Abstract
Indoleamine 2,3-dioxygenase (IDO) suppresses adaptive immunity by inhibiting T-cell proliferation and altering glucose metabolism. The tumor suppressor p53 also alters these cellular processes with similar results. The effect of IDO on p53 and on glucose metabolism was evaluated in alloreactive T cells. Mixed-lymphocyte reactions (MLRs) were performed in the presence or not of the IDO inhibitor, 1-dl-methyl-tryptophan (1-MT) and/or the p53 inhibitor, pifithrin-α (PFT). Cell proliferation, glucose consumption and lactate production were assessed. 1-MT increased cell proliferation, glucose influx and lactate production, whereas PFT enhanced cell proliferation and glucose influx, leaving lactate production unaffected. In MLR-derived T cells, protein analysis revealed that IDO activated general control non-derepressible 2 kinase and induced p53, p-p53 (p53 phosphorylated at serine 15) and p21. In addition, both IDO and p53 decreased glucose transporter 1 and TP53-induced glycolysis and apoptosis regulator and increased synthesis of cytochrome c oxidase 2. IDO also reduced lactate dehydrogenase-A and glutaminase 2 levels, whereas p53 left them unaffected. Neither 1-MT nor PFT affected glucose-6-phosphate dehydrogenase. In conclusion, in alloreactive T cells, IDO increases p53 levels, and both IDO and p53 inhibit cell proliferation, glucose consumption and glycolysis. Lactate production and glutaminolysis are also suppressed by IDO, but not by p53.
Introduction
Various inflammatory stimuli induce indoleamine 2,3-dioxygenase (IDO) expression in antigen-presenting cells (APCs), such as monocytes, macrophages and dendritic cells. In the local microenvironment, IDO, which catalyses the initial rate-limiting step of tryptophan degradation along the kynurenine pathway, depletes tryptophan and suppresses adaptive immunity (1, 2). Tryptophan depletion by IDO, through uncharged tryptophan tRNA, is sensed and activates the general control non-derepressible 2 (GCN2) kinase, which in turn phosphorylates the eukaryotic initiation factor 2α (eIF2α), altering the translation program of the T cells ultimately leading to the inhibition of their proliferation and to anergy (3). Another pathway able to sense tryptophan depletion is the mammalian target of rapamycin complex 1 (mTORC1) pathway (4, 5).
IDO-mediated immunosuppression ameliorates the clinical course of autoimmune diseases (6–8) and reduces graft rejection (9–11). Furthermore, expression of IDO in non-APC cells types, such as in tumor cells, contributes to the escape of the tumor from immunosurveillance (12), and its expression in paternally derived placental trophoblasts contributes to a successful semi-allogenic pregnancy (13, 14). Hemodialysis patients are characterized by impaired adaptive immunity and exhibit increased IDO expression, further enhanced in non-responders to hepatitis B virus vaccination (15). Consequently, IDO serves as a general control mechanism of the adaptive immune response, so that elucidation of the exact mechanisms involved in its action will lead to a better understanding of immune system physiology and pathophysiology and possibly to therapeutic applications.
Interestingly, in a previous study, we showed that IDO inhibits glucose uptake, aerobic glycolysis and glutaminolysis, which may be responsible for its immunosuppressive effect (16). In order to fulfill the bioenergetic and biosynthetic demands of proliferation, activated T cells reprogram their metabolic pathways from pyruvate oxidation via the Krebs’s cycle to the glycolytic, pentose-phosphate and glutaminolytic pathways (17).
Similar metabolic reprogramming has been observed in other rapidly proliferating cells, such as cancer cells. Indeed, most rapidly proliferating cancer cells are characterized by an increased ratio of cytoplasmic glycolysis to mitochondrial glucose oxidation, a phenomenon described almost a century ago by Otto Warburg, and hence named as the Warburg’s phenomenon or aerobic glycolysis (18).
Most cancer cells carry mutations that directly or indirectly affect the function of the tumor suppressor p53, which directly inhibits cell proliferation by inducing G1-phase cell-cycle arrest through activation of transcription of the cyclin-dependent kinase inhibitor p21WAF1 (p21) (19). Interestingly, p53 also modulates cell metabolism by inhibiting glucose uptake and aerobic glycolysis, and increasing glutaminolysis and components involved in oxidative phosphorylation (19–21).
Considering the similarities in glucose metabolism and cell proliferation between activated T cells and cancer cells, as well as the inhibition of aerobic glycolysis by IDO in T cells, we evaluated the hypothesis that IDO-induced tryptophan-depletion stress increases p53 levels, a fact that may contribute to the suppression of both proliferation and aerobic glycolysis in T cells. For this purpose the two-way mixed-lymphocyte reaction (MLR) as a model of alloreactivity was used (22), along with the specific IDO inhibitor 1-dl-methyl-tryptophan (1-MT). This is a competitive, non-toxic IDO inhibitor (23) that has been successfully used to break the immune privilege of the placenta and break tolerance against grafts (9, 13). In addition, the p53 inhibitor pifithrin-α (PFT) was used. PFT acts downstream of p53 and reversibly inhibits p53-dependent transcriptional activation (24).
