Distinct energy requirements for human memory CD4 T-cell homeostatic functions

Abstract

Differentiation and activation of CD4 memory T cells (T_mem cells) require energy from different sources, but little is known about energy sources for maintenance and surveillance activities of unactivated T_mem cells. Mitochondrial fatty acid oxidation (FAO) in human unactivated CD4 T_mem cells was significantly enhanced by inhibition of glycolysis, with respective means of 1.7- and 4.5-fold for subjects <45 yr and >65 yr, and by stimulation of AMP-activated protein kinase, with respective means of 1.3- and 5.2-fold. However, CCL19 and sphingosine 1-phosphate (S1P), which control homeostatic lymphoid trafficking of unactivated T_mem cells, altered FAO and glycolysis only minimally or not at all. Inhibition of CD4 T_mem-cell basal FAO, but not basal glycolysis, significantly suppressed CCL19- and S1P-mediated adherence to collagen by >50 and 20%, respectively, and chemotaxis by >20 and 50%. Apoptosis of unactivated T_mem cells induced by IL-2 deprivation or CCL19 was increased significantly by >150 and 70%, respectively, with inhibition of FAO and by >110 and 30% with inhibition of glycolysis. Anti-TCR antibody activation of T_mem cells increased their chemotaxis to CCL5, which was dependent predominantly on glycolysis rather than FAO. The sources supplying energy for diverse functions of unactivated T_mem cells differ from that required for function after immune activation.—Taub, D. D., Hesdorffer, C. S., Ferrucci, L., Madara, K., Schwartz, J. B., Goetzl, E. J. Distinct energy requirements for human memory CD4 T-cell homeostatic functions.

T cells require energy from ATP for homeostatic functions, such as maintenance, survival, and patrolling within the lymphoid system, as well as for specific immune responses involving proliferation, differentiation, chemotaxis to sites of tissue reactions, cytokine production, and cytotoxicity (1, 2). Glycolysis and mitochondrial oxidative phosphorylation of acetyl-CoA produced by β-oxidation of long-chain fatty acids (FAs), as well as from glycolysis-derived pyruvate are the principal sources of ATP in T cells. The relative contributions of glycolysis and FA oxidation (FAO) to T-cell energy are determined by the predominant T-cell subset and prevailing state of differentiation and activation.

Glucose uptake and increased glycolytic activity are needed to support T-cell activation and immune effector functions, and signals from coreceptors of the T-cell antigen receptor (TCR) are required for sustained optimal assembly of the pathways of glucose entry and metabolism (1, 3–5). Generation of ATP from FAO in T-cell mitochondria is evoked by stimulation of type α1 AMP-activated protein kinase (AMPKα1) through TCR signaling or elevation of intracellular [Ca²⁺] or the energy stress of decreased intracellular [ATP] (6, 7). Mouse T cells lacking AMPKα1 or its activating kinase have reduced FAO and a greater susceptibility to apoptosis induced by withdrawal of IL-2 without immune stimulation, suggesting that ATP from this source is required for survival of quiescent T cells (7). Greater expression of surface markers of activation and production of inflammatory cytokines by stimulated T cells of AMPKα1-deficient mice and of mice with transgenic up-regulation of the Glut1 glucose transporter correlated in both lines with their higher than normal levels of glucose uptake and glycolysis as a source of ATP for effector T-cell (T_eff-cell) functions (4, 7, 8).

The roles of distinct pathways of ATP generation in supporting T-cell differentiation were examined initially in mouse CD8 memory T cells (T_mem cells) and CD8 T_eff cells. Genetic deletion of a CD8 T-cell adaptor protein, which thereby profoundly depressed AMPKα1 activation and mitochondrial FAO, blocked differentiation of CD8 T_mem cells without affecting development of CD8 T_eff cells (9). Pharmacological restoration of AMPK activation, and thus mitochondrial FAO, reestablished normal generation of CD8 T_mem cells despite continued absence of the deleted adaptor protein. Inhibition of mouse CD8 T-cell mammalian target of rapamycin (mTOR) with rapamycin or an RNAi sufficiently to suppress glycolysis and enhance mitochondrial FAO increased the number of viral antigen-specific CD8 T_mem-cell precursors during the expansion phase and accelerated CD8 T_mem-cell differentiation during the contraction phase (10, 11).

