点击显示 收起
1 From the Division of Geriatrics and the Program of Immunology (AL, ND, CF, and TF) and the Division of Endocrinology (AG, FF, and ACC), University of Sherbrooke, Sherbrooke, Canada
2 Supported by the Canadian Diabetes Association (in honor of the late Marion L Monroe) and the Canadian Institutes of Health Research (MOP 53094 and 63149). ACC is a new investigator of the Canadian Institutes of Health Research. 3 Address reprint requests to AC Carpentier, Division of Endocrinology, Centre Hospitalier Universitaire de Sherbrooke, 3001 12th Avenue North, Sherbrooke, Québec, Canada J1H 5N4. E-mail: andre.carpentier{at}usherbrooke.ca.
ABSTRACT
Background: Previous studies have shown suppressive effects of polyunsaturated fatty acids (PUFAs) on T cell proliferation, but the precise mechanism for this effect has not been fully investigated in vivo in humans.
Objective: The objective was to determine whether this effect is the result of altered T cell membrane properties and impaired CD3- and CD28-mediated signaling in vivo in humans.
Design: Peripheral T cells were isolated from healthy subjects before and 2 h after an intravenous infusion of heparin plus a PUFA-rich lipid emulsion during a euglycemic hyperinsulinemic clamp to induce a 2.5-fold elevation in plasma linoleic acid concentration without significant change in plasma total free fatty acid concentrations.
Results: Intravenous infusion of heparin plus the lipid emulsion reduced peripheral T cell membrane fluidity and altered lipid raft organization, both of which were associated with reduced T cell proliferation after stimulation with CD3 plus CD28. Tyrosine phosphorylation of linker of activated T cells and activation of protein kinase B in T cells were also impaired without a reduction in T cell receptor expression. In addition, acute PUFA elevation was associated with a reduction in T cell membrane cholesterol exchange with the cellular milieu ex vivo.
Conclusions: A selective increase in plasma linoleic acid concentration and in intravascular lipolysis has a suppressive effect on peripheral T cell CD28-dependent activation, and this effect is associated with changes in plasma membrane properties. Our results have important implications for nutritional therapy in patients at high risk of septic complications and may also be of relevance to postprandial lipid metabolism disorders such as insulin resistance and type 2 diabetes.
Key Words: Polyunsaturated fatty acids • intravascular lipolysis • triacylglycerol • T cells • T cell receptor • postprandial lipid metabolism
INTRODUCTION
T cell activation and clonal expansion, critical steps in the acquired immune response, are initiated via T cell receptor (TCR) ligation by the major histocompatibility complex (1). A full state of activation and proliferation is reached only on triggering of costimulatory molecules such as CD28 (2). Early signaling events triggered by costimulation with TCR and CD28 include a rise in intracellular calcium as well as membrane reorganization and scaffolding of lipid raft-associated membrane receptors and proteins, leading up to activation of Src protein tyrosine kinases such as p56Lck (3). Activated p56Lck recruits and activates ZAP-70 from the Syk family, which in turn phosphorylates the linker of activated T cells (LAT) (4). LAT has no intrinsic activity but, once phosphorylated, it acts as a platform for signaling proteins. These early events pave the way to a wide range of intermediate signaling pathways, including the mitogen-activated protein kinase (MAPK) pathway, and to cytoskeleton reorganization, which is dependent on CD28 ligation (5). Together, activation of these signaling pathways leads to nuclear relocalization of transcriptional factors such as nuclear transcription factor AT and nuclear transcription factor B (6). Once these pathways are activated, triggered-antigen-specific T cells produce interleukin (IL) 2 and commit to proliferation.
n–3 Polyunsaturated fatty acids (PUFAs) have immunosuppressive effects on T cell functions in mice (7). Early events of TCR signaling, calcium metabolism, and IL-2 secretion are impaired in Jurkat T cells in the presence of n–3 PUFAs (8, 9). These effects appear to result from alterations in lipid raft composition of T cells, including an increase in sphingomyelin unsaturated acyl chains (9–12). Other highly prevalent dietary cis unsaturated fatty acids, such as oleate and linoleate, also impair T cell activation. For example, in vitro incubation of Jurkat or autologous peripheral T cells with serum samples from subjects undergoing intravenous infusion with heparin plus a PUFA-rich lipid emulsion during a hyperinsulinemic euglycemic clamp, to produce an elevation of plasma free fatty acid (FFA) concentrations from intravascular lipolysis of triacylglycerols, was shown to reduce intracellular calcium after CD3 or CD28 stimulation. Stulnig et al (13) described similar results in T cells after an intravenous infusion of heparin plus a PUFA-rich lipid emulsion.