Finally, in order to distinguish the effect of GCN2 kinase activation on p53 expression from a possible effect of kynurenine, a system lacking IDO was used. Isolated T cells were stimulated with anti-CD2, anti-CD3 and anti-CD28 antibodies in the presence or not of tryptophanol (TRP). This is a competitive inhibitor of the tryptophanyl-tRNA synthetase. By raising the pool of uncharged tRNA, TRP acts as a pharmacologic activator of GCN2 kinase (25).
Methods
Subjects
Blood samples were collected from 10 non-related healthy volunteers (5 men, 35±9 years old). Informed consent was obtained from each individual enrolled in the study and the hospital ethics committee gave its approval to the study protocol.
PBMC and T-cell isolation and culture
PBMCs were isolated from whole blood by Ficoll-Hypaque density gradient centrifugation (Histopaque 1077, Sigma-Aldrich, St Louis, MO, USA) and counted by optical microscopy on a Neubauer hemocytometer. Cell viability was assessed by trypan blue assay (Sigma-Aldrich).
PBMCs were resuspended in RPMI-1640 medium with l-glutamine and 10mM HEPES and supplemented with 10% FCS (Sigma-Aldrich) and an antibiotic–antimycotic solution (Sigma-Aldrich).
In experiments with TRP (Sigma-Aldrich), T cells were isolated from PBMCs. Non-T cells were indirectly magnetically labeled with a cocktail of biotin-conjugated mAbs and were depleted using the Pan-T cell Isolation Kit (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany). Isolated T cells were cultured in the same medium as the PBMCs. All cultures were performed at 37 °C in a humidified atmosphere containing 5% CO2.
Evaluation of 1-MT and PFT cytotoxicity in resting PBMCs
PBMCs were cultured in 96-well plates (1×105/well) for 7 days in the presence or not of 100 μM 1-MT (Sigma-Aldrich) and/or of 30μM PFT (Santa Cruz Biotechnology, Dallas, TX, USA). PFT was re-added to the cell cultures at day 4. The concentration of 1-MT was chosen according to previous experiments (16), whereas the PFT concentration was chosen according to previous studies (24). The cytotoxicity of 1-MT and/or of PFT in resting PMBCs was assessed by a lactate dehydrogenase (LDH)-release assay using the Cytotox Non-Radioactive Cytotoxic Assay kit (Promega Corporation, Madison, WI, USA) according to the protocol provided by the manufacturer. Cytotoxicity was calculated by the equation Cytotoxicity (%) = (LDH in the supernatant/Total LDH) × 100. Experiments were performed in PBMC cultures derived from the blood of 10 individuals. All these experiments were performed in triplicates and the results refer to the mean of the three measurements.
Assessment of l-tryptophan consumption in MLRs
MLRs were performed in 12-well plates for 7days in the presence or not of 100 μM 1-MT. The number of PBMCs for each member of the MLR couple was 5×105, leading to a total of 1×106 PBMCs in each well. After 7 days, supernatants from each MLR were collected. Tryptophan consumption was assessed in the supernatants by means of ELISA (BlueGene Biotech, Shanghai, China). The sensitivity of the above ELISA kit is 0.1ng ml−1. Ten MLRs were performed. All these experiments were performed in triplicates and the results refer to the mean of the three measurements.
Assessment of glucose uptake and aerobic glycolysis in MLRs
MLRs were performed in 12-well plates for 7 days in the presence or not of 100 μM 1-MT and/or of 30 μM PFT. The number of PBMCs from each member of the MLR couple was 5×105, leading to a total of 1×106 PBMCs in each well. After 7 days, supernatants from each MLR were collected. Glucose uptake was assessed by measuring the decrease of glucose concentration in the supernatant, that is the glucose concentration in RPMI-1640 supplemented with 10% FCS minus the final glucose concentration in the supernatants at the end of the experimental procedure. Glucose measurements were obtained using the Element Blood glucose monitor along with the test strips (Element, Infopia, Titusville, FL, USA). Aerobic glycolysis was assessed by measuring the concentration of the end product of this pathway, lactate. The lactate concentration was measured with the Blood Gas Analyzer Czito Medical (Moscow, Russia). Ten MLRs were performed. All these experiments were performed in triplicates and the results refer to the mean of the three measurements.
Assessment of cell proliferation in MLRs
MLRs were performed in 96-well plates for 7 days in the presence or not of 100 μM 1-MT and/or not of 30μM PFT. The PFT was re-added to the cell cultures at day 4. The number of PBMCs from each member of the MLR couple was 5×104, which means that there were 1×105 PBMCs in total in each well. Cultures of resting PMBCs with a population of 1×105 per well were used as controls. At the end of the 7-day period, cell proliferation was assessed via Cell Proliferation ELISA (Roche Diagnostics, Indianapolis, IN, USA) using BrdU labeling and immunoenzymatic detection according to the manufacturer’s protocol. The proliferation index was calculated as the ratio of the optical density (OD) derived from each MLR to the mean of the ODs derived from the control resting PBMC cultures of the two subjects that constituted the specific MLR. Ten MLRs were performed. All these experiments were performed in triplicates and the results refer to the mean of the three measurements.
