Research ArticleImmunometabolism

Immunometabolism regulates TCR recycling and iNKT cell functions

See allHide authors and affiliations

Science Signaling  26 Feb 2019:
Vol. 12, Issue 570, eaau1788
DOI: 10.1126/scisignal.aau1788

Glycolysis promotes recycling

Lymphocyte functions are linked to activation of cellular metabolic pathways. Yet, it remains unclear how metabolism influences the activity of invariant natural killer T (iNKT) cells, which express a T cell receptor with limited diversity. Fu et al. found that effector iNKT cells from the spleen and liver, which had an activated cell surface phenotype, were more glycolytic than naïve CD4+ T cells. T cell receptor stimulation of iNKT cells further increased glycolysis, which promoted the production of the proinflammatory cytokine interferon-γ. Pharmacological inhibitors of glycolysis reduced T cell receptor signaling and recycling to sites of antigen recognition at the cell surface. Thus, glycolysis may augment iNKT cell activation by sustaining the density of cell surface T cell receptor.

Abstract

Invariant natural killer T (iNKT) cells are innate-like T lymphocytes that express an invariant T cell receptor (TCR), which recognizes glycolipid antigens presented on CD1d molecules. These cells are phenotypically and functionally distinct from conventional T cells. When we characterized the metabolic activity of iNKT cells, consistent with their activated phenotype, we found that they had much less mitochondrial respiratory capacity but increased glycolytic activity in comparison to naïve conventional CD4+ T cells. After TCR engagement, iNKT cells further increased aerobic glycolysis, which was important for the expression of interferon-γ (IFN-γ). Glycolytic metabolism promoted the translocation of hexokinase-II to mitochondria and the activation of mammalian target of rapamycin complex 2 (mTORC2). Inhibiting glycolysis reduced the activity of Akt and PKCθ, which inhibited TCR recycling and accumulation within the immune synapse. Diminished TCR accumulation in the immune synapse reduced the activation of proximal and distal TCR signaling pathways and IFN-γ production in activated iNKT cells. Our studies demonstrate that glycolytic metabolism augments TCR signaling duration and IFN-γ production in iNKT cells by increasing TCR recycling.

INTRODUCTION

Immune cells use distinct metabolic pathways to match their states and functions. Metabolic dysregulation alters the fate and function of immune cells and is closely related to the progression of diseases like autoimmune diseases and cancers (13). The activation of naïve T cells is associated with the metabolic transition from fatty acid oxidation (FAO) and oxidative phosphorylation (OXPHOS) to aerobic glycolysis (47). Well described in tumor cells, aerobic glycolysis occurs when cells reduce pyruvate to lactate rather than oxidize it in tricarboxylic acid (TCA) cycle despite oxygen availability (8). Aerobic glycolysis is required for optimal T cell effector functions but not for the activation and expansion of naïve T cells (9). Transcription factors and signaling pathways that regulate glycolysis, such as hypoxia-inducible factor 1α, Myc, mammalian target of rapamycin (mTOR), and phosphoinositide 3-kinase–protein kinase B (or Akt), promote T cell effector differentiation (1014). Aerobic glycolysis also promotes the differentiation of CD4 T helper 1 (TH1), TH2, and TH17 cells, whereas FAO favors the differentiation of regulatory T cells (15, 16). In addition, the formation of T cell memory is also regulated by cellular metabolism. Enforcing FAO or limiting glycolysis favors generation of long-lived memory CD8 T cells, whereas skewing cellular metabolism toward aerobic glycolysis promotes differentiation and function of effector CD8 T cells (17). These studies highlight a critical link between cellular metabolism and T cell immunity, and the underlying mechanisms are not fully understood. The glycolytic enzyme glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and glycolytic metabolite phosphoenolpyruvate (PEP) regulate conventional T cell functions directly by promoting interferon-γ (IFN-γ) translation and Ca2+ signaling, respectively (9, 18). Whether other metabolic pathways or molecules regulate signal transduction and modulate T cell functions remains unclear.

Invariant natural killer T (iNKT) cells are innate-like T cells that are in the first line of immune defense against infectious diseases, cancers, autoimmune diseases, and metabolic diseases (1921). iNKT cells have different properties from conventional T cells. They are observed at low frequency in peripheral lymph nodes and white pulp of the spleen but accumulate in the red pulp of the spleen and vasculature of the liver and lung (2226). iNKT cells express a semi-invariant T cell receptor (TCR) and recognize CD1d-presented lipid antigens, as both self and foreign antigens (27, 28). They have effector phenotypes and respond rapidly to antigenic stimulation. Although iNKT cells release both TH1 and TH2 cytokines upon activation (29), iNKT cell polarization toward either TH1 or TH2 responses (3032) is attributable to distinct cytokines, antigen-presenting cells (APCs), and antigens (3337). TH1 polarization of iNKT cells could promote tumor rejection and pathogen clearance, whereas TH2 polarization of iNKT cells could promote tumor growth and inhibit autoimmune diseases (3841). The metabolic activity of iNKT cells has not been characterized. Whether metabolic alteration could cause functional polarization remains to be elucidated.

Here, we investigated how glycolytic metabolism in activated iNKT cells promoted TCR recycling, downstream signaling, and IFN-γ production. We found that in iNKT cells, glycolysis promoted the translocation of hexokinase-II (HK-II), a key enzyme of glycolysis, from cytosol to mitochondria after TCR activation. HK-II mitochondrial association was required for mTOR complex 2 (mTORC2) activation and TCR recycling. Inhibiting glycolysis in iNKT cells with 2-deoxy-d-glucose (2-DG) released HK-II from mitochondria and reduced the activation of mTORC2, Akt, and protein kinase C θ (PKCθ), which were required for TCR recycling. Thus, our data uncover a link between metabolic activation state and TCR recycling in iNKT cells that is necessary for prolonged TCR signaling and optimal IFN-γ production.

RESULTS

Activation augments iNKT cell glycolytic metabolism

To understand the metabolic activity of iNKT cells, we used a Seahorse bioanalyzer to measure their real time oxygen consumption rate (OCR) and extracellular acidification rate (ECAR). The OCR and ECAR are associated with mitochondrial respiration and lactate production, respectively. We compared metabolic activity of iNKT cells isolated from the spleen and liver that had an activated CD44+CD62L phenotype, to that of naïve CD44CD62L+ CD4 T cells (fig. S1, A and B). Whereas naïve CD4 T cells use FAO and OXPHOS for their energy supply (14), consistent with their effector phenotype (13), we found that iNKT cells used less oxygen than naïve CD4 T cells (fig. S1C) for both basal respiration and maximal respiration (fig. S1D). In addition, we found that iNKT cells had less mitochondrial mass (fig. S1E), lower mitochondrial membrane potential (fig. S1F), and reduced expression of respiratory genes (fig. S1G) than CD4 T cells. We also found reduced expression of the large neutral amino acid transporter CD98 (fig. S1, H and I) and the fatty acid β-oxidation enzyme Cpt1a mRNA (fig. S1J) in iNKT cells, comparing to CD4 T cells. These data indicated that iNKT cells were less dependent on mitochondrial OXPHOS and FAO than naïve CD4 T cells. In contrast, analysis of ECAR indicated that iNKT cells had increased basal glycolysis and maximal glycolytic capacity in comparison to CD4 T cells (fig. S1, K and L). Uptake of the fluorescent glucose analog 2-[N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl) amino]-2-deoxy-d-glucose (2-NBDG) (fig. S1M), production of lactate (fig. S1N), and expression of key proteins in glycolysis including Glut1 and PKM2 (fig. S1O) were all increased in iNKT cells when compared to naïve CD4 T cells. Overall, these data indicated that, consistent with their activation phenotypes, iNKT cells have lower mitochondrial respiratory capacity but higher glycolytic capacity than CD4 T cells.

