The stem of T cell memory

Credit: CORBIS

Effector memory T (T_EM) cells are terminally differentiated and acquire effector function immediately after re-stimulation, whereas central memory T (T_CM) cells have a longer lifespan and can differentiate into T_EM cells following antigenic challenge. A recent study published in Nature Medicine reports that an additional memory T cell subset, T memory stem (T_SCM) cells, which was previously identified in mice, exists in humans and exhibits stem cell properties.

these self-renewing, multipotent human T_SCM cells have an improved therapeutic potential compared with T_CM and T_EM cells

Mouse T_SCM cells have a naive-like CD44^lowCD62L^hi phenotype, express the CD122 (a component of the IL-15 receptor), stem cell antigen 1 (SCA1), B cell lymphoma 2 (BCL-2) and CXC-chemokine receptor 3 (CXCR3), and can be induced through activation of the WNT signalling pathway. In search of a human memory T cell subset with analogous properties, Gattinoni et al. stimulated ex vivo-isolated human naive T cells in the presence of the WNT pathway activator TWS119 and observed upregulation of CD95 (also known as FAS) and the memory-associated marker CD122 in the treated T cells, which retained a naive-like phenotype in other respects. Thus, they hypothesized that CD95 may be a marker for human T_SCM cells.

Indeed, flow cytometric analysis of human peripheral blood identified naive-like CD4⁺ and CD8⁺ T cells with the CD95⁺CD122⁺ phenotype and high levels of expression of BCL-2 and CXCR3. Like T_CM and T_EM cells, human T_SCM cells were found to have undergone clonal expansion and to be long-lived, and they rapidly acquired effector function following T cell receptor (TCR) stimulation. The gene expression pattern of T_SCM cells indicated that they are a distinct population of memory T cells that are less differentiated than T_CM and T_EM cells.

Interestingly, human T_SCM cells proliferated robustly in response to interleukin-15 stimulation (which is an attribute of memory but not naive T cells), but the majority of proliferating T_SCM cells maintained their initial naive-like phenotype, which suggests that these cells have a self-renewal capacity. By contrast, following TCR stimulation, most T_SCM cells gradually acquired a T_CM phenotype, some of them differentiated into T_EM cells and approximately 15% of them maintained their initial phenotype, indicating that T_SCM cells are multipotent.

So, do these self-renewing, multipotent human T_SCM cells have an improved therapeutic potential compared with T_CM and T_EM cells? Similarly to their mouse counterparts, human T_SCM cells proliferated more vigorously than T_CM and T_EM cells following TCR stimulation in vitro. Moreover, human T_SCM cells were found to survive longer than T_CM and T_EM cells after adoptive transfer into immunodeficient mice.

Finally, a comparison of the therapeutic potential of in vitro-generated human T_EM, T_CM and T_SCM cells (all of which were engineered to express a chimeric TCR specific for a tumour antigen) indicated that T_SCM cells — which were induced by pharmacological activation of the WNT pathway — were the most effective T cells in terms of survival and tumour clearance after transfer into tumour-bearing mice. This potential to generate engineered human T_SCM cells in vitro opens new avenues for cancer immunotherapy.