Isolation of T cells from MLRs and assessment of GCN2 kinase activity, mTORC1 activity and levels of p53, phosphorylated at serine 15 p53, p21 and enzymes involved in glucose metabolism
Ten MLRs were performed in 12-well plates for 7 days. The levels of the GCN2 kinase substrate phosphorylated at serine 51 eIF2α (p-eIF2α) and p53 and phosphorylated at serine 15 p53 (p-p53) were assessed in the presence or not of 100 μM 1-MT. Expression of p21 and of various enzymes involved in glucose metabolism was assessed in the presence or not of 100 μΜ 1-MT and/or not of 30 μM PFT. The PFT was re-added to the cell cultures at day 4. The PBMC population in each MLR sample remained the same as before. At the end of the 7-day period that MLRs lasted, T cells were isolated by negative selection using the Pan-T cell Isolation Kit (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany).
Isolated T cells were counted via optical microscopy on a Neubauer hemocytometer and cell viability was determined by the trypan blue assay. Equal numbers of T cells from each MLR were lysed using the T-PER tissue protein extraction reagent (Pierce Biotechnology) supplemented with protease and phosphatase inhibitors (Sigma-Aldrich and Roche Diagnostics). The protein was quantified via Bradford assay (Sigma-Aldrich) and western blotting was performed.
Equal quantities of protein extracts (50 μg) from each sample were loaded for electrophoresis in SDS-PAGE gels (Invitrogen, Life Technologies). Subsequently proteins were transferred to polyvinylidene difluoride (PVDF) membranes (Invitrogen, Life Technologies). Blots were incubated with the primary antibody for 16h, followed by secondary antibody (anti-rabbit IgG, HRP-linked Antibody, Cell Signaling Technology, Danvers, MA, USA) for 30min. A benchmark pre-stained protein ladder (Invitrogen, Life Technologies) was used as a marker. Bands were visualized by enhanced chemiluminescent detection using the LumiSensor Plus Chemiluminescent HRP Substrate Kit (GenScript, Piscataway, NJ, USA) and analysis was performed using the Image J software (National Institute of Health, Bethesda, MD, USA). For reprobing PVDF blots, the previous primary and secondary antibodies were safely removed via the use of Restore Western Blot Stripping Buffer (Thermo Fisher Scientific Inc., Rochford, IL, USA) according to the manufacturer’s protocol. The PVDF blot was then re-used and western blotting resumed as previously described, using a different primary antibody.
The primary antibodies used in western blotting were specific for p-eIF2α (Cell Signaling Technology, Danvers, MA, USA), p70S6 kinase phosphorylated at threonine 389 (p-p70S6K) (Cell Signaling Technology), p-53 (Cell Signaling Technology), p-53 (Cell Signaling Technology), p21 (Cell Signaling Technology), glucose transporter-1 (GLUT1) (Santa Cruz Biotechnology), glucose-6-phosphate dehydrogenase (G6PD) (Cell Signaling Technology), LDH-A (Cell Signaling Technology), TP53-induced glycolysis and apoptosis regulator (TIGAR) (Santa Cruz Biotechnology), mitochondrial glutaminase (GLS2) (Acris Antibodies, San Diego, CA, USA), synthesis of cytochrome c oxidase 2 (SCO2) (Santa Cruz Biotechnology) and β-actin (Cell Signaling Technology).
Stimulation of immediately isolated from PBMC T cells in the presence or not of TRP
T cells were immediately isolated from PBMCs using the Pan T-cell Isolation Kit (Miltenyi Biotec GmbH). Isolated T cells were counted by optical microscopy on a Neubauer hemocytometer. Cell viability was assessed by the trypan blue assay (Sigma-Aldrich).
T cells were cultured without stimuli or stimulated with anti-CD2-, anti-CD3- and anti-CD28-conjugated beads using the T-cell activation/expansion kit (Miltenyi Biotec GmbH) in a bead–to-cell ratio of 1:2. Stimulated T cells were cultured in the presence or not of TRP at a concentration of 0.25mM, sufficient for 50% reduction in proliferation of T cells stimulated with PMA and ionomycin (3).
Evaluation of TRP toxicity and of its effect on proliferation in T cells immediately after isolation from PBMCs
The toxicity of TRP for stimulated T cells was evaluated with the Cytotox Non-Radioactive Cytotoxic Assay kit (Promega Corporation), while T-cell proliferation was assessed by Cell Proliferation ELISA (Roche Diagnostics). Resting T cells, stimulated T cells or T cells stimulated in the presence of 0.25mM TRP were cultured in 96-well plates (1×105/well) for 72h. Experiments were performed in T cells derived from the blood of 10 individuals. All these experiments were performed in triplicates and the results refer to the mean of the three measurements.
Assessment of the effect of TRP on GCN2 kinase activity and p53 level in T cells immediately after isolation from stimulated PBMCs
Proteins were extracted from isolated T cells that had been stimulated in the presence or absence of 0.25mM TRP; the T cells had been stimulated during culture in 12-well plates (1×106 cells/well) for 12h in order to assess p-eIF2α and p53 expression by means of western blotting. The primary antibodies were anti-p-eIF2α, anti-p53 and anti-β-actin (all from Cell Signaling Technology). Experiments were performed in T cells derived from the blood of 10 individuals.