Although iNKT cells were basally glycolytic, we found that they further increased utilization of glycolysis when activated with antibodies against CD3 and CD28 (Fig. 1, A and B). Activation of iNKT cells in glucose-free medium (Fig. 1F) or in the presence of 2-DG (Fig. 1G), a glycolysis inhibitor, confirmed that TCR activation increased aerobic glycolysis. Moreover, we found that after overnight activation, iNKT cells significantly increased expression of glycolytic genes (fig. S2, A and B), lactate production (fig. S2C), and uptake of 2-NBDG (fig. S2D). In contrast, activated iNKT only slightly increased basal mitochondrial respiration, and activation did not alter maximal respiration (Fig. 1, C and D). Although both ECAR and OCR were elevated 30 min after iNKT cell activation, the increased ratio of ECAR/OCR indicated that activated iNKT cells were still predominantly glycolytic (Fig. 1E). After overnight activation, we also found that iNKT cells significantly increased expression of genes involved in glutaminolysis and OXPHOS (fig. S2, A and B) and increased mitochondrial mass and mitochondrial membrane potential (fig. S2E). Together, our data demonstrate that TCR engagement not only primarily stimulated iNKT cell glycolytic metabolism but also increased utilization of mitochondrial pathways after prolonged activation.

Fig. 1 iNKT cells activate glycolysis after TCR stimulation.

(A to D) Seahorse extracellular flux analysis of ECAR (A) and OCR (C) on iNKT cells activated by antibodies against CD3 and CD28, as indicated. Basal glycolysis, glycolytic capacity, basal respiration, and maximal respiration were calculated, as indicated in (A) and (C). Normalized values (B and D) are means ± SEM of seven to nine biological replicates pooled from two independent experiments. (E to G) Seahorse extracellular flux analysis of OCR and ECAR on iNKT cells activated as indicated by antibodies against CD3 and CD28 in normal medium (E) or in glucose-free medium (F) or in the presence of 2-DG (G). Data are means ± SEM of nine biological replicates pooled from three independent experiments. **P < 0.01 and ***P < 0.001 by Student’s t test.

Glycolysis stimulates proliferation and IFN-γ production in iNKT cells

To understand what role glycolysis has in activating iNKT cell functions, we used 2-DG to inhibit glycolysis in CD1d-PBS57 tetramer–activated iNKT cells. This more physiologically relevant invariant TCR stimulation activated iNKT cells as efficiently as antibodies against CD3 and CD28 (Fig. 2A and fig. S3A). We found that 2-DG inhibited in a dose-dependent manner IFN-γ production by iNKT cells activated with either CD1d-PBS57 tetramer or antibodies against CD3 and CD28 (Fig. 2A and fig. S3A). However, we noted that 2-DG less potently suppressed iNKT cell interleukin-4 (IL-4) production. Whereas more than 90% of IFN-γ production was inhibited by 2-DG (5 mM), only about half IL-4 production was reduced by same dose of 2-DG (Fig. 2A and fig. S3A). Thus, the increased IL-4/IFN-γ ratio indicated that 2-DG polarized iNKT cells toward a TH2 response (Fig. 2B and fig. S3B). 2-DG also inhibited α-galactosylceramide (αGC)–induced IFN-γ production in vivo (Fig. 2, C to E). However, it did not reduce the mean fluorescence intensity of intracellular IL-4, which confirmed that 2-DG promoted iNKT cell TH2 polarization (Fig. 2, C to E). We obtained similar results after glucose depletion (Fig. 2, F and G) and exogenous addition of glucose-6-phosphate (G6P) (Fig. 2, H and I), an inhibitor of the glycolytic enzyme HK. During glycolysis, glucose is converted to pyruvate, which enters the TCA cycle in normal cells. Thus, 2-DG would also inhibit production of pyruvate and sequential oxidation of pyruvate in TCA cycle. However, 2-DG did not alter OXPHOS in activated iNKT cells (Fig. 1G). In addition, we found that pyruvate was unable to completely restore the IFN-γ production or IL-4/IFN-γ ratio in 2-DG–treated iNKT cells (Fig. 2, J and K). These results excluded the possibility that reduction of IFN-γ production induced by 2-DG attributed to insufficient pyruvate entering the TCA cycle. Together, our results demonstrated that glycolysis promoted IFN-γ production by iNKT cells.

Fig. 2 2-DG inhibits proliferation of iNKT cells and polarizes iNKT cells toward TH2 response.

(A and B) Cytometric bead array (CBA) analysis of IFN-γ and IL-4 production by iNKT cells activated by immobilized mCD1d-PBS57 tetramer overnight with 2-DG, as indicated. Data (A) and relative IL-4/IFN-γ ratio (B) are means ± SEM of nine independent biological replicates pooled from three independent experiments. (C to E) Flow cytometry analysis of intracellular IFN-γ and IL-4 in hepatic iNKT cells activated by αGC in vivo with 2-DG, as indicated. Dot plots (C) are representative of three independent experiments. Quantified percent cytokine+ iNKT cells (D) and cytokine mean fluorescence intensity (E) are means ± SEM of seven mice per group. (F and G) CBA analysis of IFN-γ and IL-4 production by iNKT cells activated with antibodies against CD3 and CD28 in medium with or without glucose, as indicated. Data (F) and IL-4/IFN-γ ratio (G) are means ± SEM of 9 biological replicates pooled from three independent experiments. (H and I) CBA analysis of IFN-γ and IL-4 production by iNKT cells activated with antibodies against CD3 and CD28 with or without G6P. Data (H) and IL-4/IFN-γ ratio (I) are means ± SEM of nine biological replicates pooled from three independent experiments. (J and K) CBA analysis of IFN-γ and IL-4 production by iNKT cells activated with immobilized CD1d-PBS57 tetramer overnight in the presence of 2-DG and pyruvate replenishment, as indicated. Data (J) and IL-4/IFN-γ ratio (K) are means ± SEM of nine biological replicates pooled from three independent experiments. (L) Flow cytometry analysis of surface CD69, CD25, and CD40L abundance on iNKT cells activated with immobilized CD1d-PBS57 tetramer overnight and 2-DG, as indicated. Data are means ± SEM of nine independent replicates pooled from three independent experiments. (M) Flow cytometry analysis of Ki67 expression in iNKT cells activated by immobilized mCD1d-PBS57 tetramer for 2 days with or without 2-DG. Data are means ± SEM of nine biological replicates pooled from three independent experiments. *P < 0.05, **P < 0.01, and ***P < 0.001 by Student’s t test.

