PD-1 alters T-cell metabolic reprogramming by inhibiting glycolysis and promoting lipolysis and fatty acid oxidation

Abstract

During activation, T cells undergo metabolic reprogramming, which imprints distinct functional fates. We determined that on PD-1 ligation, activated T cells are unable to engage in glycolysis or amino acid metabolism but have an increased rate of fatty acid β-oxidation (FAO). PD-1 promotes FAO of endogenous lipids by increasing expression of CPT1A, and inducing lipolysis as indicated by elevation of the lipase ATGL, the lipolysis marker glycerol and release of fatty acids. Conversely, CTLA-4 inhibits glycolysis without augmenting FAO, suggesting that CTLA-4 sustains the metabolic profile of non-activated cells. Because T cells utilize glycolysis during differentiation to effectors, our findings reveal a metabolic mechanism responsible for PD-1-mediated blockade of T-effector cell differentiation. The enhancement of FAO provides a mechanistic explanation for the longevity of T cells receiving PD-1 signals in patients with chronic infections and cancer, and for their capacity to be reinvigorated by PD-1 blockade.

Figure 1. PD-1 inhibits transport and utilization of glucose during T-cell activation.

CD4⁺ primary human T cells were either left unstimulated (UT) or were incubated with magnetic beads conjugated with αCD3/αCD28/IgG (T cells co-stimulated (T_CC)) or magnetic beads conjugated with αCD3/αCD28/PD-L1-Ig fusion protein (T cells co-stimulated+PD-1 (T_CC+PD1)). (a) IFN-γ production was assessed by ELISA (*P<0.05 T_CC+PD1 versus T_CC; n=3 experiments; Student’s t-test). (b,c) Expression of Glut1 after culture under the indicated conditions and time intervals was examined by flow cytometry, and glucose uptake was examined by 2-{¹⁴C(U)}-deoxy-D-glucose incorporation. At each time point glucose uptake in T_CC+PD1-stimulated cells was compared with T_CC-stimulated cells (*P<0.05, T_CC+PD1 versus T_CC; n=3 experiments; Student’s t-test). (d) Analysis of key metabolites involved in glycolysis was performed in cells and culture supernatants as described in Methods. The amounts of the indicated metabolites in unstimulated, T_CC and T_CC+PD1 cells were plotted in whisker boxes. The lower and upper sides of the box indicate the first and third quartile, respectively. The horizontal line inside the box indicates the median value, whereas the lower and upper bars indicate the minimum and maximum of distribution, respectively; (+) mean value; (o) extreme data point. Results of five measurements generated from five independent experiments (*P<0.05, T_CC+PD1 versus T_CC, Welch’s t-test; ^♦P<0.05, T_CC versus UT and T_CC+PD1 versus UT, analysis of variance). (e) HK2 mRNA was analysed by real-time quantitative PCR and relative expression of mRNA of each time point and culture condition over the levels expressed in unstimulated cells (defined as 1) is shown. Analysis was performed over prolonged time of culture (120 h) to investigate whether quantitative difference versus altered kinetics of HK2 expression was induced by PD-1 (*P<0.05, T_CC+PD1 versus T_CC, Student’s t-test; n=3 experiments).

Figure 2. PD-1 inhibits transport and catabolism of glutamine and branched-chain amino acids during T-cell activation.

(a) SNAT1 and SNAT2 mRNA was analysed by real-time quantitative PCR and relative expression of mRNA of each time point and culture condition over the levels expressed in unstimulated cells (defined as 1) is shown (*P<0.05, T_CC+PD1 versus T_CC, Student’s t-test; n=3 experiments). (b) The amounts of glutamine and the glutaminolysis metabolite, glutamate, valine and the metabolic product of valine catabolism, 4-methyl-2-oxopentanoate, were analysed in UT, T_CC and T_CC+PD1 cells and their culture supernatants were plotted in whisker boxes (*P<0.05, T_CC+PD1 versus T_CC, Welch’s t-test; ^♦P<0.05, T_CC+PD1 versus UT and T_CC versus UT, analysis of variance). Results of five replicate samples generated from five independent experiments. (c) Mitochondrial membrane potential (ΔΨ_m) was assessed after the indicated time intervals of culture by staining with the potentiometric dye TMRE and analysis by flow cytometry; mean fluorescence intensity (MFI) of each sample is shown in the tables. Results are representative of three experiments.

