Low-ΔΨm cells display a gene expression and metabolic profile of memory CD8⁺ T cells

Using RNA-seq analysis to provide a global assessment of differential gene expression between ΔΨm subsets, we identified a total of 1895 genes that were differentially-expressed (1151 up-regulated and 744 down-regulated) in the low-ΔΨm compared to high-ΔΨm fractions (Figure 2A and Table S2). Low-ΔΨm cells were specifically enriched for the expression of genes encoding factors known to promote memory CD8⁺ T cell formation, including B cell CLL/lymphoma 6 (Bcl6), the Wnt-β-catenin signaling transducers transcription factor 7 (Tcf7) and lymphoid enhancer factor (Lef1), and the FOXO-target gene Kruppel-like factor 2, lung (Klf2) (Figure 2A). Quantitative PCR analyses confirmed these differences in gene expression (Figure 2B). In contrast, transcripts encoding key regulators of effector differentiation such as PR domain zinc finger protein 1 (Prdm1) and eomesodermin (Eomes) were enriched in high-ΔΨm cells. Furthermore, genes that encode for cytotoxic effector molecules such as perforin (Prf1) and granzyme B (Gzmb) were also enriched within the high-ΔΨm subset (Figure 2B). We also found that multiple distinct inhibitory T cell receptors such as programmed death-1 (PD-1), cytotoxic T-lymphocyte-associated protein 4 (CTLA-4), B- and T-lymphocyte attenuator (BTLA) and lymphocyte-activation gene 3 (LAG-3) were enriched in the high-ΔΨm subset compared with the low-ΔΨm subset (Figure 2A).

Open in a separate window

Figure 2

Low-ΔΨm cells display a gene expression and metabolic profile of memory CD8⁺ T cells

(A) Volcano plot of RNA-sequencing analysis of ΔΨm-sorted CD8⁺ T cell subsets. Analysis includes a total of 1895 genes that were either up-regulated (1151 genes) or down-regulated (744) by at least a 2-fold change. Highlighted are canonical genes associated with glycolysis, fatty-acid oxidation (FAO), effector, memory differentiation and co-inhibitory receptors (B) Bar graphs showing quantitative RT-PCR expression of indicated memory and effector genes in CD8⁺ T cells. (C) Global metabolomic heatmap showing metabolites detected in CD8⁺ T cells within ΔΨm subsets. (D) Scatter plot showing quantification of key metabolites (FAO) in CD8⁺ T cells from the indicated ΔΨm subset. (E) Bar graphs showing quantitative RT-PCR analysis for expression of Cpt1a, Acaca, Fasn and Slc2a1 within ΔΨm subsets. (F) ECAR of low-ΔΨm and high-ΔΨm CD8⁺ T cells. (G) OCR of low-ΔΨm and high-ΔΨm CD8⁺ T cells in response to indicated mitochondrial modulators: oligomycin; FCCP, R&A-rotenone and antimycin A. Data are presented as mean ± SEM and * = P < 0.05; ** = P < 0.01; *** = P < 0.001; **** = P < 0.0001 (two-tailed t test).

Recent studies have demonstrated that metabolic programs may in part determine memory and effector CD8⁺ T cell fate (MacIver et al., 2013; Pearce et al., 2009; Sukumar et al., 2013). To elucidate the metabolic program associated with low-ΔΨm CD8⁺ T cells, we performed an unbiased metabolomic profiling of sorted ΔΨm subsets. This metabolic fingerprinting revealed that ΔΨm CD8⁺ subsets had differences in 39 distinct metabolic intermediates (Figure 2C). Relative levels of ribose-5-phosphate, sedoheptulose-5-phosphate and ribulose-5-phosphate were measured in the low-ΔΨm and high-ΔΨm CD8⁺ T cell subsets using LC/MS, but only ribulose-5-phosphate levels were significantly different (Figure S2A). Notably, multiple carnitine species were more abundant in low-ΔΨm cells compared to their high-ΔΨm counterparts (Figure 2D). In addition, we measured increased levels of short, medium, and long-chain free fatty-acid (FFA) metabolites in the low-ΔΨm subset (Figure 2D).