After activation, iNKT cells proliferated and increased expression of several surface activation markers including CD69, CD25, and CD40L. We found that 2-DG did not alter the abundance of CD69 and CD40L but inhibited the activation-induced increase in CD25 (Fig. 2L). Although the frequency of iNKT cells that expressed the cell cycle marker Ki-67 after stimulation for 2 days was also reduced by 2-DG (Fig. 2M), we found no change in proliferation was apparent after overnight activation (fig. S4A). In addition, 2-DG did not reduce the total, IFN-γ+, or IL-4+ iNKT cell numbers after stimulation in vitro (fig. S4, B and C). Therefore, reduced IFN-γ production was not due to the impaired proliferation or reduced cell numbers. Together, our results demonstrated that glycolysis was required for proliferation and optimal IFN-γ production in iNKT cells.

Glycolysis favors IFN-γ production by promoting TCR recycling

In contrast to TCR stimulation, phorbol 12-myristate 13-acetate (PMA) plus ionomycin activates iNKT cells by activating PKC independent of TCR signaling. When we stimulated iNKT cells with PMA plus ionomycin, we found that 2-DG did not influence IFN-γ or IL-4 production (Fig. 3, A and B). These results suggested that 2-DG diminished IFN-γ production by inhibiting proximal TCR signaling. In contrast, we found that the proliferation of iNKT cells (fig. S5A) and abundance of CD25 (fig. S5B) were inhibited by 2-DG, and even cells were activated by PMA plus ionomycin. Therefore, 2-DG inhibited iNKT cell proliferation and CD25 expression by mechanisms downstream TCR signaling. Although IL-4 production requires only short TCR stimulation in iNKT cells, optimal IFN-γ production requires sustained TCR stimulation (42). We found that interrupting TCR signaling with anti-CD1d antibody at 4 hours after stimulation inhibited αGC-induced IFN-γ production but had less of an effect on IL-4 production (fig. S6). These results confirmed that IFN-γ production required prolonged TCR signaling in iNKT cells, as expected (42). It is possible that 2-DG may limit prolonged TCR signaling required for iNKT cell IFN-γ responses.

Fig. 3 2-DG inhibits iNKT cell TCR signaling.

(A and B) Flow cytometry analysis of intracellular IFN-γ and IL-4 in iNKT cells activated with PMA + ionomycin (P + I) and 2-DG for 4 hours, as indicated. Frequency data (A) and cytokine mean fluorescence intensity (MFI) (B) are means ± SEM of nine biological replicates pooled from three independent experiments. (C and D) TIRFM analysis of the distribution (C) and amount (D) of surface TCR at the synapses of iNKT cells activated with CD1d-PBS57 tetramer in the presence of phosphate-buffered saline (PBS) control or 2-DG at indicated time points. Images (C) are representative of six independent experiments (C). Quantified values are means ± SEM of 60 cells per group pooled from six independent experiments (D). Scale bars, 5 μm. A.U., arbitrary units. (E and F) Western blot analysis of phosphorylation of pLCK-Tyr394, pLAT-Tyr191, pPKCθ-Ser676, pP38 MAPK-Thr180/Tyr182, and pErk1/2-Thr202/Tyr204 in lysates of iNKT cells activated with antibodies against CD3 and CD28 for 45 min with or without 2-DG. Blots (E) are representative of three independent experiments. Quantified band intensity values (F) are means ± SEM pooled from three independent experiments. (G) CBA analysis of inhibition of IFN-γ and IL-4 production by 2-DG compared to PBS control added to iNKT cells after activating for 2 and 8 hours. Data are means ± SEM of nine biological replicates pooled from three independent experiments. *P < 0.05, **P < 0.01, and ***P < 0.001 by Student’s t test or by Wilcoxon test.

To investigate the influence of 2-DG on the duration of TCR signaling, we measured the accumulation of the TCRβ chain within immune synapses on the surface of iNKT cells using total internal reflection fluorescence microscopy (TIRFM) (43, 44). The stimulation of iNKT cells with CD1d-PBS57 tetramer–coated coverslips promoted clustering and accumulation of TCRβ at the contact site. When iNKT cells were treated with 2-DG before activation, accumulation of TCRβ was reduced at 45 and 90 min after stimulation (Fig. 3, C and D). However, there was no effect of 2-DG on the initial surface TCRβ clustering and accumulation after short 15-min stimulation. These results suggested that 2-DG did not influence the initiation of TCR clustering but inhibited the sustained TCR surface accumulation within immune synapses.

To determine whether 2-DG suppressed TCR signaling, we measured the activation of proximal and distal TCR signaling proteins after activating iNKT cells for 45 min. We found that the phosphorylation of Lck, linker of activated T cells (LAT), PKCθ, Erk1/2, and p38 mitogen-activated protein kinase (MAPK) were all significantly reduced in iNKT cells treated with 2-DG before TCR engagement (Fig. 3, E and F). Moreover, these effects were similar when 2-DG was added to iNKT cells either at the beginning of or 5 min after TCR engagement (fig. S7). These results indicated that 2-DG interfered with TCR signaling. Because IFN-γ production requires prolonged TCR signaling (42), we anticipated that inhibiting TCR signaling shortly after activation could efficiently dampen the IFN-γ response. When 2-DG was added to iNKT cells 2 hours after activation, it efficiently inhibited IFN-γ production. Consistent with our hypothesis, we found that 2-DG had less of an inhibitory effect when it was added 8 hours after iNKT cell activation (Fig. 3G).

Although 2-DG diminished TCR accumulation at immune synapses, it did not change the total amount of surface and intracellular TCRβ in iNKT cells (Fig. 4A). Thus, we investigated whether 2-DG influenced the internalization and recycling of iNKT TCRs (45, 46). By flow cytometry, we found that 2-DG profoundly inhibited TCR recycling but not TCR internalization in iNKT cells that were activated for longer than 45 min (Fig. 4, B and C). We found similar effects of 2-DG on TCR recycling in iNKT cells activated with antibodies against CD3 and CD28 (fig. S3, C and D), and the effects of 2-DG were not restored by pyruvate replenishment (Fig. 4D). Thus, 2-DG suppressed TCR recycling in iNKT cells activated by multiple stimuli, and its effects were not due to the shortage of pyruvate entering TCA cycle. We further investigated the distribution of intracellular TCRβ close to the cell surface using TIRFM after blocking surface TCR with unlabeled antibody before intracellular TCRβ staining. We found that TCR stimulation recruited intracellular pools of TCRβ to immune synapses of iNKT cells and that 2-DG treatment reduced the accumulation of membrane-adjacent intracellular TCRβ (Fig. 4, E and F). Because of the properties of the TIRF evanescent wave, the fluorescence intensity of TCRβ can be used as a measure of surface proximity. Thus, intracellular TCRβ closer to membrane were brighter than those deeper within iNKT cells. Treatment of iNKT cells with 2-DG reduced mean fluorescence intensity of intracellular TCRβ vesicles, suggesting they may be located further from the plasma membrane surface (Fig. 4G). In addition, we found that exogenous G6P also inhibited surface and intracellular TCRβ accumulation at immune synapses (Fig. 4, H to L). Together, these results implied that TCR recycling to immune synapses was impaired when glycolysis was inhibited in iNKT cells, which suggests that glycolysis favors IFN-γ production in iNKT cells by promoting TCR vesicle recycling.