These observations raise the possibility that acute intravascular lipolysis of PUFA-containing triacylglycerols may lead to potentially important suppressive effects on the acquired immune system. To our knowledge, however, no previous study has addressed the relation between a potentially immunosuppressive effect of in vivo intravascular lipolysis of triacylglycerols rich in PUFA and T cell membrane fluidity, lipid raft organization, and CD3/CD28 signaling in humans. This study aimed to determine whether enhanced intravascular lipolysis of PUFA-containing triacylglycerols impairs CD3/CD28-dependent signaling, T cell proliferation, and T cell membrane properties. T cell cholesterol turnover was also assessed because cholesterol is important to lipid raft structure, which in turn sustains proliferation.
SUBJECTS AND METHODS
Subjects
Ten (3 women, 7 men) healthy nonobese white subjects [
The effect of insulin on plasma FFA metabolic partitioning, during enhanced intravascular lipolysis, was determined with the use of an intravenous infusion of heparin plus a PUFA-rich lipid emulsion. The complete experimental design and metabolic results of this study are reported elsewhere (15). The following description applies to the immunologic part of the study, which is the focus of the present report. Participants were admitted to our metabolic investigation center, on each occasion and for the duration of the study, between 0730 and 0830 after fasting overnight for 12 h. On arrival, body weight and height were measured, and lean body mass was determined by electrical bioimpedance (Hydra ECF/ICF; Xitron Technologies, San Diego, CA). An intravenous catheter was placed in one forearm for the infusions and another was placed, in a retrograde fashion, in the contralateral arm maintained in a heating box (55 °C) for blood sampling.
Sustained elevation of intravascular lipolysis of triacylglycerols was induced with an intravenous infusion of heparin (250 U/h; Hepaléan, Organon Teknika, Scarborough, Canada) plus a PUFA-rich lipid emulsion (40 mL/h, Intralipid 20%; Baxter, Mississauga, Canada) (HL infusion), as described previously (16). Hyperinsulinemia was obtained by using a primed (0.8 mU/kg) continuous high (1.2 mU · kg–1 · min–1) insulin infusion (Novolin GE; NovoNordisk, Mississauga, Canada) with 10 mEq KCl/h. Fasting plasma glucose was maintained during the clamp with a variable infusion of dextrose 20% adjusted to plasma glucose concentration as measured every 5 min. Octreotide acetate (30 µg/h; Omega, Montreal, Canada) and human recombinant growth hormone (3 ng · kg–1 · min–1, Humatrope; Eli Lilly, Toronto, Canada) were also infused during the clamp (17).
After 30 min of bed rest, blood samples were taken at 10-min intervals at baseline and between 90 and 120 min of the intravenous infusion for the measurement of plasma insulin, total triacylglycerol, glycerol, and FFA concentrations. Blood samples were collected into tubes containing Na2EDTA and Orlistat (30 µg/mL; Roche, Mississauga, Canada) to prevent in vitro triacylglycerol lipolysis. Additional blood samples were collected into evacuated tubes containing sodium citrate, at 0 and 120 min, to isolate circulating mononuclear cells, as described below.