Given the importance of the balance between fatty acid oxidation (FAO) and fatty acid synthesis (FAS) on T cell differentiation (Berod et al., 2014), we next evaluated whether low-ΔΨm cells display differences in the expression of key enzymes involved in FAS and FAO. Low-ΔΨm cells displayed increased expression of carnitine palmitoyl transferase 1a (Cpt1a), the rate-limiting enzyme involved in transport of fatty-acids into the mitochondria for subsequent β-oxidation (van der Windt et al., 2012) (Figure 2E). Additionally, the expression of both acetyl-coA-carboxylase (Acaca) and fatty acid synthase (Fasn) were reduced in the low-ΔΨm subset. These data indicate that low-ΔΨm cells may preferentially engage in FAO as opposed to FAS (Figure 2E). RNA-Seq analysis revealed that gene expression of many of the enzymes involved in the glucose metabolism such as hexokinase 2 (Hk2), lactate dehydrogenase (Ldha), pyruvate dehydrogenase (Pdh), and pyruvate dehydrogenase kinase, isoform 2 (Pdk2) are differentially-expressed between the low-ΔΨm and high-ΔΨm CD8⁺ T cell groups (Table S2). Quantitative PCR analysis confirmed that low-ΔΨm cells displayed reduced expression of both Hk2 and Ldha, (Figure S2B). Interestingly, RNA-seq analysis revealed that Pdk1 expression was not significantly altered in ΔΨm subsets and quantitative PCR analysis revealed that the expression of Pdk2 was also not significantly increased in low-ΔΨm CD8⁺ T cells (Figure S2B). The high-ΔΨm subset expressed greater levels of the glucose transporter solute carrier family 2, facilitated transporter member 1 (Slc2a1) (Figure 2E). To determine whether these gene expression and metabolic measurements were associated with functional changes in cellular metabolism, we next evaluated the extracellular acidification rate (ECAR), which quantifies proton production as a surrogate for lactate production and thus reflects glycolytic flux and the oxygen consumption rate (OCR), a measure of overall mitochondrial respiration. Consistent with gene expression data, high-ΔΨm cells demonstrated evidence of increased glycolytic flux (Figure 2F). These cells also exhibited greater basal OCR (Figure S2C). Low-ΔΨm cells exhibited greater spare respiratory capacity, a feature of long-lived memory CD8⁺ T cells (van der Windt et al., 2012) (Figure 2G). Thus, low-ΔΨm cells exhibit many of the metabolic hallmarks associated with memory T cells, including increased expression of Cpt1a and increased levels of free fatty acids, increased spare respiratory capacity, and relatively low glycolytic flux. Taken together, we conclude that sorting T cells based solely on ΔΨm identifies cell populations with distinct gene and metabolic profiles associated with either effector T cell (high-ΔΨm) or long-lived memory (low-ΔΨm) T cell subsets.