Fig. 4 Glycolysis promotes TCR recycling.

(A) Flow cytometry analysis of total surface and intracellular TCR in iNKT cells activated with CD1d-PBS57 tetramer and PBS or 2-DG for the indicated times. Data are means ± SEM of nine biological replicates pooled from three independent experiments. (B and C) Flow cytometry analysis of TCR internalization (B) and recycling (C) in iNKT cells activated with CD1d-PBS57 tetramer and PBS or 2-DG for the indicated times. Data are means ± SEM of 12 biological replicates pooled from four independent experiments. (D) Flow cytometry analysis of TCR recycling in iNKT cells activated with CD1d-PBS57 tetramer and PBS, 2-DG, or 2-DG plus pyruvate for the indicated times. Data are means ± SEM of nine biological replicates pooled from three independent experiments. (E to G) TIRFM analysis of the distribution (E), density (F), and mean fluorescence intensity (G) of intracellular TCR in iNKT cells activated with CD1d-PBS57 tetramer and 2-DG as indicated for 45 min. Images (E) are representative of three independent experiments with threshold images are inserted. Quantified values (F and G) are means ± SEM of 100 cells per group pooled from three independent experiments. (H and I) TIRFM analysis of the distribution (H) and amount (I) of surface TCR at the synapses of iNKT cells activated with CD1d-PBS57 tetramer and G6P as indicated for 45 min. Images (H) are representative of three independent experiments. Quantified values (I) are means ± SEM of 60 cells per group pooled from three independent experiments. (J to L) TIRFM analysis of the distribution (J), density (K), and mean fluorescence intensity (L) of intracellular TCR in iNKT cells activated with CD1d-PBS57 tetramer and G6P as indicated for 45 min. Images (J) are representative of three independent experiments. Quantified values (K and L) are means ± SEM of 80 cells per group pooled from three independent experiments. Scale bars, 5 μm. *P < 0.05, **P < 0.01, and ***P < 0.001 by Student’s t test.

We detected reduced amount of Ifng mRNA and, to a lesser extent, Il4 mRNA in 2-DG–treated iNKT cells (fig. S8). This result is consistent with previous findings that glycolysis promotes transcription of Ifng in TH1 cells (47). However, deletion of lactate dehydrogenase A (LDHA) in CD4 T cells reduces Ifng mRNA by diminishing histone acetylation in Ifng locus (47), and inhibiting glycolysis with 2-DG did not reduce the surface TCR accumulation at synapses of CD4 T cells (fig. S9). Therefore, glycolysis may promote IFN-γ production by distinct mechanisms in effector phenotype iNKT cells and in naïve CD4 T cells, possibly due to their distinct metabolic status.

Translocation of HK-II to mitochondria promotes TCR accumulation at synapses

Among the four isoforms of HK, HK-II is the major contributor to aerobic glycolysis in tumor cells (48). In iNKT cells, abundant expression of HK-II was detected, and that further increased after long-term activation (fig. S2B). Different from other isoforms, HK-II translocates from cytosol to mitochondria to favor glycolysis (49). The translocation of HK-II to mitochondria is promoted by glucose and Akt, whereas it is inhibited by 2-DG, G6P, and glucose depletion (49, 50). In iNKT cells, we found that TCR engagement promoted HK-II colocalization with the mitochondria protein ATPB, whereas 2-DG and G6P reduced mitochondrial association of HK-II (Fig. 5, A and B). HK-II is associated with mitochondria through a 15–amino acid sequence in its N-terminal region, which interacts with voltage-dependent anion channel (VDAC) in the mitochondrial outer membrane (51). To investigate whether the translocation of HK-II influenced the function of iNKT cells, we used a HKVBD peptide that blocked the binding of HK-II to VDAC (52, 53) and found that it reduced the colocalization of HK-II with mitochondria, without inhibiting HK-II expression (Fig. 5, A and B, and fig. S10). This HKVBD peptide inhibited IFN-γ production and promoted iNKT cell TH2 polarization (Fig. 5, C and D). In addition, when cells were exposed to HKVBD peptide before TCR stimulation, the abundance of surface TCRβ within the immune synapse was reduced (Fig. 5, E and F). Peptide pretreatment also reduced the amount of intracellular TCRβ (Fig. 5, G to I) adjacent to the cell surface at the CD1d-PBS57 tetramer contact site. Thus, our data suggest that glycolytic metabolism promotes mitochondrial translocation of HK-II, which enhances TCR recycling and IFN-γ production.

Fig. 5 Dissociation of HK-II from mitochondria inhibits IFN-γ production and TCR recycling in iNKT cells.

(A and B) TIRFM analysis of the distribution (A) and colocalization (B) of mitochondrial marker ATPB and HK-II in iNKT cells activated with CD1d-PBS57 tetramer and HKVBD, 2-DG, or G6P as indicated for 45 min. Images (A) are representative of three independent experiments. Quantified values (B) are means ± SEM of 60 cells per group pooled from three independent experiments. (C and D) CBA analysis of IFN-γ and IL-4 production by iNKT cells activated with CD1d-PBS57 tetramer and HKVBD peptide, as indicated. Data (C) and IL-4/IFN-γ ratio (D) are means ± SEM of nine biological replicates pooled from three independent experiments. (E and F) TIRFM analysis of the distribution (E) and amount (F) of surface TCR at the synapses of iNKT cells activated with CD1d-PBS57 tetramer and HKVBD peptide as indicated for 45 min. Images (E) are representative of three independent experiments. Quantified data (F) are means ± SEM of 60 cells per group pooled from three independent experiments. (G to I) TIRFM analysis of the distribution (G), density (H), and mean fluorescence intensity (I) of intracellular TCR in iNKT cells after activation with CD1d-PBS57 tetramer and HKVBD peptide for 45 min. Images (G) are representative of three independent experiments. Quantified data (H and I) are means ± SEM of 80 cells per group pooled from three independent experiments. Scale bars, 5 μm. *P < 0.05 and ***P < 0.001 by Student’s t test.