Statistical analysis
The normality of the evaluated variables was assessed and confirmed by the one-sample Kolmogorov–Smirnov test. For comparison of means between two conditions, a paired t-test was used. For comparison of means among more than two conditions, the sphericity assumption was evaluated by Mauchly’s test and, if violated, degrees of freedom were corrected using Greenhouse–Geisser or Huynh–Feldt estimates of sphericity. Comparison of means was assessed by one-way repeated-measures ANOVA followed by the Bonferroni’s correction test. Results are expressed as mean ± SD and a P < 0.05 was considered statistically significant.
For the western blotting, results were expressed as ODs, so P-values were calculated by comparing the means of OD. Statistical analysis relative to the control OD values was avoided to prevent violation of the prerequisite for normal distribution of the compared variables when applying parametric statistical tests. Results are depicted according to the OD, and the error bars corresponded to 95% CI of difference of means. However, for the reader’s convenience, in the text the results are also expressed after normalization of means for the control group.
Results
1-MT and PFT were almost non-toxic in resting PBMCs, as was TRP in stimulated, isolated T cells
In resting PBMCs, the LDH-release assay revealed a cytotoxicity of 11.17±1.50% after no treatment, 12.00±3.32% after treatment with 1-MT, 11.50±1.65% after treatment with PFT and 12.43±2.57% after treatment with both. No statistically significant difference of means was detected among the above groups (Fig. 1A).
The effect of 1-MT and/or PFT treatment on glucose consumption, lactate production and cell proliferation in MLRs. Ten MLRs were performed in the presence or not of the IDO inhibitor 1-MT and in the presence or not of the p53 inhibitor PFT. Neither 1-MT nor PFT exhibited cytotoxicity for PBMCs (A). In MLRs, both 1-MT and PFT increased glucose consumption significantly (B). Regarding lactate production, 1-MT increased it, whereas PFT left it unaffected (C). Both 1-MT and PFT increased cell proliferation (D). Error bars correspond to 95% CI of difference of means.
In stimulated isolated T cells, cytotoxicity was 11.00±2.05%. TRP treatment augmented cytotoxicity, but only slightly to 13.50±1.41% (P = 0.008) (Fig. 2A).
The effect of TRP on cell proliferation, GCN2 kinase activity and p53 levels in stimulated isolated T cells. The T cells were isolated from PBMCs and stimulated with anti-CD2, anti-CD3 and anti-CD28 antibodies in the presence or not of the GCN2 kinase activator TRP. The TRP was almost non-toxic for stimulated T cells (A) but decreased T-cell proliferation significantly (B). GCN2 kinase activity, assessed by phosphorylation of its substrate eIF2α at serine 51, and p53 were evaluated. The western blotting lanes correspond to two representative experiments of the ten performed (C). TRP significantly increased p-eIF2α (D) and p53 (E). Error bars correspond to 95% CI of difference of means.
1-MT decreased l-tryptophan consumption, reduced GCN2 kinase activity, did not alter mTORC1 activity and decreased p53 and p-p53 in MLR-derived T cells
In MLRs, 1-MT decreased l-tryptophan consumption significantly. The l-tryptophan concentration in the supernatants of MLRs was 2.83±0.43ng mL−1 in the absence of 1-MT, whereas in the presence of 1-MT, the l-tryptophan concentration was 6.34±0.73ng mL−1 (P < 0.001) (Fig. 3A).
The effect of 1-MT treatment of MLRs on l-tryptophan concentrations and on the GCN2 kinase activity, mTORC1 activity, p53 and p-p53 levels in alloreactive T cells. Ten MLRs were performed in the presence or not of the IDO inhibitor 1-MT. In the absence of 1-MT, l-tryptophan in the supernatants of MLRs decreased (A). In T cells derived from the MLRs, GCN2 kinase activity, assessed by phosphorylation of its substrate eIF2α at serine 51, mTORC1 activity, assessed by phosphorylation of its substrate p70S6K at threonine 389, p53 and phosphorylated p53 at serine 15 (p-53) were evaluated. The depicted western blotting lanes correspond to two MLRs out of ten that were performed (B). In the absence of 1-MT, p-eIF2α (C), p53 (E) and p-p53 (F) were all significantly increased, whereas p-p70S6K remained unaffected (D). Error bars correspond to 95% CI of difference of means.
In MLR-derived T cells, 1-MT reduced the GCN2 kinase activity significantly, as assessed by phosphorylation of its substrate eIF2α. OD analysis of the western blotting bands revealed that p-eIF2α was 4.74±3.42-fold higher in MLR-derived T cells in the absence of 1-MT (P < 0.001) (Fig. 3B and C).
However, in MLR-derived T cells, mTORC1 activity remained unaffected by 1-MT, as assessed by phosphorylation of its substrate p70S6K. The level of p-p70S6K was 1.06±0.15-fold higher in MLR-derived T cells in the presence of 1-MT (P = 0.593) (Fig. 3B and D).
1-MT decreased p53 expression and the p-p53 level significantly. The expression of p53 was 1.64±0.40-fold higher in MLR-derived T cells in the absence of 1-MT (P < 0.001) (Fig. 3B and E). The level of p-p53 was 2.41±0.88-fold higher in MLR-derived T cells in the absence of 1-MT (P < 0.001) (Fig. 3B and F).