Low-ΔΨm defines stem-cell-like activity in an array of cell types

Given that ΔΨm-based sorting allowed for the enrichment of CD8⁺ T cells with distinct metabolic and in vivo functional activities, we hypothesized that this strategy might have a similar discriminatory power in a variety of other cell types, including T cell subsets and hematopoietic stem cell (HSC). A recent report demonstrated that adult tissue memory T cells with stem cell-like properties reside within the CD62L⁺ T_CM population (Graef et al., 2014). Therefore, using a gating strategy based on T_CM surface markers (CD8⁺CD44⁺CD62L⁺), we further partitioned the T_CM subset into low-ΔΨm and high-ΔΨm T_CM populations (Figure S3A and Figure S3B). We found that low-ΔΨm T_CM display increased mRNA levels of transcription factors and genes associated with memory precursors (Figure S3C). Following adoptive transfer, low-ΔΨm T_CM demonstrated a 6-fold greater engraftment in the spleen of vaccinated mice compared with their high-ΔΨm counterparts (Figure 3A). Thus, we were able to isolate a functionally-distinct fraction of stem-cell-like T cells from an otherwise phenotypically homogenous T_CM population. We next addressed whether ΔΨm sorting would allow us to isolate metabolically-fit T cells within distinct effector T cell populations (Tc17, Th1 and Th17 cells). We generated Tc17 and segregated them based on ΔΨm by FACS sorting (Figure S3D). Upon adoptive transfer, we found that low-ΔΨm Tc17 cells displayed enhanced survival capacity compared with their high-ΔΨm counterparts (Figure 3B). Similarly, using CD4⁺ cells specific for the tissue differentiation antigen TRP-1 (Muranski et al., 2011), we polarized T cells under Th1 and Th17 conditions and segregated them into low-ΔΨm and high-ΔΨm (Th1 and Th17) subsets. Again expression of factors known to promote longevity, including Bcl6, Tcf7 and Lef1, were enriched in the low-ΔΨm fraction relative to the high-ΔΨm fraction within these effector CD4⁺ subsets (Figure 3C and 3D).

Open in a separate window

Figure 3

Stem-like characteristics of low-ΔΨm cells are conserved within other cell types

(A–B) Flow cytometic analysis after adoptive-transfer of 10⁵ high-ΔΨm or low-ΔΨm CD8⁺ T cells isolated from spleen of B6 mice. (C–D) Bar graphs showing expression of quantitative RT-PCR analysis of indicated genes in Th1 (C) and Th17 cells (D) CD4⁺ T cells in indicated ΔΨm subsets. (E) Representative contribution of sorted donor mouse HSCs (CD45.1) mixed with competitor bone marrow (CD45.2) at one month (F) and over a 6-month period. (n=5 mice per group). Data are presented as mean ± SEM and ** = P < 0.01; *** = P < 0.001; **** = P < 0.0001 (two-tailed t test).

There is a growing appreciation that intracellular metabolism may be critical for the maintanence of HSC and induced pluripotent stem cells (iPSCs) (Folmes et al., 2011; Ito et al., 2012; Simsek et al., 2010; Takubo et al., 2010), embryonic stem cells (Schieke et al., 2008) and cancer stem cells (CSCs) (Ye et al., 2011). To determine whether this metabolic sorting has applicability in stem cells, we analyzed long-term hematopoietic stem cells (LT-HSC) purified from mouse bone marrow based on conventional surface markers (Lin⁻/Sca1⁺/c-kit⁺/CD34⁻). We isolated mouse LT-HSC and characterized the phenotype of the corresponding low-ΔΨm and high-ΔΨm subsets (Figure S4A). Several reports have demonstrated that CD150 high HSC are associated with greater self-renewal and long-term reconstitution potential (Morita et al., 2010). Therefore, we evaluated the expression of CD150 expression in the low-ΔΨm and high-ΔΨm HSCs. We found that CD150, a LT-HSC-self renewal marker was expressed by 80% of the low-ΔΨm HSC population, while high-ΔΨm HSCs did not express any levels of this marker (Figure S4B and Figure S4C). We next purified low-ΔΨm and high-ΔΨm subsets from within mouse Lin⁻/Sca1⁺/c-kit⁺/CD34⁻ cells and mixed these with competitor bone marrow prior to reinfusion into lethally-irradiated recipient mice. As measured by donor reconstitution, long-term engraftment capacity of low-ΔΨm LT-HSCs was significantly more robust than high-ΔΨm LT-HSCs (Figure 3E and 3F). Our results support the view that sorting HSC based on ΔΨm can identify long-term self-renewing HSC for transplant capable of effective long-term repopulation.