Dissociation of HK-II from mitochondria inhibits mitochondrial translocation and activation of mTORC2

mTOR is a serine/threonine kinase that forms two distinct functional complexes termed mTORC1 and mTORC2; both of them promote glycolysis. Raptor and rictor are the defining component of mTORC1 and mTORC2, respectively (5457). mTORC2 localizes to mitochondria-associated endoplasmic reticulum membrane (MAM) where HK-II is associated with VDAC (58, 59). In addition, mTORC2 interacts with the IP3R-Grp75–VDAC complex and controls the association of HK-II with VDAC (59). However, whether mitochondrial HK-II influences mTORC2 activity remains unclear. We found that activation of iNKT cells with antibodies against CD3 and CD28 promoted colocalization of mTOR with ATPB+ mitochondria, whereas releasing HK-II from mitochondria by pretreatment of iNKT cells with HKVBD peptide, 2-DG, or G6P reduced the mitochondrial association of mTOR (Fig. 6, A and B). We also found impaired translocation of mTOR to lysosomes in iNKT cells treated with 2-DG, G6P, and HKVBD peptide (fig. S11). Because mTORC1 localizes in the lysosome and mTORC2 localizes to the MAM (60), these data suggest that glycolysis inhibitors reduce mitochondria-associated mTOR complexes, most likely mTORC2. Consistent with our hypothesis, we noted that 2-DG, G6P, and HKVBD peptide reduced the abundance of the mTORC2 complex components mTOR and rictor in crude mitochondrial fractions from activated iNKT cells (Fig. 6, C and D). These results indicated blockage of glycolysis-impaired localization of mTORC2 at MAM. As readout for mTORC2 activity, when we measured the phosphorylation of Akt at S473 in activated iNKT cells, we found that 2-DG, G6P, and HKVBD peptide treatment diminished phosphorylation at this site (Fig. 6, E and F). In addition to Akt, other protein kinase A, G, and C (AGC) subfamily kinases are direct substrates phosphorylated by mTORC2 (6164), including serum/glucocorticoid-regulated kinase and PKCθ at S696 (65). We found that the phosphorylation of PKCθ at S696 was reduced by release of HK-II from mitochondria, which further indicated that the activation of mTORC2 was impaired by these inhibitors (Fig. 6, E and F). Thus, our results demonstrate that the translocation of HK-II to the mitochondria also promotes translocation of mTORC2 components to mitochondria, which enhances the activation of mTORC2.

Fig. 6 Dissociation of HK-II from mitochondria inhibits mTORC2 activation.

(A and B) TIRFM analysis of the distribution (A) and colocalization (B) of the mitochondrial marker ATPB and mTOR in iNKT cells activated with CD1d-PBS57 tetramer and HKVBD peptide, 2-DG, or G6P as indicated for 45 min. Images (A) are representative of three independent experiments. Quantified values are means ± SEM of 90 cells per group pooled from three independent experiments. Scale bars, 5 μm. (C and D) Western blot analysis of mTOR and rictor in crude mitochondrial fractions from expanded iNKT cells activated with antibodies against CD3 and CD28, and HKVBD, 2-DG, or G6P as indicated for 45 min. Blots (C) are representative of three independent experiments. Quantified band intensity values (D) are means ± SEM pooled from three independent experiments. (E and F) Western blot analysis of pAKT-Ser473 and pPKCθ-Ser676 in lysates from iNKT cells activated with antibodies against CD3 and CD28, and HKVBD, 2-DG, or G6P as indicated for 45 min. Blots (E) are representative of three independent experiments. Quantified band intensity values (F) are means ± SEM polled from three independent experiments. *P < 0.05, **P < 0.01, and ***P < 0.001 by Student’s t test or by Wilcoxon test.

AKT and PKCθ pathways promote TCR recycling and IFN-γ production in iNKT cells

Both Akt and PKCθ have been reported to regulate vesicle trafficking (6668). To determine which pathways promoted iNKT cell function, we inhibited Akt with MK2206 and PKCθ with sotrastaurin and measured cytokine production. Both MK2206 and sotrastaurin significantly diminished IFN-γ production and promoted TH2 polarization of iNKT cells (Fig. 7, A, B, H, and I). To understand whether these same pathways promoted TCR recycling, we treated iNKT cells with MK2206 and sotrastaurin and used TIRFM to determine the TCRβ accumulation at the contact site. We found that inhibitor treatment reduced the accumulation of surface and intracellular TCR at the immune synapse after TCR stimulation (Fig. 7, C to G and J to N). Overall, our results demonstrate that Akt and PKCθ activity in iNKT cells are required for TCR cell surface accumulation and optimal IFN-γ production.

Fig. 7 AKT and PKCθ promote TCR recycling and IFN-γ production in iNKT cells.

(A and B) CBA analysis of IFN-γ and IL-4 production by iNKT cells activated with CD1d-PBS57 tetramer and PKCθ inhibitor sotrastaurin (Sotra), as indicated. Data (A) and IL-4/IFN-γ ratio (B) are means ± SEM of nine biological replicates pooled from three independent experiments. (C and D) TIRFM analysis of the distribution (C) and amount (D) of surface TCR at the synapses of iNKT cells activated with CD1d-PBS57 tetramer and Sotra as indicated for 45 min. Images (C) are representative of three independent experiments. Quantified values (D) are means ± SEM of 120 cells per group pooled from three independent experiments. (E to G) TIRFM analysis of the distribution (E), density (F), and mean fluorescence intensity (G) of intracellular TCR in iNKT cells activated with CD1d-PBS57 tetramer and Sotra as indicated for 45 min. Images (E) are representative of three independent experiments. Quantified values (F and G) are means ± SEM of 60 cells per group pooled from three independent experiments. (H and I) CBA analysis of IFN-γ and IL-4 by iNKT cells activated with CD1d-PBS57 tetramer and Akt inhibitor MK2206 as indicated. Data production (H) and IL-4/IFN-γ ratio (I) are means ± SEM of nine biological replicates pooled from three independent experiments. (J to N) TIRFM analysis of distribution (J) and amount (K) of surface TCR, or the distribution (L), density (M), and mean fluorescence intensity (N) of intracellular TCR at the synapses of iNKT cells activated with CD1d-PBS57 tetramer and MK2206 as indicated for 45 min. Images (J and L) are representative of three independent experiments. Quantified values (K, M, and N) are means ± SEM of 60 cells per group pooled from three independent experiments. Scale bars, 5 μm. *P < 0.05, **P < 0.01, and ***P < 0.001 by Student’s t test.