TRP decreased cell proliferation, whereas it enhanced GCN2 kinase activity and p53 levels in stimulated isolated T cells
In T cells isolated from PBMCs, stimulation with anti-CD2, anti-CD3 and anti-CD28 antibodies induced proliferation. In this case the proliferation index was 4.93±0.96 (P < 0.001). However, TRP decreased the proliferation index of stimulated isolated T cells significantly, to 2.92±0.51 (P < 0.001) (Fig. 2B).
In stimulated isolated T cells, TRP significantly enhanced the GCN2 kinase activity, as assessed by the phosphorylation of its substrate eIF2α. During TRP treatment, the level of p-eIF2α was 8.71±3.75-fold higher (P < 0.001) (Fig. 2C and D).
Likewise, in stimulated isolated T cells, TRP significantly enhanced p53 levels. After TRP treatment, the p53 level was 6.08±0.6-fold higher (P = 0.008) (Fig. 2C and E).
In MLRs, 1-MT increased glucose consumption, lactate production and cell proliferation; similarly, PFT increased glucose consumption and cell proliferation but did not affect lactate production
Both 1-MT and PFT increased glucose consumption in MLRs. In untreated MLRs, glucose consumption was 31.33±20.47mg dl−1. 1-MT significantly increased glucose consumption to 96.83±23.74mg dl−1 (P < 0.001). Similarly, PFT significantly increased glucose consumption to 112.50±27.64mg dl−1 (P < 0.001). Interestingly, the combined treatment of MLRs with 1-MT and PFT increased glucose consumption further to 134.33±30.85mg dl−1, a value significantly higher than in 1-MT-treated MLRs (P = 0.005) or in PFT-treated MLRs (P = 0.001) (Fig. 1B).
Treatment of MLRs with 1-MT significantly increased lactate production from 4.07±0.42 mmol l−1 to 5.27±0.86 mmol l−1 (P = 0.002). On the contrary, PFT treatment left lactate production unaffected at 4.08±0.50 mmol l−1 (P = 1.0). The combined treatment of MLRs with 1-MT and PFT increased lactate production to 5.53±0.70 mmol l−1 (P < 0.001) but not to a higher extent than the effect exerted by 1-MT alone (P = 0.486) (Fig. 1C).
Both 1-MT and PFT increased cell proliferation in MLRs. After treatment with 1-MT, the proliferation index increased from 1.95±0.28 to 2.56±0.31 (P < 0.001) and after treatment with PFT to 2.36±023 (P < 0.001). Combined treatment with 1-MT and PFT also significantly enhanced the proliferation index to 2.64±0.23 (P < 0.001) but not more than each separate treatment (P = 1.0 and P = 0.99 when compared with the 1-MT- and PFT-treated groups respectively) (Fig. 1D).
In T cells derived from 1-MT-treated MLRs, GLUT1, LDH-A and GLS2 were increased, TIGAR and SCO2 were decreased, whereas G6PD remained stable; PFT also increased GLUT1, and decreased TIGAR and SCO2 but left LDH-A, GLS2 and G6PD unaffected
Compared with untreated MLRs, in T cells derived from 1-MT-treated MLRs, GLUT1 expression was 3.18±1.45-fold higher (P < 0.001). In T cells from PFT-treated MLRs, GLUT1 expression was 3.71±1.78-fold higher (P < 0.001). In the presence of both 1-MT and PFT, GLUT1 expression was 3.80±2.12-fold higher (P < 0.001). Comparing the 1-MT-treated group or the PFT-treated group with the group that received treatment with both inhibitors, no further increase in GLUT1 expression was noticed (P = 0.262 and P = 1.0 respectively) (Figs 4 and 5A).
Western blot depicting the effect of 1-MT and or PFT treatment of MLRs on GLUT1, G6PD, TIGAR, LDH-A, GLS2 and SCO2 in alloreactive T cells. Ten MLRs were performed in the presence or not of the IDO inhibitor 1-MT and the presence or not of the p53 inhibitor PFT. Then T cells were isolated and western blot performed. The western blotting lanes correspond to three representative experiments of the ten performed. A, B and C correspond to the different MLRs from which the T cells were isolated. 1 corresponds to no treatment of the MLRs, 2 to treatment with 1-MT, 3 to treatment with PFT and 4 to treatment with both 1-MT and PFT.
The effect of 1-MT and/or PFT treatment of MLRs on GLUT1, G6PD, TIGAR, LDH-A, GLS2 and SCO2 in alloreactive T cells. Ten MLRs were performed in the presence or not of the IDO inhibitor 1-MT and the presence or not of the p53 inhibitor PFT. Then T cells were isolated and western blot performed. Both 1-MT and PFT significantly increased GLUT1 expression (A). Neither 1-MT nor PFT affected G6PD expression (B). Both 1-MT and PFT significantly decreased TIGAR expression (C). Regarding LDH-A, 1-MT increased its expression, whereas PFT left it unaffected (D). Similarly, 1-MT significantly increased GLS2 expression, whereas PFT did not affect it (E). Both 1-MT and PFT significantly decreased SCO2 expression (F). Error bars correspond to 95% CI of difference of means.
Neither 1-MT nor PFT treatment of MLRs significantly altered G6PD expression in T cells. Compared with T cells derived from untreated MLRs, G6PD expression was altered only by a factor of 1.01±0.33 in T cells derived from 1-MT-treated MLRs (P = 1.0), by a factor of 1.36±0.58 in T cells derived from PFT-treated MLRs (P = 1.0) and by a factor of 1.14±0.46 in T cells derived from MLRs treated with both 1-MT and PFT (P = 1.0) (Figs 4 and 5B).