High-ΔΨm is associated with effector cytokine production in T cells

We sought to determine whether ΔΨm can modulate cytokine production in effector T cell subsets. To test this, we generated effector T cell subsets such as Tc17, Th1 and Th17, further sorted these subsets into low-ΔΨm and high-ΔΨm subsets and measured cytokine production within each lineage. High-ΔΨm Tc17 secreted 3-fold higher levels of IL-17A compared with the low-ΔΨm Tc17 subset (Figure 4A and 4B). Furthermore, under Th1 polarizing conditions high-ΔΨm Th1 CD4⁺ T cells demonstrated a four-fold increase in levels of IFN-γ compared with the low-ΔΨm Th1 subset (Figure 4C and 4D). Finally, the majority of high-ΔΨm Th17 cells expressed increased levels of both IL-17A and IL-17F compared to their low-ΔΨm counterparts (Figure 4E and 4F). These results suggest that an increase in ΔΨm is associated with increased effector cytokine secretion in a variety of lymphocyte lineages. Together, these data show that in addition to inducing aerobic glycolysis, lymphocytes also may need to increase their ΔΨm to achieve full effector cytokine production and function.

Open in a separate window

Figure 4

High-ΔΨm is associated with effector cytokine production in T cells

(A) Representative intracellular staining and (B) quantification of IL-17A by Tc17 polarized Pmel-1 CD8⁺ cells generated in vitro after 4 hr restimulation with phorbol myrystate acetate (PMA) and ionomycin in the presence of brefeldin A. (C) Representative intracellular staining and (D) quantification of IFN-γ by Th1 TRP-1 CD4⁺ cells generated in vitro (E) Representative intracellular staining of IL-17A and IL-17F and (F) quantification of IL-17A by Th17-polarized TRP-1 CD4⁺ cells generated in vitro. Data are presented as mean ± SEM and ** = P < 0.01; **** = P < 0.0001 (two-tailed t test).

Low-ΔΨm CD8⁺ T cells demonstrate decreased oxidative-stress

The functional demands and longevity of stem cell-like memory T cells likely require programs to reduce oxidative stress and maintain long-term genomic stability. We therefore tested if the increased functionality observed in low-ΔΨm CD8⁺ T cells might reflect an intrinsic reduction in oxidative stress levels. Direct reduction of ΔΨm in CD8⁺ T cells resulted in reduced ROS levels as measured by the redox-sensitive fluorophore dihydrodichlorofluorescein diacetate (DCFDA) (Figure S4D). Low-ΔΨm CD8⁺ T cells had a large reserve pool of oxidized glutathione (GSSG) compared with high-ΔΨm CD8⁺ T cells (Figure S4E). Furthermore, the low-ΔΨm CD8⁺ T cells also maintained increased mRNA expression of ROS-detoxifying enzymes including catalase (Cat), glutathione peroxidase 4 (Gpx4), superoxide dismutase 1 (Sod1) and superoxide dismutase 2 (Sod2) (Figure S4F) that may contribute to low-ΔΨm CD8⁺ T cell redox regulation. The increased levels of glutathione and anti-oxidant genes may help to support the in vivo expansion and redox balance of low-ΔΨm CD8⁺ T cells. These cells also had reduced levels of DNA-damage, as assessed by Ser139 phosphorylated histone varant H2A.X (γH2AX) foci formation on nuclear DNA (Figure 5A). In contrast, high-ΔΨm T cells exhibited increased γH2AX staining, and demonstrated relative up-regulation of genes involved in DNA replication, DNA repair (Figure 5B) and cell-cycle arrest (Figure 5C).