DISCUSSION

Intracellular metabolism influences immune cell fates and functions (1, 2). Understanding how may indicate a new way to modulate host immunity against tumor and autoimmune diseases (3, 69). It remains unclear how cellular metabolism influences the function of peripheral innate-like iNKT lymphocytes, which have many different properties from conventional T cells. As a metabolic regulator, mTOR activity is important for the development of iNKT cells (70, 71). Here, we demonstrated that glycolytic metabolism activated mTORC2 necessary for the recycling of cell surface TCRs, which augmented IFN-γ production in iNKT cells. We isolated effector phenotype iNKT cells and showed that they were more glycolytically active than naïve CD4 T cells (fig. S1, K and L). In agreement with their activated cell surface phenotypes, iNKT cells also exhibited less mitochondrial respiratory activity than CD4 T cells (fig. S1, C and D). Similar to other activated T cell subsets (9), this preference for glucose utilization might favor their rapid responses to stimuli. Using live cell metabolic analysis, we demonstrated that iNKT cells increased glycolysis and OXPHOS after TCR engagement. However, aerobic glycolysis was the predominant bioenergetic pathway used in iNKT cells after TCR stimulation (Fig. 1E). We found that further increased glycolysis in iNKT cells, in turn, sustained TCR signaling by promoting TCR recycling. This positive feedback loop enhances the duration of TCR signaling, which augments TH1 response in iNKT cells. Our results demonstrate a previously unrecognized link between glycolysis and prolonged TCR signaling in iNKT cells. However, iNKT cells also increased OXPHOS after overnight activation (fig. S2), which suggests that this pathway may promote iNKT cell functions at late phase of activation. Here, we propose that the mitochondrial localization of HK-II stimulates mTORC2 activity in iNKT cells. HK-II is a key enzyme catalyzing the first step of glycolysis (72). In addition to its catalytic function, HK-II augments mitochondria functions, autophagy, and cell survival through protein-protein interactions (48, 73). In response to glucose and activation signals, HK-II translocates from cytosol to mitochondria to augment glycolysis in distinct cell types including hamster ovary cells, dendritic cells (DCs), and skeletal muscle cells (49, 52, 74). We found that inhibition of glycolysis with 2-DG and G6P prevented HK-II translocation to mitochondria (Fig. 5A), which also inhibited translocation of mTOR and rictor, key components of the mTORC2 complex (Fig. 6, A and C). Both mTORC2 and HK-II are associated with VDAC on mitochondria, and active mTORC2 localizes on contact site between mitochondria and ER via interacting with IP3R-Grp75–VDAC complex, where it promotes phosphorylation and binding of HK-II to VDAC (48, 59). Whether release of HK-II from VDAC would disrupt interactions between mTORC2 and IP3R-Grp75–VDAC complex requires further investigation. Although T cell–specific deletion of rictor markedly decreased stage 2 and stage 3 iNKT cells, it did not influence the IL-4 and IFN-γ productions in peripheral iNKT cells (75). The paradoxical result could be caused by genetic compensation in knockout mice. mTORC2 uniquely promotes IL-17–secreting iNKT cell differentiation in thymus (71, 75). In contrast, all our studies were performed with either splenic iNKT cells or hepatic iNKT cells, which are iNKT1 cells (76). Acute inactivation of glycolysis with inhibitors demonstrated the role of mTORC2 in promoting mature iNKT cell IFN-γ production. In contrast, we found that 2-DG did not influence the cytokine production of iNKT cells when they were stimulated with PMA plus ionomycin. Thus, the effects of glycolysis inhibition in iNKT cells required the TCR signaling. Our results suggested that mTORC2 may link glycolytic activity to intracellular signal transduction. Glucose is not the only metabolite that regulates TCR signaling. Both the amino acid arginine and the lipid cholesterol promote TCR activation (7780). It is also possible that TCR signal transduction could be influenced by the products of cellular metabolism at distinct sites.

In activated CD4 T cells, glycolysis promotes IFN-γ production in a posttranscriptional manner (9). Inhibition of glycolysis diminishes IFN-γ translation through binding Ifng mRNA with spare GAPDH. In our studies, Ifng mRNA was reduced by 2-DG because of its inhibitory effect on TCR signaling, consistent with shortened TCR stimulation, inhibiting Ifng transcription in iNKT cells (42). Inhibiting TCR signal transduction with 2-DG significantly reduced expression of glycolytic genes in iNKT cells after overnight activation (fig. S8). Thus, we could not exclude the possibility that further shutdown of glycolysis at a later phase of activation may free more GAPDH to repress IFN-γ translation. In addition, the glycolytic metabolite PEP sustains intracellular calcium by blocking sarco/endoplasmic reticulum Ca2+-ATPase (SERCA) (18). However, PEP was unable to restore the IFN-γ production in our studies on activated iNKT cells (fig. S12A). Thus, PEP shortage is not the only reason reducing IFN-γ production in iNKT cells when glycolysis is inhibited. Our results also showed that impaired translocation of HK-II dampened AKT and PKCθ activities as well. Overall, these studies demonstrate that multiple molecules associated with glycolysis may stimulate iNKT cell effector function.

In addition to inhibiting glycolysis, 2-DG also inhibits protein glycosylation (81). However, N-glycosylation can inhibit TCR signaling (8284). Thus, reducing glycosylation with 2-DG should favor, not suppress TCR signaling. In addition, when we used d-mannose to restore glycosylation in 2-DG–treated cells (81, 85), d-mannose failed to restore IFN-γ production (fig. S12B). These results further suggest that 2-DG did not inhibit TCR signaling by preventing protein glycosylation.

As tissue resident innate like lymphocytes (22), iNKT cells play important roles in maintaining tissue homeostasis. The high glycolytic activity would favor their production of inflammatory cytokines, as the first line of defense against tissue infection. We noted that glycolysis is differentially involved in IFN-γ and IL-4 production, which suggests that IFN-γ production might be a more energy-consuming process than IL-4 production (Fig. 2A). We found that iNKT cells were plastic and polarized their cytokine secretion functions from TH1/TH2 toward TH2 when glucose was limited. Thus, during infections metabolic alteration of microenvironment may skew iNKT cell polarization. Because many of the proinflammatory cells use glycolysis for energy production, such as M1 macrophages, DCs, TH1 T cells, cytotoxic T lymphocytes, and NK cells (17, 52, 8690), this switch in iNKT activity may help to maintain immune balance. Similarly, glucose is exhausted in tumor microenvironment, which influences functions of resident immune cells (18, 91, 92). Our studies would suggest that glucose exhaustion by tumor cells would predictably shorten TCR signaling and inhibits IFN-γ production in iNKT cells. Reduced IFN-γ production in iNKT cells has been reported in patients with tumor (9395). Overcoming impaired glycolysis may help to restore the IFN-γ production and promote the antitumor effect of iNKT cells. Overall, our studies demonstrate the mechanisms by which glycolysis influences the function of mature iNKT cells.

MATERIALS AND METHODS

Mice

Wild-type mice were purchased from the Model Animal Research Center of Nanjing University. Vα14 Tg.cxcr6gfp/+ mice have been previously described (96). All mice used were on the C57BL/6J background and 10 to 12 weeks old. Mice were housed under specific pathogen–free conditions. All animal procedures were approved by the University of Science and Technology of China Institutional Animal Care and Use Committee. All experiments were performed in accordance with the approved guidelines.