Similar investigations of mouse CD4 T-cell subset differentiation showed that energy for development of immune properties of Th1, Th2, and Th17 T_eff cells is derived from high levels of Glut1-mediated glucose uptake and glycolysis, whereas energy for acquisition of the immune activities of inducible T_reg cells comes from mitochondrial FAO (8, 12). Dependence of differentiation of these CD4 T-cell subsets on distinctive metabolic profiles also was demonstrated by the predominance of T_eff cells in Glut1-transgenic mice with enhanced T-cell glycolysis and by selective increases in T_reg-cell generation when AMPKα1 was stimulated in a mouse model of asthma (8).

CD4 T_mem cells are necessary for facilitated production of antigen-specific CD4 T_eff cells, as well as for optimal generation of CD8 T_mem cells, high-affinity memory B cells, and long-lived plasma cells (13–17). Despite these central roles of CD4 T_mem cells in adaptive immunity, there is little information about which sources of energy are required for CD4 T_mem-cell homeostatic functions of maintenance, survival, and trafficking in the lymphoid system or for their interactions with other immune cells. The present results describe energy-producing metabolic activities of human blood unactivated CD4 T_mem cells, their respective sources of energy for diverse homeostatic functions, and the effects of aging and activation on these energy profiles.

MATERIALS AND METHODS

Isolation of human CD4 T_mem cells

The volunteers included old (n=27, mean±sd age 78±6.4 yr, range 66–94 yr) and sex-matched young (n=23, mean±sd age 35±6.1 yr, range 22-45 yr; 4 were studied twice) healthy subjects, who signed consent forms approved by the review committees at each institution. Calcium- and magnesium-free phosphate-buffered saline (PBS⁻²)-diluted heparinized blood was centrifuged on Ficoll-Paque (GE Healthcare Life Sciences, Pittsburgh, PA, USA) as described previously (18). The total population of CD4 T_mem cells in mixed mononuclear leukocytes at the Ficoll-Paque:buffer interface was purified by immunomagnetic negative selection with 2 passages through LS-type bead columns in a magnetic field; CD4 T_mem-cell purity was 96 to 98% by flow cytometric assessment of CD45RA and CD45RO (Miltenyi-Biotec, Auburn, CA, USA).

Quantification of CD4 T_mem-cell glycolysis and β-oxidation of mitochondrial FAs

For both assays, selected wells of 48-well plates (Corning Life Sciences, Lowell, MA, USA) were precoated for 24 h at 4°C with 0.3 ml of human placental type IV collagen (Sigma-Aldrich, St. Louis, MO, USA) at 1 mg/ml in 0.15 M NaCl titrated to pH 5.0 with 5 M acetic acid or 1 μg of anti-human CD3 antibody, without or with mouse monoclonal anti-human CD28 antibody, as described previously (19). A portion of the CD4 T_mem cells were activated by preincubation in fetal bovine serum (FBS)-RPMI 1640 for 24 h on adherent anti-human CD3 antibody plus anti-human CD28 antibody prior to transfer of replicate 0.5-ml suspensions of 0.5 × 10⁶ activated T_mem cells to uncoated wells. Unactivated CD4 T_mem cells were preincubated for 24 h in RPMI 1640 with 10% (v:v) charcoal- and dextran-extracted FBS, 100 U/ml of penicillin G, and 50 μg/ml of streptomycin (CD-FBS-RPMI 1640) prior to transfer of replicate 0.5-ml suspensions of 0.5 × 10⁶ T_mem cells to collagen-coated wells. Inhibitors were introduced into selected wells, followed by incubation at 37°C in 5% CO₂ for 1 h, the addition of stimuli, and further incubation at 37°C in 5% CO₂ for 16 h. The stimuli for unactivated T_mem cells were 30 nM human chemokine (C-C motif) ligand 19 (CCL19; Peprotech, Rocky Hill, NJ, USA) and 100 nM sphingosine l-phosphate (S1P; Sigma-Aldrich) or anti-human CD3 plus anti-human CD28 antibodies, and the stimulus for activated T_mem cells was 30 nM human CCL5 [also designated regulated and normal T-cell expressed and secreted (RANTES); Peprotech].