Figure 3. PD-1 induces upregulation of CPT1A and fatty acid β-oxidation.

(a) CPT1A mRNA was analysed by real-time quantitative PCR and relative expression of mRNA of each time point and culture condition over the levels expressed in unstimulated cells (defined as 1) is shown (*P<0.05, T_CC+PD1 versus T_CC, Student’s t-test; ^♦P<0.05, T_CC+PD1 versus UT and T_CC versus UT, analysis of variance (ANOVA); n=3 experiments). (b) CD4⁺ primary human T cells were cultured under the indicated conditions. Cell lysates were prepared at the indicated time points and expression of CPT1A and β-actin were assessed by SDS–PAGE and immunoblot. Results are representative of three experiments. (c) Fatty acid β-oxidation rate after culture under the indicated conditions for various time intervals was examined. Values of T_CC and T_CC+PD1 cells were compared with unstimulated (UT) cells (^♦P<0.05, ANOVA; n=3) and values of T_CC+PD1 were compared with T_CC cells (*P<0.05, Student’s t-test; n=3).

Figure 4. PD-1 induces lipolysis and utilization of endogenous fatty acids for β-oxidation.

(a) The amounts of n3DPA;22:5n3 and glycerol-3-phosphate (G3P) in the cells and the amounts of DHA;22:6n3 and 3-hyroxybutyrate (BHBA) in the culture supernatants of the indicated culture conditions were analysed and values were plotted in whisker boxes. Values of T_CC and T_CC+PD1 cells were compared with unstimulated (UT) cells (^♦P<0.05, analysis of variance (ANOVA); n=5) and values of T_CC+PD1 were compared with T_CC cells (*P<0.05, Welch t-test; n=5). (b) Purified human T cells were cultured under the indicated conditions. Cell lysates were prepared at the indicated time points and expression of FASN, ATGL and β-actin were assessed by SDS–PAGE and immunoblot. Results are representative of three experiments. (c) ATGL mRNA was analysed by real-time quantitative PCR and relative expression of mRNA of each time point and culture condition over the levels expressed in unstimulated (UT) cells (defined as 1) is shown. Expression levels of ATGL mRNA in T_CC+PD1-stimulated cells were compared with TCC-stimulated cells (*P<0.05, Student’s t-test) and expression levels in T_CC and T_CC+PD1-stimulated cells were compared with unstimulated (UT) cells (^♦P<0.05, ANOVA). Results are mean±s.e.m. of three experiments. (d) T cells were cultured with [9,10-³H]palmitate for 48 h, over which time the radiolabelled fatty acid was used to generate cellular lipids. Subsequently, cells were washed to remove unincorporated label and were cultured with beads conjugated with aCD3/CD28/IgG or with beads conjugated with aCD3/CD28/PD-L1-Ig and the amount of β-oxidation of labelled fatty acids generated from endogenous lipid was determined each day. Values of T_CC and T_CC+PD1 cells were compared with unstimulated (UT) cells (^♦P<0.05, ANOVA; n=3) and values of T_CC+PD1 were compared with T_CC cells (*P<0.05, Student’s t-test; n=3).

Figure 5. T cells receiving PD-1 signals have substantial mitochondrial spare respiratory capacity.

(a,b) CD4⁺ primary human T cells were cultured under the indicated conditions and 72 h of culture extracellular acidification rate (ECAR) and oxygen consumption rates (OCR) were assessed. (c) OCR/ECAR ratio was also measured. (d) OCR for each group was measured in real time under basal conditions and after addition of the indicated mitochondrial inhibitors. (e) Spare respiratory capacity (SRC) indicated by the difference of maximum OCR over basal OCR (arrows in d) calculated as percentage of basal OCR was also determined. Values of T_CC and T_CC+PD1 cells were compared with unstimulated (UT) cells (^♦P<0.05, analysis of variance; n=4); values of T_CC+PD1 cells were compared with T_CC cells (*P<0.05, Student’s t-test; n=4).

Figure 6. Inhibition of PI3K/Akt and MEK/Erk pathways is involved in the metabolic reprogramming of activated T cells.