Under the same conditions, TIGAR expression decreased by a factor of 0.61±0.06 (P < 0.001) in T cells derived from 1-MT-treated MLRs as compared with untreated MLRs. In T cells from PFT-treated MLRs, TIGAR decreased by a factor of 0.34±0.21 (P < 0.001), whereas in T cells derived from MLRs treated with both inhibitors, TIGAR protein levels decreased by a factor of 0.27±0.22 (P < 0.001). Compared with the 1-MT-treated group, the groups treated with PFT or with a combination of 1-MT and PFT exhibited a more-intensive decrease of TIGAR expression (P < 0.001 and P = 0.007 respectively). No difference was detected regarding TIGAR expression between PFT-treated cells and cells treated with both 1-MT and PFT (P = 0.453) (Figs 4 and 5C).
Compared with untreated MLRs, in T cells derived from 1-MT-treated MLRs, LDH-A expression increased significantly, by a factor of 4.55±5.1 (P < 0.001). In T cells from PFT-treated MLRs, LDH-A expression increased by a factor of 1.51±1.06, and it was not significantly different (P = 1.0). In T cells derived from MLRs treated with both 1-MT and PFT, LDH-A expression increased significantly, by a factor of 4.49±5.59 (P = 0.002). Compared with the 1-MT-treated group, the group treated with PFT exhibited decreased expression of LDH-A (P = 0.003), whereas no such a difference was detected in the group treated with both 1-MT and PFT (P = 0.298) (Figs 4 and 5D).
Regarding GLS2 expression as assessed by comparing the untreated group with the 1-MT-treated group, it was significantly increased in the latter by a factor of 2.94±2.17 (P < 0.001). In the PFT-treated group, GLS2 expression experienced an insignificant decrease, by a factor of 0.95±0.33 (P = 0.452). Treatment with both 1-MT and PFT demonstrated a notable increase in GLS2 expression, by a factor of 2.03±1.84 (P = 0.032). Compared with the 1-MT-treated group, the group treated with PFT exhibited decreased expression of GLS2 (P < 0.001). Compared with the 1-MT-treated group, in the group treated with both 1-MT and PFT, a statistically significant decrease of GLS2 expression was detected (P = 0.05) (Figs 4 and 5E).
A significant decrease was also observed in SCO2 expression in 1-MT-treated MLRs, by a factor of 0.42±0.26 (P = 0.002), in contrast to untreated alloreactive cells. In the PFT-treated group, SCO2 expression decreased by a factor of 0.32±0.19 (P = 0.001). Finally, in the group treated with both 1-MT and PFT, SCO2 expression decreased by a factor of 0.24±0.14 (P < 0.001). Compared with the 1-MT-treated group, the groups treated with PFT or with the combination of 1-MT and PFT exhibited a more-intensive decrease of SCO2 expression (P = 0.007 and P = 0.05 respectively). No difference was detected regarding SCO2 expression between T cells derived from MLRs treated with PFT and from MLRs treated with both 1-MT and PFT (P = 0.937) (Figs 4 and 5F).
In T cells derived from MLRs treated with either 1-MT or PFT, expression of p21 was decreased
Treatments of MLRs with either 1-MT or PFT significantly reduced p21 expression in T cells. In T cells derived from 1-MT-treated MLRs, PFT-treated MLRs and MLRs treated with both 1-MT and PFT, p21 expression in every instance was roughly two-fold higher than in T cells derived from untreated MLRs (1-MT: 1.98±0.47, P < 0.001; PFT: 2.47±0.62, P < 0.001 and 1-MT and PFT: 2.03±0.93, P < 0.01). Compared with MLRs treated with 1-MT or PFT alone, the combined treatment with both 1-MT and PFT did not significantly alter p21 expression in T cells (P = 1.0 in both cases) (Fig. 6A and B).
The effect of 1-MT and/or PFT treatment of MLRs on p21 expression in alloreactive T cells. Ten MLRs were performed in the presence or not of the IDO inhibitor 1-MT and the presence or not of the p53 inhibitor PFT. Then T cells were isolated and western blot performed. The western blotting lane corresponds to three representative experiments of the ten performed. A, B and C correspond to the different MLRs from which the T cells were isolated. 1 corresponds to no treatment of the MLRs, 2 to treatment with 1-MT, 3 to treatment with PFT and 4 to treatment with both 1-MT and PFT (A). Both 1-MT and PFT significantly decreased p21 expression in alloreactive T cells (B).
Discussion
IDO plays a significant role in immune system homeostasis and therefore revealing the exact mechanism of its action in suppressing T cells is of great importance (1, 2). Inhibition of cell proliferation is a well-defined mechanism of IDO action in T cells, and the same molecule inhibits aerobic glycolysis in alloreactive T cells as shown in a previous study conducted by our team (16). Since the tumor suppressor p53 shares the above properties of IDO (19–21), the role of IDO in increasing p53 level in isolated T cells from MLRs was evaluated first.