Open in a separate window

Figure 5

Low-ΔΨm CD8⁺ T cells demonstrate decreased oxidative stress, increased long-term in vivo persistence and superior anti-tumor activity

(A) DNA damage in DAPI stained (blue) ΔΨm-sorted CD8⁺ T cells using indirect immunofluorescence staining (green) for γ-H2A histone family member X (γ-H2AX). Bottom right shows % of cells with H2AX foci per high powered field. (B–C) Quantitative RT-PCR analysis of genes associated with (B) DNA repair and (C) Cell cycle inhibition in CD8⁺ T cells. (D) Flow cytometric analysis for reactive oxygen species (ROS) levels using 2’,7’–dichlorofluorescein diacetate (DCFDA) fluorescence expressed in arbitrary units (AU) ΔΨm-sorted subsets. (E) Flow cytometric analysis of DNA damage using phospho-Histone H2A.X (Ser139) antibody in ΔΨm-sorted subsets. (F) Quantitative RT-PCR expression levels of indicated exhaustion factors and (G) Flow cytometric analysis of exhaustion markers in ΔΨm-sorted bulk CD8⁺ T cells. Numbers in the panel G indicate MFI (H) Quantitative RT-PCR expression levels of indicated pro-apoptotic gene (Bax) within ΔΨm-sorted bulk CD8⁺ T cells (I) Representative flow cytometry analysis of Annexin and PI staining in ΔΨm-sorted bulk CD8⁺ T cells. Data are presented as mean ± SEM and * = P < 0.05; ** = P < 0.01; *** = P < 0.001; **** = P < 0.0001 (two-tailed t test).

To test if the hypothesis that ROS generation may compromise function of self-renewal and longevity programs is generalizable to other T cell subsets and to HSC, we measured whether increased ROS and DNA damage levels were present within high-ΔΨm subsets in varieties of other cell types. Consistent with findings using bulk activated CD8⁺ T cells, we found that the high-ΔΨm subset of the central memory CD8⁺ T cell subset (T_CM), effector T cell subsets (Tc17, Th1 and Th17 cells) and LT-HSC exhibited elevated ROS levels (Figure 5D) and DNA damage (Figure 5E) compared to their low-ΔΨm counterparts. Elevated levels of oxidative stress and damaged DNA resulting from increased ROS in stem cells have been linked with impaired stem cell self-renewal (Suda et al., 2011). Our results are consistent with a model where high continuous and unregulated levels of ROS may limit the long-term survival of lymphocytes and hematopoietic stem cells.

In addition to direct detrimental effects of high ROS and DNA damage levels, recent data have demonstrated that chronic infection and cancer can trigger terminal differentiation and T cell exhaustion (Wherry, 2011). Therefore we evaluated the expression of various exhaustion markers associated with terminal differentiation in metabolically-sorted cells. We found that multiple negative co-inhibitory receptors such as PD-1, CTLA-4, LAG-3 and BTLA that regulate T cell exhaustion were enriched at both the RNA (Figure 5F) and protein level (Figure 5G) in the high-ΔΨm T cells compared with the low-ΔΨm T cells. The expression of multiple negative inhibitory receptors in high-ΔΨm T cells may contribute to the observed reduced long-term survival. Additionally high-ΔΨm CD8⁺ T cells underwent increased cell death compared with low-ΔΨm CD8⁺ T cells as evidenced by increased expression of the pro-apoptotic gene Bax (Figure 5H) and increased Annexin-V staining (Figure 5I). These results demonstrate that high-ΔΨm cells display increased oxidative stress, DNA damage response, cell-cycle arrest and molecular features of T-cell exhaustion, each of which predispose T cells to apoptotic death. In contrast, we demonstrate that low-ΔΨm cells are more efficiently suited to combat oxidative stress, are protected from DNA damage, and are resistant to apoptosis.

Interleukin-2 augments ΔΨm through mTOR signaling

The presence of an inflammatory cytokines (Chang et al., 2014; Finlay et al., 2012; Kaech and Cui, 2012) and increased mTOR activity (Araki et al., 2009; Chi, 2012; Kaech and Cui, 2012; Shi et al., 2011; Shrestha et al., 2014) are both required for the full differentiation of functional effector T cells. Whether these cytokines regulate ΔΨm, however, remains unknown. The presence of the cytokine interleukin-2 (IL-2) during T cell priming increased the basal OCR and ΔΨm in a concentration-dependent fashion (Figure S5A and S5B). We found that low-ΔΨm cells displayed reduced activity of mTORC1 signaling as evidenced by reduced level of phosphorylation of 4-E-BP1 and S6 phosphorylation compared to their high-ΔΨm counterparts (Figure S5C). We next tested whether the reduction of mTORC1 signaling using the mTOR inhibitor, rapamycin, would reduce ΔΨm. T cell priming in the presence of rapamycin significantly reduced ΔΨm in CD8⁺ T cells (Figure S5D). Our results suggest that the presence of IL-2 and increase in mTORC1 signaling may increase ΔΨm to regulate the effector T cell response.