(a) CD4⁺ primary human T cells were cultured with tosyl-activated magnetic beads conjugated with aCD3/aCD28/IgG (T_CC) in the presence of either LY294002 (LY, 10 μM), UO126 (UO, 10 μM) or their combination. Cell lysates were prepared at the indicated time points and expression of CPT1A and β-actin was assessed by SDS–PAGE and immunoblot. Results are representative of three experiments. (b–d) In parallel experiments, rate of fatty acid β-oxidation was examined. Values of T_CC+inhibitor cultures were compared with T_CC (*P<0.05, Student’s t-test; n=3) and values of T_CC and T_CC+inhibitor cultures were compared with unstimulated (UT) (^♦P<0.05, analysis of variance (ANOVA); n=3). (e) At 72 h of culture under the indicated conditions, spare respiratory capacity (SRC) was determined. Values of T_CC and T_CC+inhibitor cells were compared with unstimulated (UT) cells (^♦P<0.05, ANOVA; n=3) and values in T_CC+inhibitor cells were compared with T_CC (*P<0.05, Student’s t-test; n=3). (f) T cells were cultured with tosyl-activated magnetic beads conjugated with aCD3/aCD28/PD-L1-Ig or with tosylatcivated magnetic beads conjugated with aCD3/aCD28/IgG in the presence of either LY294002 (LY, 10 μM), UO126 (UO, 10 μM) or their combination. Cell lysates were prepared at the indicated time points and expression of ATGL and β-actin was assessed by SDS–PAGE and immunoblot.

Figure 7. CTLA-4 inhibits glycolytic reprogramming without increasing the rate of fatty acid β-oxidation.

CD4⁺ primary human T cells were either left unstimulated (UT) or were incubated with magnetic beads conjugated with αCD3/αCD28/IgG (T cells co-stimulated (T_CC)) or magnetic beads conjugated with αCD3/αCD28/αCTLA-4 mAbs (T cells co-stimulated+ CTLA-4 (T_CC+CTLA4)). (a) Expression of Glut1 after culture under the indicated conditions and time intervals was examined by flow cytometry. Results are representative of three experiments. (b–f) mRNA for HK2, SNAT1, SNAT2, CPT1A and ATGL was analysed by real-time quantitative PCR and relative expression of mRNA of each time point and culture condition over the levels expressed in unstimulated cells (defined as 1) is shown. (g) Fatty acid β-oxidation rate after culture under the indicated conditions for various time intervals was examined. (h,i) Analysis of lactate (h) and 3-hyroxybutyrate (i), end metabolites of glycolysis and fatty acid β-oxidation, respectively, was performed in culture supernatants. (For the studies shown in all panels: *P<0.05, T_CC+CTLA4 versus T_CC, Student’s t-test; ^♦P<0.05, T_CC versus UT and T_CC+CTLA4 versus UT, analysis of variance; n=3 experiments).

Figure 8. PD-1 alters the metabolic programme of activated T cells from glycolysis to FAO.

CD4⁺ human T cells were pre-activated with anti-CD3-and-anti-CD28 mAbs for 4 h or for 24 h and subsequently were collected, were left to rest for 3 h and re-cultured with tosyl-activated magnetic beads conjugated with αCD3/αCD28/PD-L1-Ig (Tpr4hr→PD-1 and Tpr24hr→PD-1). In the same experiment CD4⁺ T cells without pre-activation were stimulated with tosyl-activated magnetic beads conjugated with αCD3/αCD28/IgG (T_CC) and used as positive control for optimal stimulation without PD-1. (a) Expression of Glut1 after culture under the indicated conditions and time intervals was examined by flow cytometry. (b,c) mRNA for HK2 and CPT1A was analysed by real-time quantitative PCR and relative expression of mRNA of each time point and culture condition over the levels expressed in unstimulated cells (defined as 1) is shown. (d) Fatty acid β-oxidation rate after culture under the indicated conditions for various time intervals was examined. (e,f) Analysis of lactate (e) and 3-hydroxybutyrate (f), end metabolites of glycolysis and fatty acid β-oxidation, respectively, was performed in culture supernatants (for the studies shown in all panels: *P<0.05, Tpr4hr→PD-1 versus T_CC or Tpr4hr→PD-1 versus T_CC, Student’s t-test; ^♦P<0.05, T_CC versus UT, Tpr4hr→PD-1 versus UT and Tpr4hr→PD-1 versus UT, analysis of variance; n=3 experiments).