vivo on exposure to antigen and can enter the memory compartment to persist long-term. A number of factors, including cellular phenotype, influence the behavior of the infused line. Primate studies have shown that antigen-specific CD8+ T cell clones only persisted long-term in vivo if they were derived from central memory T cells, but not from effector memory T cells, reacquiring phenotypic and functional properties of memory T cells.1 Other studies have suggested that adoptive transfer of ex vivo-expanded effector memory T cells will have poor survival and clinical efficacy, reporting instead that less differentiated T cells with longer telomeres exhibit longer persistence. These data imply that prolonged ex vivo expansion, required, for example, for T cell cloning, adversely affects subsequent in vivo expansion and survival. However, our trials administering ex vivo-expanded, polyclonal EBV-specific T cell lines demonstrated that expanded effector memory T cells, infused into a lymphodepleted host, can expand massively in vivo, enter the memory compartment, and persist for up to seven years after infusion. Furthermore, in a study infusing trivirus-specific CTLs with effector memory phenotype, we saw expansion of CTLs specific for the latent viruses CMV and EBV. By contrast, adenoviral-specific CTL persisted only in patients who were acutely infected with the agent2 We recently compared non-specifically activated T cells (ATC) with EBV-specific CTLs derived from the same initial peripheral blood collection and expressing distinguishable chimeric GD2-specific chimeric antigen receptors (CARATC and CAR-CTL). In this study, ATCs were cultured for 14 to 21 days. Between 0.9% and 6.1% retained a central memory (CCR7+, CD62L+) phenotype, up to 30% had an effector memory phenotype (CCR7−, CD62L+), and the remainder had a terminally/fully differentiated effector phenotype. By contrast, EBV-CTL were cultured for 30 to 44 days and expressed no CCR7, but up to 50% were CD62L+, and contained cells that were terminally/fully differentiated effectors and effector memory cells. These EBV-CTLs also all had a CD45RO memory phenotype, while about 13% to 60% of ATCs expressed CD45RA, a marker of naïve T cells. Despite these differences in memory subsets, it was the CAR-CTLs that had the clearly greater persistence and could be shown to retain functionality, while CAR-ATC rapidly disappeared from the circulation and could not be recovered. Hence, factors other than phenotype, such as antigenic stimulation and costimulation almost certainly influence cell fate after infusion, and determine whether or not effector memory cells can re-access the central memory pool. Ultimately, strategies that combine selection of optimal phenotype with the provision of antigen stimulation and co-stimulation and a cytokine milieu that favors homeostatic expansion will likely lead to the most effective outcomes following adoptive T cell transfer.
Lineage relationships, homeostasis, and recall capacities of central– and effector–memory CD8 T cells in vivo