Cell isolation and activation

iNKT Cells, gating as GFPhi cells, were sorted from livers or spleens of Vα14 Tg.cxcr6gfp/+ mice by FACSAria (BD Biosciences). The purity of iNKT cells was more than 90% (fig. S1B). CD4 T cells gating out iNKT cells were sorted from the spleens. Purified iNKT cells were stimulated with plate-coated mCD1d-PBS57 tetramer (1 μg per well) or plate-coated antibodies against CD3 (1 μg per well) and CD28 (1 μg per well), with or without 2-DG (5 mM), G6P (25 mM), or HKVBD peptide (20 μM). Cytokines in supernatants were measured by CBA kit (BD Biosciences) after overnight activation. In proliferation assays, cells were stimulated with plate-coated mCD1d-PBS57 tetramer for 2 days, and cell proliferation was measured by Ki67 staining (BD Biosciences). To compare the expression of activation markers, iNKT cells were stimulated with plate-coated mCD1d-PBS57 tetramer overnight. In pyruvate supplementary assays, pyruvate (10 mM) was added to iNKT cells stimulated with plate-coated mCD1d-PBS57 tetramer overnight with or without 2-DG. In TCR blocking assays, enriched iNKT cells were stimulated with αGC-pulsed RBL.CD1d, and TCR signaling was blocked by antibodies against CD1d (10 μg/ml) at the indicated time points. To investigate the in vivo effect of 2-DG on iNKT cells, mice were administered intraperitoneally with 6 mg of 2-DG or PBS buffer, 1 hour before injecting 2 μg of αGC per mouse. Mice were sacrificed 3 hours later, and intracellular IL-4 and IFN-γ in hepatic iNKT cells were measured by flow cytometry. 2-DG, G6P, and pyruvate were purchased from Sigma-Aldrich. AKT inhibitor MK2206 and PKCθ inhibitor sotrastaurin were purchased from Selleck. HKVBD peptide was synthesized according to previous studies (52).

Antibodies and flow cytometry

Fluorochrome-labeled or unlabeled monoclonal antibodies against mouse TCRβ (H57-597), CD1d (1B1), CD4 (GK1.5), CD69 (H1.2F3), CD40L (MR1), CD25 (PC61), CD3 (145-2C11), CD28 (37.51), IFN-γ (XMG1.2), and IL-4 (11B11) were purchased from BD or BioLegend. CD1d-PBS57 tetramer was provided by the National Institutes of Health (NIH) Tetramer Core Facility. To analyze total TCR expression, including both surface TCR and intracellular TCR, cells were stained with Alexa 647–labeled anti-TCRβ (5 μg/ml) for 60 min on ice and then were fixed in 4% paraformaldehyde for 15 min on ice and permeabilized with buffer containing 5% fetal bovine serum (FBS) and 0.1% Triton X-100 for 30 min on ice. Intracellular TCR was then stained with Alexa 647–labeled anti-TCRβ (5 μg/ml) for 60 min. Samples were acquired by a BD FACSVerse flow cytometry, and data were analyzed with FlowJo software (TreeStar).

Metabolic assays

OCR and ECAR were measured using an XF96 extracellular analyzer (Seahorse Bioscience). Seahorse cell plates were coated with poly-l-lysine for 40 min, washed, and then coated with or without antibodies (1 μg per well) against CD3 and CD28 overnight at 37°C. Sorted mouse iNKT cells were resuspended in medium [minimal Dulbecco’s minimum essential medium (DMEM) with 1 mM sodium pyruvate, 2 mM glutamine, and with or without 10 mM glucose, as indicated (pH 7.4)] and were seeded in the plates (5 × 105 cells per well in 180-μl medium). After centrifugation at 1500 rpm for 3 min, the plate was loaded into the instrument to determine the real-time OCR and ECAR. To measure mitochondrial respiration, cells were consecutively exposed to mitochondria-perturbing reagents, oligomycin (Oligo, 1 μM), carbonyl cyanide p-trifluoromethoxyphenylhydrazone (FCCP) (1 μM), and lastly, rotenone plus antimycin A (Anti + rot, 0.5 μM). Oligomycin inhibits adenosine triphosphate (ATP) synthase and reduces mitochondrial respiration associated with ATP production, FCCP uncouples oxygen consumption from ATP production and raises oxygen consumption to a maximal value, and antimycin A plus rotenone target the electron transport chain and reduce OCR to a minimal value. The difference between OCR before oligomycin injection and OCR after rotenone plus antimycin A injection is defined as basal respiration, and difference between OCR after FCCP injection and OCR after rotenone plus antimycin A injection is defined as maximal respiration. For glycolysis stress test, medium without glucose or pyruvate was used, and glucose (Glu, 10 mM), oligomycin (Oligo, 1 μM), and 2-DG (10 mM) were consecutively added to cells. Glucose induces glycolysis, oligomycin inhibits ATP synthase and further shifts the energy production to glycolysis, and 2-DG inhibits glycolysis. The difference between ECAR after glucose injection and ECAR after 2-DG injection is defined as basal glycolysis, and difference between ECAR after oligomycin injection and ECAR after 2-DG injection is defined as maximal glycolytic capacity. To measure the OCR and ECAR of iNKT cells in different medium, iNKT cells were activated for 30 min and then were used for the measurements. To measure glucose uptake, inactivated iNKT cells and iNKT cells activated by immobilized anti-CD3 plus anti-CD28 overnight were incubated with 2-NBDG (100 μM; Invitrogen) for 10 min. To measure mitochondria activities, iNKT cells were incubated with MitoTracker Green (100 nM) for 15 min or incubated with tetramethylrhodamine (TMRM) (100 nM) for 30 min. The concentrations of lactate in culture medium were measured by lactate assay kit (Sigma-Aldrich).

Western blot

iNKT cells were expanded with αGC in vitro in the presence of IL-2 (100 U/ml) for 7 days. Sorted iNKT cells or in vitro expanded iNKT cells were treated with or without 2-DG (5 mM), G6P (25 mM), and HKVBD (20 μM) for 30 to 60 min, respectively, and then were activated with immobilized antibodies against CD3 and CD28 or CD1d-PBS57 tetramer for 45 min in the presence or absence of inhibitors. In some experiments, 2-DG was added to culture medium at the beginning of or 5 min after activation, as indicated. Cells were harvested and lysed with sample buffer and boiled for 10 min. Proteins were separated by electrophoresis and detected by Western blot. To isolate crude mitochondria, in vitro expanded iNKT cells were resuspended in 1-ml precooled isolation buffer [10 mM KCl2, 1.5 mM MgCl2, and 10 mM tris-HCl (pH 7.4)] and then were homogenized in a precooled 2-ml Dounce Homogenizer according to previous study (97). Homogenate was centrifuged 1000g for 5 min at 4°C, and then supernatant was collected and centrifuged 8000g for 30 min at 4°C. Pellet containing mitochondrial fraction was then washed twice. Antibodies against p-Lck, Lck, p-LAT, LAT, p-AKT, AKT, p-PKCθ, PKCθ, p-ERK1/2, ERK1/2, p-p38 MAPK, p38 MAPK, mTOR, rictor, LDHA, PKM2, Glut1, HK-II, and β-actin were purchased from Cell Signaling Technology, Sigma-Aldrich, or Proteintech.

RNA isolation and qPCR

Total mRNA was isolated and reverse-transcribed with Reverse Transcription System (Promega A3500). Real-time quantitative polymerase chain reaction (qPCR) was performed with PikoReal 96 (Thermo Fisher Scientific). Samples were run in triplicate and were normalized to β-actin to determine relative expression level. Primer sequences are listed in table S1.