To measure glycolysis by quantification of extracellular concentrations of l-lactate, the T_mem cells were sedimented by centrifugation of plates at 1000 g for 10 min at 4°C, followed by removal of 0.2-ml portions of each supernatant. For the ELISA, each well of a 96-well plate received either 100 μl of an l-lactate standard ranging in concentration from 15.7 μM to 1 mM or 10 μl of a diluted T_mem-cell supernatant plus 90 μl of assay buffer. The reactions then were developed according to kit directions (Cayman Chemical, Ann Arbor, MI, USA),and optical density was determined at 490 nm in a VersaMax ELISA reader (Molecular Devices, Sunnyvale, CA, USA).

To measure FAO, etomoxir (0.2 mM; Calbiochem-EMD Chemicals, Gibbstown, NJ, USA) and dorsomorphin dihydrochloride (1 μM; Tocris Bioscience, Minneapolis, MN, USA) were introduced into replicate sets of 0.5-ml suspensions of unactivated T_mem cells to block mitochondrial uptake and β-oxidation of FAs, respectively, followed in 2 h by CCL19 or S1P for collagen-coated wells and the FAO stimulus 1 mM AICAR (Tocris Bioscience, Minneapolis, MN, USA) or the inhibitor of glycolysis 5 mM 2-deoxy-d-glucose (Sigma-Aldrich). Etomoxir and dorsomorphin dihydrochloride also were introduced into replicate sets of 0.5-ml suspensions of activated T_mem cells, followed in 2 h by CCL5 or AICAR or 2-deoxy-d-glucose. After 1 h of preincubation, each well received 1 μCi of (9,10-³H)-palmitic acid (ICN Radiochemicals, Costa Mesa, CA, USA) in 10 μl of 10% FA-free BSA (Sigma-Aldrich) with 20 μM nonradioactive palmitic acid (Sigma-Aldrich). After incubation for 24 h, the plates were centrifuged at 1000 g for 10 min, and 150 μl of supernatant from each well was applied to a 1-ml Dowex 1×8-200 column (Dow Water and Process Solutions, Edina, MN, USA) that was developed with 2.5 ml of water, as described previously (20, 21). Tritium in 1 ml of each eluate was quantified in a Beckman LS6500 liquid scintillation counter (Beckman Coulter, Fullerton, CA, USA).

Assessment of CD4 T_mem-cell chemotaxis and adherence

Unactivated CD4 T_mem cells were incubated overnight in CD-FBS-RPMI 1640 to deplete cellular S1P before stimulation or for 24 h in FBS-RPMI 1640 on adherent anti-human CD3 plus anti-human CD28 to activate T_mem cells before stimulation with CCL5 as for the metabolic studies. Transwell plate-permeable upper inserts with a 5-μm-diameter pore filter (Corning Life Sciences) were preincubated overnight at 4°C in human type IV collagen, washed, and dried as described previously (18). Some portions of T_mem cells were preincubated for 1 h at 37°C without and with 0.2 mM etomoxir plus 1 μM dorsomorphin or 1 mM AICAR or 5 mM 2-deoxy-d-glucose or 50 nM rapamycin (Fisher Scientific, Pittsburgh, PA, USA). Each upper insert received 10⁶ unactivated T_mem cells in 0.1 ml of CD-FBS-RPMI 1640 or 10⁶ activated T_mem cells in 0.1 ml of FBS-RPMI 1640 and was placed in a well containing 0.6 ml of CD-FBS-RPMI 1640 without (background control) or with 100 nM S1P or 30 nM CCL19 for unactivated T_mem cells or with 30 nM CCL5 for activated T_mem cells. After incubation at 37°C in 5% CO₂ for 4 h, the number of T cells in each lower compartment was determined by manual microscopic counting of samples blinded by coding. The results are expressed as a percentage of the initial number of T_mem cells in the upper insert.