T cell differentiation is characterized by an increase in mitochondrial metabolism (Sena et al., 2013). Our data suggest that increase in ΔΨm and ROS generation are dependent on the IL2-mTOR axis. Our assertion that differences in ΔΨm may regulate cell-fate decisions was further supported by recent observations that increased mitochondrial metabolism and ROS are important for adipocyte differentiation (Tormos et al., 2011), T cell differentiation (Sena et al., 2013; Shrestha et al., 2014; Weinberg et al., 2015) and differentiation of embryonic stem cells (Schieke et al., 2008).

Low-ΔΨm CD8⁺ T cells demonstrate long-term in vivo persistence and superior antitumor-activity

The metabolic, biochemical, and transcriptional profiles of low-ΔΨm T cells suggest that these cells might have significant advantages in long-term functional reconstitution assays. Long-term persistence is one of the key biological characteristics of memory T cells (Chang et al., 2014; Graef et al., 2014; Kaech and Cui, 2012). We therefore sought to evaluate long-term survival following adoptive transfer of an equal number of low-ΔΨm and high-ΔΨm CD8⁺ T cells. When assessed 300 days after adoptive transfer, low-ΔΨm CD8⁺ cells displayed a dramatic increase in persistence compared with high-ΔΨm cells, as determined by the frequency and number of pmel-1 T cells in the spleen (Figure 6A and 6B). Long-term persistence of pmel-1 CD8⁺ T cells was associated with expected on-target autoimmune sequelae through sustained damage to melanocytes. This melanocyte targeting led to the development of vitiligo, a condition that was significantly greater in recipient mice receiving low-ΔΨm cells (Figure 6C). We have also previously demonstrated that there is a strong correlation between the functionality of CD8⁺ T cells targeting gp100 and the degree of ocular autoimmunity (Palmer et al., 2008). Therefore, we analyzed the extent of ocular injury in recipient animals 300 days after transfer. We noted that animals receiving low-ΔΨm CD8⁺ T cells exhibited more severe uveitis when compared to the recipients of high-ΔΨm cells (Figure 6D). The ability of T cells to robustly proliferate and persist after transfer has been shown to be critical to anti-tumor responses in mice and humans receiving adoptive T cell-based therapies (Gattinoni et al., 2009; Sukumar et al., 2013). To test the anti-tumor effect of ΔΨm-sorted T cells, we adoptively transferred low-ΔΨm and high-ΔΨm T cell subsets into wild-type mice bearing subcutaneous syngeneic, large vascularized established B16 melanomas. Consistent with their memory phenotype, low-ΔΨm pmel-1 CD8⁺ T cells displayed significantly improved in vivo anti-tumor functionality and mediated improved tumor regression (Figure 6E). Thus, metabolic sorting based on low mitochondrial membrane potential can identify, within a given population, a subset of T cells with high capacity for self-renewal and long-term survival as immune memory. Consistent with their stem-like characterisitcs, these low-ΔΨm T cells also have the abity to differentiate in vivo into highly potent effector cells that can mediate autoimmune tissue damage and control large, established tumors.