For this purpose the IDO inhibitor 1-MT was used (9, 13). Indeed IDO increased l-tryptophan consumption in the supernatants of MLRs. This was sensed in MLR-derived T cells by GCN2 kinase, which was shown to be activated because increased levels of its phosphorylated substrate p-eIF2α were detected. In parallel, the p53 levels also increased. The levels of p-p53 were also elevated (Fig. 3). Phosphorylation of p53 at serine 15 impairs the ability of the mouse double minute 2 (MDM2) homolog to bind p53, promoting both the accumulation and activation of p53 (26).
Regarding the other pathway that senses amino acid deprivation, T cells fail to proliferate in response to antigen once tryptophan or other essential amino acids become sparse, associated with reduced mTORC1 signaling (4, 5). However, in the model of human MLR used in this study, mTORC1 activity, assessed by the phosphorylation of its substrate p70S6K, was not affected by IDO-induced tryptophan depletion. This is in accordance with an elegant study demonstrating that GCN2 kinase mediates proliferation arrest of T cells in response to IDO, a result that could not be recapitulated with mTORC1 inhibitors, such as rapamycin (3).
However, besides GCN2 kinase activation, kynurenine, the first breakdown product in the IDO-dependent tryptophan degradation pathway, is able to affect T cells by activating the arylhydrocarbon receptor (AHR) (27, 28). In order to elucidate whether GCN2 activation alone is adequate for increasing p53 in T cells, we used a kynurenine-free system that lacked APCs, and the GCN2 kinase activator TRP (25). Indeed, in activating isolated T cells, TRP activated the GCN2 kinase and increased p53 levels, indicating that GCN2 activation alone is adequate for p53 up-regulation in T cells. Interestingly, TRP also inhibited T-cell proliferation (Fig. 2).
The fact that 1-MT treatment of MLRs reduces the level of p53 in T cells indicates that IDO increases p53 levels. Moreover IDO, through tryptophan depletion, activates GCN2 kinase. The effect of IDO in increasing p53 is also confirmed by the elevated levels of p53 in T cells treated with the GCN2 kinase activator TRP.
Next we evaluated which effect of IDO may be mediated by p53. For this purpose, in parallel to 1-MT as an IDO inhibitor, PFT was used in order to inhibit p53 (24). If the action of PFT in MLRs, where p53 was found to be increased due to IDO, recapitulates the effects of 1-MT, which inhibits IDO and reduces p53, then it could be concluded that p53 may contribute to the effects of IDO.
Firstly, we examined the effect of IDO on glucose consumption, lactate production and cell proliferation. Both IDO and p53 reduced glucose consumption and decreased proliferation. However, only IDO and not p53 decreased lactate production (Fig. 1).
Subsequently, we evaluated the effect of IDO on various key enzymes of glucose metabolism and examined which of these effects are also induced by p53 in MLR-derived T cells (Fig. 4). Regarding the main glucose transporter of T cells (GLUT1), both IDO and p53 inhibited its expression (Fig. 5). This is in accordance with the results for glucose consumption in MLRs (Fig. 1). Albeit in other cell types, p53 is known to repress the transcriptional activity of the GLUT gene (29). In T cells, IDO-induced down-regulation of GLUT1 is accompanied by a decrease in the expression of this transporter on the cell surface (16).
A key mechanism by which p53 inhibits glycolysis is the up-regulation of TIGAR, which inhibits the enzyme 6-phosphofructokinase 2/fructose-2,6-biphosphatase (PFK2) that dephosphorylates the metabolite fructose-2,6-biphosphate (F-2,6-P2). Since F-2,6-P2 is a potent activator of 6-phosphofructokinase 1 (PFK1), which converts fructose-6-phosphate (F-6-P) to fructose-1,6-biphosphate (F-1,6-P2) in the third step of glycolysis, a p53-dependent decrease in F-2,6-P2 by TIGAR suppresses glycolysis (30). In MLR-derived T cells, both IDO and p53 induced TIGAR expression and thus it is likely to inhibit glycolysis (Fig. 5).
LDH-A is responsible for lactate production by pyruvate in T cells. In MLR-derived T cells, IDO but not p53 decreased LDH-A levels (Fig. 5). This is in agreement with the results for lactate production in MLRs (Fig. 1). It should be noted that in 1-MT-treated, rapidly proliferating T cells, the increased expression of LDH-A and the production of lactate could be protective. In the context of increased intracellular glucose due to increased GLUT1 expression, a massive entry of the produced pyruvate into the Krebs’s cycle and the consequent increased availability of the electron donors reduced NADH (nicotinamide adenine dinucleotide) and reduced FADH2 (flavin adenine dinucleotide) for oxidative phosphorylation could lead to increased oxidative stress (31).
Regarding the other pathway of pyruvate metabolism, that is the Krebs’s cycle, the key enzyme that regulates its entry into the mitochondrion and Krebs’s cycle is pyruvate dehydrogenase (PDH). This is inhibited by reversible phosphorylation by the mitochondrial PDH kinase (PDK). PDH phosphate phosphatase (PDP) reverses that phosphorylation and re-activates PDH (32). However, we have already shown that IDO does not affect PDH expression or p-PDH levels in alloreactive T cells (16). Thus it seems that the little pyruvate formed due to IDO-induced inhibition of glucose uptake and glycolysis is preferentially diverted to the Krebs’s cycle instead of being converted to lactate. This assures a minimum of ATP production under the starvation conditions, and it is IDO dependent but p53 independent.