To quantify adherence, replicate suspensions of 0.5 × 10⁶ T_mem cells in 0.5 ml of CD-FBS-RPMI 1640 were preincubated for 1 h at 37°C without or with 0.2 mM etomoxir plus 1 μM dorsomorphin or 1 mM AICAR or 5 mM 2-deoxy-d-glucose and added to collagen-coated wells, where some received 100 nM S1P or 30 nM CCL19. After 2 h at 37°C in 5% CO₂, wells were washed 3 times with 1 ml of PBS⁻², and the numbers of adherent T_mem cells in 10 fields at ×160 were counted to determine the mean level per field.

Determination of CD4 T_mem-cell apoptosis

Replicate suspensions of 0.5 × 10⁵ T_mem cells in 0.5 ml of CD-FBS-RPMI 1640 were preincubated for 1 h at 37°C without or with 0.2 mM etomoxir plus 1 μM dorsomorphin or 5 mM 2-deoxy-d-glucose prior to addition to either wells that had been precoated with collagen and then received 30 nM CCL19, or uncoated wells that received 10 μg of neutralizing rat monoclonal IgG2a/κ anti-human IL-2 antibody (clone MQ1-17H12; BioLegend, San Diego, CA, USA). After incubation for 16 h at 37°C in 5% CO₂, the T_mem cells were washed twice in 1 ml of PBS⁻² and lysed. Intact nuclei and other cellular fragments were removed by centrifugation, and nucleosomes in the cytosol were measured by an ELISA kit based on antihistone antibody capture and conjugated anti-DNA antibody detection, with reading of color absorbancy at 405 nm (Roche Applied Science, Indianapolis, IN, USA).

RESULTS

Mitochondrial long-chain FAO, but not glycolysis, was significantly greater in CD4 T_mem cells of old than young healthy subjects without stimulation and for every condition examined (Fig. 1). FAO in T_mem cells of old and young healthy subjects was increased significantly by AICAR, which enhances activity of the FAO rate-limiting enzyme AMPKα1, with respective mean increments of 417 and 33%. The increases in FAO evoked by 2-deoxy-d-glucose inhibition of glycolysis in T_mem cells of old and young subjects were similarly ranked at 349 and 74%, respectively (Fig. 1). CCL19 and S1P, which control lymphoid trafficking of T cells and immune-activating antibodies to the TCR, augmented FAO by T_mem cells of old or young subjects only minimally or not at all. Similarly, CCL19, S1P, and also AICAR, here only minimally increased glycolysis by T_mem cells from both old and young subjects or had no effect (Fig. 1). In sharp contrast, antibodies to the TCR very significantly increased glycolysis by T_mem cells from old and young subjects. Thus, TCR-dependent immune activation of T_mem cells recruits energy from glycolysis, but not FAO, whereas the chemotactic responses of CD4 T_mem cells during homeostatic lymphoid patrolling are not consistently associated with increases in FAO or glycolysis.

Figure 1.

Energy-generating activities of CD4 T_mem cells from healthy young and old subjects. Left panel: FAO. Each column and bar depicts the mean ± sd of results for 4 old and 4 young matched subjects, where each value is the counts per minute of ³H-water produced by 10⁶ CD4 T_mem cells/ml. Right panel: glycolysis. Each column and bar depicts the mean ± sd of results for 4 old and 4 young matched subjects, where each value is the extracellular concentration of sodium l-lactate. 2-DG, 2-deoxy-d-glucose; Ab, antibody; Coll, type IV collagen; Glu, glucose. The significance of differences between mean levels for challenged CD4 T_mem cells and resting CD4 T_mem cells (0) from the same subjects was calculated by an unpaired t test. ⁺P < 0.05, *P < 0.01, **P < 0.001.