Open in a separate window

Figure 6

Low-ΔΨm CD8⁺ T cells demonstrate increased long-term in vivo persistence and superior anti-tumor activity

(A–B) Flow cytometric analysis and quantification of persistence of transferred Thy1.1⁺ CD8⁺ T cells 300 days after adoptive transfer of 10⁵ high-ΔΨm and low-ΔΨm cells in sub-lethally irradiated mice (mean ± s.e.m. of 4 mice). (C–D) Vitiligo scores and melanophage uveitis scoring in mice that received ΔΨm CD8⁺ T cell subsets. (E) Adoptive transfer of 10⁶ high-ΔΨm or low-ΔΨm cells and subsequent infection with gp100-vaccinia virus in sub-lethally irradiated mice bearing 10-day established subcutaneous B16 tumors (n=5 mice per group). Data are presented as mean ± SEM and * = P < 0.05; ** = P < 0.01 (two-tailed t test). (E) ** P =0.009 (Wilcoxon rank sum).

Mitochondrial membrane potential sorting

CD8⁺ T cells from pmel-1 mice were stimulated in vitro with 1 µM hgp100_25–33 peptide and expanded for 4 days in culture medium containing 10 ng•ml⁻¹ IL-2 (Chiron). Membrane potential was assessed using the potentiometric dye tetramethyl rhodamine methyl ester (TMRM; Sigma Aldrich) at a final concentration of 25 nm for 30 min at 37°C. For sorting of cells with different ΔΨm, CD8⁺ gated cells were sorted into two pools corresponding to the cells with the lowest (7–10%) and highest (7–10%) TMRM fluorescence. Cell sorting was performed with a FACS Aria instrument (BD Biosciences). Generation of Tcm, Th1, Th17 and LT-HSC ΔΨm subsets was described in Supplementary Experimental Procedures.

Adoptive cell transfer and tumor experiments

Tumor experiments were performed as previously described (Gattinoni et al., 2009). Unless otherwise indicated, mice (n ≥ 5 per group) were injected s.c. with 2–5 × 10⁵ B16 melanoma cells. Ten to 14 days later, treated mice received i.v. injections of pmel-1 CD8⁺ T cells sorted based on ΔΨm and indicated doses of recombinant vaccinia virus expressing hgp100. Treated mice received 500 cGy of sublethal irradiation prior to ACT. In addition to vaccination, all treated mice also received intraperitoneal injections of rhIL-2 administered twice daily for 3 days after transfer.

Extracellular acidification rate (ECAR) and basal oxygen consumption rate (OCR)

OCR and ECAR were measured at 37°C using an XF24 extracellular analyzer (Seahorse Bioscience) as previously described (van der Windt et al., 2012). Oxygen consumption rates (OCR) and extracellular acidification rates (ECAR) were measured in XF media (nonbuffered RPMI 1640 containing 25 mM glucose, 2 mM L-glutamine, and 1 mM sodium pyruvate) under basal conditions and in response to 1 µM oligomycin, 2.0 µM fluoro-carbonyl cyanide phenylhydrazone (FCCP), or 100 nM rotenone with 1 µM antimycin A (Sigma).

Real-time RT-PCR

We isolated RNA with the RNeasy mini kit (Qiagen) and generated cDNA by reverse transcription (Applied Biosystems). Real-time RT-PCR was performed for all genes with primers from Applied Biosystems by Prism 7900HT (Applied Biosystems). Gene expression was calculated relative to Actb expression.

Intracellular ROS measurements

For analysis of intracellular ROS, CD8+ T cells were incubated with 5 µM dichlorofluorescein diacetate (DCFDA, Invitrogen). Cells were incubated on a shaker at 37 °C for 30 min, followed immediately by flow cytometry analysis using a LSRII instrument (Becton-Dickinson).

This research was supported with gifts from the Milstein family foundation and Mr. Jinyuan Li, chairman of the Tiens charitable foundation, the Leducq Foundation and by the Intramural Research Programs of the NCI and NHLBI. We thank A. Mixon and S. Farid of the Flow Cytometry Unit for help with sorting; and D. Goldstein and M. Malik for help with Metabolon. We thank Steven A. Rosenberg for his support. This study was done in partial fulfillment of a PhD in Biomedical Sciences (to S.J.P.) at the Georgetown University.