Taken together, we conclude that sorting T cells based solely on m identifies cell populations with distinct gene and metabolic profiles associated with either effector T cell (high-m) or long-lived memory (low-m) T cell subsets

Taken together, we conclude that sorting T cells based solely on m identifies cell populations with distinct gene and metabolic profiles associated with either effector T cell (high-m) or long-lived memory (low-m) T cell subsets. Low-m defines stem-cell-like activity in an array of cell types Given that m-based sorting allowed for the enrichment of CD8+ T cells with distinct metabolic and functional activities, we hypothesized that this strategy might have a similar Rapamycin (Sirolimus) discriminatory power in a variety of other cell types, including T cell Rapamycin (Sirolimus) subsets and hematopoietic stem cell (HSC). of viral-associated illnesses and cancer. Graphical Abstract INTRODUCTION Immunotherapy using adoptive transfer of tumor-specific T cells mediates durable and complete disease regression Rapamycin (Sirolimus) in some patients with metastatic cancer (Brentjens et al., 2013; June et al., 2015; Porter et al., 2011; Riddell and Greenberg, 1995). Mounting evidence has shown that metabolism supports and drives many basic features of T cells including cellular activation, proliferation, differentiation, effector function (Gerriets et al., 2014; Gerriets and Rathmell, 2012; MacIver et al., 2013; Michalek et al., 2011a; Michalek et al., 2011b; Pearce et Hhex al., 2013; Pearce et al., 2009; Sena et al., 2013; Shi et al., 2011), and anti-tumor immunity. This has led to a growing interest in leveraging this understanding to improve the efficacy of T cell transfer therapies, such as adoptive transfer immunotherapy in the treatment of cancer. In pre-clincial models it has been shown that highly-glycolytic T cells are short-lived after adoptive transfer and have impaired anti-tumor immunity (Sukumar et al., 2013), whereas T cells with a metabolic profile characterized by elevated fatty-acid oxidation (Pearce et al., 2009) and enhanced mitochondrial spare respiratory capacity (SRC) have greater long-term survival (van der Windt et al., 2012). Although there is increasing evidence that metabolism can affect the survival and anti-tumor function of T cells, identifying a simple and clinically-feasible method to isolate T cells with favorable metabolic features has proved challenging. Because mitochondria are the central metabolic organelle in cells, we hypothesized that the measurement of a single mitochondrial-associated parameter may help to identify T-cells with a favorable bioenergetic profile that can survive for long periods after adoptive transfer for T-cell based immunotherapy. Here, we describe a clinically-feasible method to isolate functionally-robust T cells based on a single metabolic parameter: mitochondrial membrane potential (m). Mitochondria produce energy by establishing an electrochemical proton motive force (p) across their inner cell membrane, which in turn fuels the synthesis of ATP by driving the proton turbine F0F1 ATPase (Ehrenberg et al., 1988; Sena et al., 2013; Wang and Green, 2012; Weinberg et al., 2015). We show that CD8+ T cells that are found to have low-m display enhanced persistence, augmented autoimmunity and greater anti-tumor immunity relative to high-m cells. These findings demonstrate that metabolic-sorting can complement sorting based on conventional cell surface markers in identifying cells with the capacity for long-term survival and ongoing Rapamycin (Sirolimus) effector function after adoptive-transfer. This novel immunometabolomic approach to cell sorting may have important and immediate therapeutic applications in enhancing cell-based therapies for patients with viral-associated illness, advanced cancer, and disorders of hematopoiesis. RESULTS AND DISCUSSION m based sorting segregates short-lived effector from memory T cell precursors To understand the molecular programs regulating long-term persistence and anti-tumor functions of the CD62L+ memory T cell population, we compared a genome-wide microarray analysis of minimally-differentiated stem-cell memory T cells (TSCM, CD62L+ CD44- Sca-1+) with more highly-differentiated effector memory T cells (TEM, CD62L? CD44+) and found significant differences in the expression of genes related to metabolic processes (Figure 1A and Table S1). We then FACS-purified T cells using the mitochondrial potential-sensitive dye TMRM, a lipophilic cationic dye that accumulates in the mitochondrial matrix in proportion to the magnitude of m electronegativity (Ehrenberg et al., 1988). We vaccinated pmel-1 T cell receptor (TCR) transgenic mice, whose CD8+ T cells recognize an epitope derived from Rapamycin (Sirolimus) the shared melanocyte/melanoma differentiation antigen (Ag) gp100, with a recombinant vaccinia virus encoding hgp100 (gp100-VV). At the peak of the primary immune response following vaccination, we FACS-sorted the bulk population of T cells into low-m and high-m fractions and subsequently transferred equal numbers of cells into either wild-type (WT) or (recombination activating gene-2) Rag2?/? recipient mice, which were then infected with gp100-VV (Figure 1B). Cells derived from the low-m cell fraction were enriched in TSCM and central memory.