Adherence to human type IV collagen is a homeostatic activity of immunologically unactivated T_mem cells that was enhanced significantly (P<0.01) for both age sets by CCL19 and S1P (Fig. 2). Augmentation of adherence of T_mem cells from both age sets by the chemotactic factors that control homeostatic patrolling was significantly suppressed by inhibition of FAO, but not by inhibition of glycolysis to a similar extent by 2-deoxy-d-glucose. Stimulation of FAO by AICAR (Fig. 1) significantly increased chemotactic factor-augmented adherence of T_mem cells from old subjects, but it suppressed (Fig. 2, left panel) or did not affect (Fig. 2, right panel) adherence of T_mem cells from young subjects. As for adherence, chemotaxis to CCL19 and S1P of immunologically unactivated CD4 T_mem cells from both age sets was suppressed significantly by inhibition of FAO, but not by inhibition of glycolysis (Fig. 3). Also as for adherence, AICAR enhancement of FAO significantly augmented chemotaxis of T_mem cells from old subjects, but it suppressed (Fig. 3, left panel) or did not affect (Fig. 3, right panel) chemotaxis of T_mem cells from young subjects (Fig. 3).

Figure 2.

Dependence of CD4 T_mem-cell adherence to type IV collagen on energy from FAO but not glycolysis. Left panel: CCL19 stimulation. Right panel: S1P stimulation. Each column and bar depicts the mean ± sd of results for 4 old and 4 young matched subjects. All values are the percentage of adherence of T_mem cells to type IV collagen with a stimulus (CCL19 or S1P) alone (0) or a stimulus plus a metabolic agent compared to that in the absence of the stimulus or an agent (100%). Mean ± SD of adherence (n=4) without a stimulus or agent (100%) was 33 ± 2.1 T_mem cells/×160 field for young and 33 ± 4.8/×160 field for old subjects. E + D, etomoxir plus dorsomorphin. ⁺P < 0.05, *P < 0.01, **P < 0.001 vs. stimulus without a metabolic agent.

Figure 3.

Dependence of CD4 T_mem-cell chemotaxis on energy from fatty acid oxidation but not glycolysis. Left panel: CCL19 stimulation. Right panel: S1P stimulation. Each column and bar depicts the mean ± sd of results for 4 old and 4 young matched subjects. All values are the percentage of chemotaxis of T_mem cells to a stimulus with a metabolic agent relative to that without an agent (100%). Mean ± sd chemotaxis without a metabolic agent (100%) was 15 ± 2.3% of total CD4 T_mem cells with CCL19 and 5.6 ± 3.5% with S1P for young and 10 ± 2.2% with CCL19 and 3.0 ± 0.9% with S1P for old subjects. ⁺P < 0.05, *P < 0.01, **P < 0.001 vs. stimulus without a metabolic agent (100%).

T_mem cells were partially activated immunologically by incubation for 24 h with anti-TCR antibodies to simulate the state of CD4 T_eff cells and confirm their functional dependence on energy from glycolysis (4, 5, 8). Mean ± sd FAO with anti-TCR antibodies plus CCL5 was 9344 ± 7156 cpm/10⁶ T cells from young donors and 20,484 ± 9858 cpm/10⁶ T cells from old donors (P=0.049 for old vs. young, n=6). Mean ± sd glycolysis with anti-TCR antibodies plus CCL5 was 14.9 ± 1.2 mM lactate for 10⁶ T cells/ml from young donors and 13.7 ± 1.2 mM lactate for 10⁶ T cells/ml from old donors (not significant for old vs. young, n=3; P<0.001 for both values relative to basal levels of glycolysis in unactivated T_mem cells). Thus, the metabolic activities of T_mem cells activated with anti-TCR antibodies plus CCL5 were similar to those quantified initially with anti-TCR antibodies alone (Fig. 1). The chemotactic responses of unactivated CD4 T_mem cells from both age sets to the immune inflammatory stimulus CCL5, which preferentially attracts activated T cells, were suppressed significantly by inhibition of FAO but not by inhibition of glycolysis (Fig. 4). CCL5-stimulated chemotaxis of activated T_mem cells from both age sets was significantly greater than that for the corresponding population of unactivated T_mem cells (P<0.01). In contrast to the exclusive dependence on FAO of chemotaxis of unactivated T_mem cells, CCL5-stimulated chemotaxis of activated T_mem cells was very significantly suppressed by inhibition of glycolysis but only minimally suppressed by inhibition of FAO (Fig. 4).