TCR recycling assay

Purified iNKT cells were treated with or without 2-DG (5 mM) for 30 to 60 min and then were activated by immobilized CD1d-PBS57 tetramer for indicated time points with or without 2-DG. Surface TCR was blocked by saturated unconjugated anti-TCRβ antibody (10 μg/ml). After washing, cells were transferred to a 37°C water bath for indicated times to allow TCR recycling and then placed on ice. Recycled TCR was stained with APC-conjugated anti-TCRβ (1 μg/ml). The recycling pool of TCR was calculated as the percentage of total surface TCR, using previously described equation ((HC)/T) 100%, where H is the recycled TCR detected at indicated times, C is the recycled TCR detected at 0 min, and T is the total surface TCR (98).

TCR internalization assay

Purified iNKT cells were treated with or without 2-DG (5 mM) for 30 to 60 min, and then were activated by immobilized CD1d-PBS57 tetramer for different time points with or without 2-DG. Cells were incubated for 30 min on ice with APC-conjugated anti-TCRβ (1 μg/ml). After washing, the cells were transferred to a 37°C water bath for the indicated times to allow TCR internalization and then placed on ice. Noninternalized TCR is stripped from the cell surface using ice-cold acid stripping buffer (RPMI 1640 plus 1% bovine serum albumin, pH 3.5 adjusted with HCl) for 15 min on a shaking platform. The internalized TCR was detected by flow cytometry, and the mean fluorescence intensity of internalized TCR was calculated by subtracting the mean fluorescence intensity of isotype control.

Total internal reflection fluorescence microscopy

Round coverslips (18 mm; Thermo Fisher Scientific) were cleaned and coated with poly-l-lysine for 15 min. CD1d-PBS57 tetramer (3 μg per glass) or tetramerized antibody against CD3 (3 μg per glass) were used to coat the coverslips overnight at 37°C. For CD3 tetramer preparation, antibody against CD3-biotin was mixed with streptavidin in a 1:4 molar ratio at room temperature for 2 hours. GFPhi iNKT cells were treated with or without 2-DG (5 mM), G6P (25 mM), HKVBD (20 μM), sotrastaurin (0.25 μM), and MK2206 (0.1 μM) for 60 min, respectively, and then were activated by on tetramer-coated coverslips for 15, 45, or 90 min in the presence or absence of inhibitors. To detect surface TCR distribution, cells were stained with Alexa 647–labeled anti-TCRβ (5 μg/ml) for 60 min on ice and then fixed in 4% paraformaldehyde for 15 min on ice. To view the intracellular TCRβ, surface TCR was blocked with saturated unconjugated anti-TCRβ (10 μg/ml), and then cells were fixed in 4% paraformaldehyde and permeabilized with buffer containing 5% FBS plus 0.1% Triton X-100 for 30 min on ice. Intracellular TCR was stained with Alexa 647–labeled anti-TCRβ (5 μg/ml) for 60 min. To investigate distribution of mitochondria, HK-II, and mTOR, cells were fixed with 100% methanol for 5 min, permeabilized with 0.1% Triton X-100 for 5 min, and then were blocked with PBS buffer containing 10% normal goat serum and 0.1% Tween 20 for 1 hour. Cells were stained with anti-ATPB (3D5, 5 μg/ml), anti-HK-II (H.738.7, 5 μg/ml), and anti-mTOR (7C10, 5 μg/ml) for 3 hours on ice. Samples were visualized with a Leica SR GSD microscope with the 160× oil objective and 100-nm penetration depth. Quantitative analyses of fluorescence intensity, vesicle density, and colocalization coefficient were performed using ImageJ software. To analyze the vesicle fluorescence intensity and vesicle density, same threshold range was set for all images to perform automatic particle analysis. Vesicle density was calculated by dividing the number of vesicles by the area of immune synapse. To analyze the colocalization coefficient, a “rolling ball” algorithm was used to subtract background, and then Coloc 2 plugin was used to calculate the Manders’ coefficients.

RNA-seq analysis

RNA sequencing (RNA-seq) was performed using map analysis and plotting server (MAPS), as previously described (99). Briefly, total RNA was extracted from iNKT cells using TRIzol reagent (Invitrogen) and was reverse-transcribed using a biotinylated oligo (dT) primer and SuperScript III (Invitrogen). Terminal transferase was used to block 3′ end of complementary DNA (cDNA) in the presence of dideoxynucleotides (ddNTP). After second-strand cDNA synthesis, 23 cycles of PCR were performed to amplify cDNAs. PCR products with lengths of 200 to 400 nt were subjected to deep sequencing on a HiSeq 2500 (Illumina). The reads of RNA-seq were mapped to the mouse reference genome using Bowtie (v1.1.1).

Statistical analysis

Datasets were analyzed using Prism (GraphPad Software, version 5.01).

SUPPLEMENTARY MATERIALS

www.sciencesignaling.org/cgi/content/full/12/570/eaau1788/DC1

Fig. S1. Reduced respiratory capacity but increased glycolytic activity in iNKT cells than in CD4 T cells.

Fig. S2. Metabolic reprogramming after long-term activation.

Fig. S3. 2-DG inhibits IFN-γ production and TCR recycling in iNKT cells activated by anti-CD3 plus anti-CD28.

Fig. S4. 2-DG does not reduce cell numbers of iNKT cells.

Fig. S5. Influences of 2-DG on proliferation and activation markers of iNKT cells activated by PMA plus ionomycin.

Fig. S6. Shortening TCR signaling inhibits IFN-γ production.

Fig. S7. 2-DG inhibits TCR signaling proteins even after TCR engagement.

Fig. S8. 2-DG inhibits gene transcription.

Fig. S9. 2-DG does not influence accumulation of TCR at synapses of CD4 T cells.

Fig. S10. Inhibitors do not influence the expression of HK-II.

Fig. S11. Inhibition of glycolysis reduces translocation of mTOR to lysosomes.

Fig. S12. PEP and d-mannose do not restore IFN-γ production in 2-DG–treated iNKT cells.

Table S1. List of primer sequences.

REFERENCES AND NOTES

Acknowledgments: We thank NIH Tetramer Core Facility for providing us CD1d-PBS57 tetramer and A. Bendelac for providing us Vα14 Tg mice. Funding: This work was supported by National Key R&D Program of China (2017YFA0505300 to L.B.), National Natural Science Foundation of China (91542203 and 81771671 to L.B.), the Strategic Priority Research Program of the Chinese Academy of Sciences (XDA12030208 to L.B.), and the Fundamental Research Funds for the Central Universities (to L.B.). Author contributions: S.F. and S.Z. contributed equally to the paper. S.F., S.Z., C.T., S.B., J.Z., C.Z., D.X., L.W., Z.L., and J.L. performed the experiments. R.Z., Z.T., and T.X. provided materials and helped with the data analysis. S.F., S.Z., H.Z., and L.B. designed the experiments and wrote the manuscript. Competing interests: The authors declare that they have no competing financial interests. Data and materials availability: The RNA-seq data have been deposited at the GEO: GSE119346 (www.ncbi.nlm.nih.gov/geo/). All other data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials.
View Abstract

Stay Connected to Science Signaling

Navigate This Article