The level of the first enzyme of the pentose-phosphate pathway, G6PD, remained unaffected by both IDO and p53 (Fig. 5). However, p53 is known to bind to G6PD, inactivating it (33). Thus it is likely that IDO-induced up-regulation of p53 by inactivating G6PD also inhibits nucleotide synthesis, which is required for rapidly proliferating cells, such as stimulated T cells. The down-regulation of p53 in 1-MT-treated rapid proliferating T cells is expected to release G6PD activity which, except for promoting the synthesis of nucleotides, could be cytoprotective as well. The pentose-phosphate pathway produces NADPH, which is required for the generation of the potent antioxidant reduced glutathione (GSH) (34).
Another key aspect to be considered is the increased glutaminolysis observed during T-cell activation (17). In MLR-derived T cells, IDO but not p53 down-regulated the mitochondrial GLS2 (Fig. 5). However, albeit in other cell types, studies have shown that p53 up-regulates GLS2 (35, 36). IDO-induced down-regulation of GLS2 prevents the supply of the Krebs’s cycle with an additional substrate, starving the T cells more. In addition by suppressing glutaminolysis, IDO could suppress the synthesis of lipids and amino acids, which takes place by the reverse of two Krebs’s cycle reactions that convert α-ketoglutarate (α-KG) to citrate into the mitochondria (37–40). Once more, IDO blocks the synthesis of biomolecules required for cell proliferation.
SCO2 expression is known to be up-regulated by p53 and is required for the assembly of cytochrome c oxidase II subunit into the cytochrome oxidase II complex of the mitochondrial electron transport chain (41). In MLR-derived T cells, both IDO and p53 increased SCO2 expression (Fig. 5). It is likely that under the described starvation conditions the IDO-induced up-regulation of oxidative phosphorylation chain components ensures that the little available glucose will be used in the most energetically effective way. The tumor suppressor p53 serves as a switch, which induces glucose starvation and reduces glycolysis, but on the other hand enhances oxidative phosphorylation in order to ensure a minimal ATP production and possibly prevent a total collapse of the cell. All the described IDO-induced and p53-induced alterations of glucose metabolism we found are depicted in Fig. 7.
Effects of IDO and p53 on certain enzymes of glucose metabolism in alloreactive T cells. IDO, through l-tryptophan consumption, activates GCN2 kinase and increases p53 levels in alloreactive T cells. Both IDO and p53 down-regulate GLUT1 and decrease the influx of glucose, but they up-regulate TIGAR, which by inhibiting PFK1 activity decelerates aerobic glycolysis. The expression of the rate-limiting enzyme of the pentose-phosphate pathway G6PD remained unaffected. Through the decline of LDH-A expression, IDO but not p53 suppresses lactate production. Also through the reduction of GLS2 expression, IDO but not p53 prevents glutamine from entering the Krebs’s cycle. Finally, IDO induces the expression of SCO2, a component of the oxidative phosphorylation chain. Overall, it is likely that IDO starves T cells of glucose and reduces glycolysis. By suppressing LDH-A expression, IDO ensures that the decreased amount of the produced pyruvate will enter into the more energetically effective Krebs’s cycle, instead of being converted to lactate. Glutaminolysis is also suppressed by IDO. Under conditions of starvation, the IDO-induced up-regulation of oxidative phosphorylation chain components ensures that the little available glucose will be used by the most energetically effective way. p53 may serve as a switch, which induces glucose starvation and reduces glycolysis, but on the other hand enhances oxidative phosphorylation in order to ensure a minimal ATP production and possibly prevent a total collapse of the cell. However, as it is discussed in the text, these IDO-induced alterations, in addition to diminished ATP production, affect pathways involved in the biosynthesis of new biomolecules, another prerequisite for rapid proliferation of activated T cells. Abbreviations: Ac-CoA, acetyl-CoA; α-KG, α-ketoglutarate; OAA, oxalocetate; Gln, glutamine; Glu, glutamate; R5P, ribose 5-phosphate.
Finally, in MLR-derived T cells, both IDO and p53 induced the up-regulation of p21 expression (Fig. 6). This observation is in agreement with the fact that both IDO and p53 suppressed cell proliferation in MLRs (Fig. 1). It is known that p53 directly inhibits cell proliferation by inducing G1-phase cell-cycle arrest through activation of transcription of the cyclin-dependent kinase inhibitor p21 (19). The p21-mediated suppression of cell proliferation by IDO lowers cell requirements for energy and new biomolecules and simultaneously IDO-induced alterations in the cell metabolism machinery are affected in the same way.
In conclusion, in alloreactive T cells, IDO increases p53 levels, and both IDO and p53 inhibit cell proliferation, glucose consumption and glycolysis. Lactate production and glutaminolysis are also suppressed by IDO, but not by p53. It remains to be elucidated if the described effects of IDO are p53 mediated.
Acknowledgements
The project and the experimental design were conceptualized by T.E. after significant discussions with I.S. The experiments were conducted by T.E and G.P. T.E., G.P., G.A., A.S. and V.L. analyzed the data. T.E., G.P., G.A., A.S., V.L. and I.S. contributed equally towards writing the manuscript.
Conflict of interest statement: The authors declared no conflicts of interest. The authors alone are responsible for the content and writing of the paper.
References