Figure 4.

Greater dependence of activated CD4 T_mem cell chemotaxis on energy from glycolysis than fatty acid oxidation. Left panels: CCL5 stimulation of chemotaxis of unactivated CD4 T_mem cells (100%) and suppression of the responses by inhibitors of FAO (E+D) but not glycolysis (2-deoxy-d-glucose). Right panels: CCL5 stimulation of chemotaxis of TCR-activated CD4 T_mem cells (100%) and suppression of responses by the metabolic inhibitors. Each column and bar depicts the mean ± sd of results for 4 old and 4 young matched subjects. E + D, etomoxir plus dorsomorphin. ⁺P < 0.05, *P < 0.01, **P < 0.001 vs. stimulus without a metabolic inhibitor (100%).

To verify this change in energy source for chemotaxis that occurs after T_mem-cell immune activation, we examined CCL5-stimulated chemotaxis of CD4 T_mem cells from the same healthy subjects before and after exposure to activating anti-TCR antibodies (Table 1). Rapamycin was employed in place of 2-deoxy-d-glucose to inhibit glycolysis without affecting FAO by an mTOR-dependent mechanism. The chemotaxis of CD4 T_mem cells from both young and old subjects to CCL5 was significantly greater after TCR-dependent activation (Table 1). Chemotaxis of unactivated T_mem cells from both age sets to CCL5 was exclusively dependent on energy from FAO. After TCR-dependent activation of T_mem cells from the same subjects, T_mem-cell chemotaxis to CCL5 was highly dependent on energy from glycolysis but derived little energy from FAO.

Table 1.

Different energy requirements for chemotaxis to CCL5 before and after activation of T_mem cells

Subject age	TCR activation	Chemotaxis to CCL5 (% of Tmem cells)	Percentage of control chemotaxis
			E + D	Rapamycin
Old	0	7.3 ± 2.9	71 ± 7.6*	94 ± 4.0
Old	+	13 ± 3.3	85 ± 4.6*	49 ± 6.7**
Young	0	12 ± 4.3	73 ± 7.6*	90 ± 6.1
Young	+	24 ± 2.1	92 ± 3.5+	46 ± 8.1**

Values are means ± sd of results for 3 old and 3 young matched subjects. T_mem cells were activated by incubation on adherent anti-CD3 + anti-CD28 antibodies for 24 h. E + D, etomoxir plus dorsomorphin.

⁺P < 0.05,

^*P < 0.01,

^**P < 0.001.

In contrast to the exclusive dependence on energy from FAO for homeostatic adherence and chemotaxis, optimal prevention of apoptosis of unactivated CD4 T_mem cells required energy from both FAO and glycolysis. Under study conditions, CCL19 and a neutralizing antibody to IL-2 modestly increased apoptosis of T_mem cells from both age sets (Fig. 5, left panel). Baseline and elicited apoptosis of T_mem cells from old subjects was significantly greater than corresponding apoptosis of T_mem cells from young subjects. Significant increases in elicited apoptosis of unactivated T_mem cells from young and old donors were evoked by inhibition of FAO or glycolysis (Fig. 5). Thus, optimal survival of unactivated CD4 T_mem cells exposed to IL-2 withdrawal or CCL19 depended on energy from both glycolysis and mitochondrial FAO.

Figure 5.

Energy requirements for prevention of apoptosis of T_mem cells. Left panel: apoptosis without (0) and with exposure to type IV collagen (Coll) plus CCL19 or to a neutralizing a-IL-2 antibody (Ab). Each column and bar depicts the mean ± sd of results for 4 old and 4 matched young subjects, where each value is in absorbancy units at 405 nm from the ELISA. ⁺P < 0.05 vs. young group. Right panel: metabolic requisites for survival. Each column and bar depicts the mean ± sd of results for 4 old and 4 young matched subjects, where each value is the percentage of apoptosis relative to that for each stimulus without a metabolic agent (100%). ⁺P < 0.05, *P < 0.01, **P < 0.001 vs. stimulus without a metabolic agent (100%).

DISCUSSION

Stimulation of T_mem cells by concentrations of chemotactic factors that maintain homeostatic patrolling of the lymphoid system did not substantially affect levels of FAO or glycolysis (Fig. 1). In contrast, TCR-dependent immune activation of T_mem cells enhanced energy generation from glycolysis without altering FAO. Although stimulation of homeostatic responses of T_mem cells was not associated with increased levels of either FAO or glycolysis, each such function was dependent on energy generated by basal FAO or both basal FAO and glycolysis (Figs. 2, 3, and 5). Despite the known heterogeneity of the T_mem-cell subset (22), the capacities of the total set of CD4 T_mem cells from old and young subjects to adhere to and migrate on type IV collagen were significantly dependent on basal FAO but not glycolysis (Figs. 2 and 3). In contrast, minimization of apoptosis required energy from both basal FAO and glycolysis (Fig. 5). The dependence of optimal survival of unactivated T_mem cells after different challenges on both basal glycolysis and FAO also was clearly distinguishable from the predominant requirement of activated T_mem-cell chemotaxis on enhanced glycolysis (Fig. 4). FAO in T_mem cells is one of their few biochemical activities that is greater in magnitude and more efficiently coupled to functions for old than young subjects (Figs. 1–3).

Restriction of the chemotactic factors examined in unactivated CD4 T_mem-cell adherence and migration assay systems to CCL19 and S1P, which control homeostatic trafficking of CD4 central T_mem cells in the lymphoid system, was a selective pressure applied to focus on the central memory subset and avoid immune activation that might occur with inflammatory chemotactic factors. In recent related studies of the energy requirements for chemotaxis of activated human T_eff cells to the chemokine CCL5 (aka RANTES), there was significant enhancement of glucose uptake, glycolysis, and AMPKα1 activity (23). Further, inhibition of either glycolysis or AMPKα1, and, therefore, presumably FAO, suppressed chemotaxis to CCL5. Our findings correlate with these data, as TCR-dependent activation of human CD4 T_mem cells reduced the dependence of their chemotaxis to CCL5 on FAO and created a predominant requirement for energy from glycolysis (Fig. 4, Table 1). Thus, unlike the sole dependence on FAO of homeostatic chemotactic responses of T_mem cells to CCL19 and S1P (Fig. 3), the chemotaxis of activated T_mem cells and T_eff cells to CCL5 required energy principally from glycolysis and to a much lesser extent from FAO (Fig. 4 and Table 1). This also implies that such inflammatory chemotactic responses of T cells involve new synthesis of proteins and perhaps other types of structural molecules, which are derived from the anabolic aspects of glycolysis in contrast to the catabolism associated with FAO.

AICAR is a stimulus of AMPKα1 activity and thereby significantly enhanced FAO, but not glycolysis, in T_mem cells from old and young subjects. Unexpectedly, opposite functional consequences of similar AICAR-enhanced mitochondrial FAO were observed in T_mem cells from old and young subjects. As for the possible outcomes in other types of cells (24), the energy generated from increased mitochondrial FAO in T_mem cells from old subjects (Fig. 1) was adaptively well coupled to greater adherence and chemotaxis, whereas that generated in T_mem cells from young subjects was maladaptively wasted or resulted in diminished adherence and chemotaxis (Figs. 2 and 3). The latter functional consequence may be attributable to cellular injury by mitochondrial reactive oxygen species (ROS) instead of positive metabolic and inflammatory signaling (25). The ratio of the level of mitochondrial ROS to that of protective antioxidants will determine the net functional outcome in stimulated cells (25, 26). Extensive additional studies will be necessary to elucidate the differences between pathways of ROS generation and destruction in T_mem cells from old and young subjects. However, the available data already suggest that activators of AMPKα1 activity, which have been developed for some human endocrine and metabolic diseases may be useful enhancers of the development and functions of T_mem cells of the elderly in the settings of defense against infections and responses to vaccines.