Plasma assays
Glucose was assayed at bedside (Beckman Glucose Analyzer II; Beckman Instruments Corporation, Fullerton, CA). Insulin and growth hormone were measured with specific radioimmunoassays (Linco Inc, St Charles, MO; Nichols Institute Diagnostics, San Juan Capistrano, CA). Total plasma FFAs and triacylglycerols were measured with colorimetric assays (Wako Industrials, Neusshafen, Germany; Thermo DMA, Arlington, TX). Plasma glycerol was extracted and derivatized with bis(trimethylsilyl)-trifluoroacetamide + 10% trimethylchlorosilane (Regis Technologies, Morton Grave, IL), and plasma glycerol was measured by gas chromatography–mass spectrometry (GC-MS) with the use of an Agilent GC model 5890A (Agilent Technologies, Avondale, PA) coupled to an MS detector (model 5971 quadrupole MSD; Agilent) equipped with a fused silica column (25 m x 0.20 mm, 0.33 µm, Supelco SPB-5, Supelco, Oakville, Canada) and a splitless injector. Electron impact ionization with an electron beam energy of 70 eV was used in selected ion monitoring mode at mass-to-charge ratios of 117 and 205 for glycerol and of 118 and 206 for [1-13C]glycerol (internal standard). To measure plasma palmitate, linoleate, and oleate, heptadecanoic acids was added as an internal standard to 100 µL plasma and mixed with 500 µL methanol. After centrifugation (2000 x g, 15 min, 4 °C), the supernatant fluid was filtered and injected on a column (5 µm, 4.0 x 125 mm, Hypersil ODS; Agilent Technologies) on a liquid chromatography–mass spectrometry detector series 1100 (Agilent) with monitoring of ions 279 (18:2), 281 (18:1), 255 (16:0), and 269 (17:0 as internal standard). Standard curves were generated for 16:0, 18:1, and 18:2 with the use of purified standards of known concentration. Intraassay and interassay CVs were <6.1% for all assays.
Isolation of T cells and cultures
Circulating mononuclear cell fractions were isolated from blood samples by Ficoll-Hypaque density sedimentation, as already described (18), and depleted of monocytes by adhesion to plastic tissue-culture flasks coated with autologous serum (1 h, 37 °C). B lymphocytes and residual phagocytic cells were removed by absorption to a prewarmed nylon wool column (1 h, 37 °C). The resulting T cell preparations were phenotyped by flow cytometry and shown to contain <3% contaminating B or natural killer (NK) cells (19) and consisted of >96% CD3-positive cells with <1.0% surface immunoglobulin M (B cells)-, CD16 (NK cells)-, and CD14 (monocytes)-positive contaminating cells. Cell viability was >95%, as estimated by trypan blue exclusion.
Flow cytometry analysis
Freshly separated T cells (1 x 106 lymphocytes/mL) were resuspended in phosphate-buffered saline (PBS) and labeled with fluorescein isothiocyanate (FITC)–conjugated anti-TCR-/ß or with anti-CD28 antibodies (Becton Dickinson, Montreal, Canada). After being washed, the secondary antibody conjugated to FITC was added for another 30 min. T cell apoptosis was determined by FITC-conjugated Annexin-V and propidium iodide (5 µL/mL of a 50 µg/mL stock solution; Sigma-Aldrich, St Louis, MO) staining for 30 min at 4 °C in the dark. Binding buffer was added to each tube, and cells were analyzed immediately by flow cytometry (FACScalibur; Becton Dickinson, Montreal, Canada) as already described (20). Data were expressed as the percentage of Annexin positive T cells relative to propidium iodide negative T cells.
Ex vivo T cell stimulation
Freshly prepared T cells were kept for 1 h in RPMI medium at 37 °C. Peripheral blood lymphocytes (2 x 105 cells/mL) were cultured in triplicate for up to 96 h, and T cell proliferation in response to exposure to anti-CD3 mAb (Leu-4, 5 µg/mL, clone UCHT-1; Sigma-Aldrich), with or without anti-CD28 (5 µg/mL, clone 28.2; Becton Dickinson, Montreal, QC), was determined by [3H]thymidine incorporation as previously described (21). Control cells (20 x106 lymphocytes) were left untreated.
Western blotting
Proteins (20 µg) from total cell lysates were diluted in Laemmli buffer containing DL-dithiothreitol, heated in a boiling water bath for 5 min, spun for 5 min (16 000 x g), resolved by sodium dodecyl sulfate polyacrylamide gel electrophoresis under reducing conditions (10% acrylamide gels), and transferred to nitrocellulose membranes (Amersham, Baie d'Urfé, Canada). Membranes were treated for 1 h at room temperature with tris-buffered saline (20 mmol tris/L, 137 mmol NaCl/L; pH 7.6) containing 0.1% (by vol) Tween 20 and a solution of 5% (wt:vol) skim milk powder, followed by incubation with the relevant antibodies: anti-phosphotyrosine (1:1000, clone 4G10; Upstate Biotechnology, Lake Placid, NY), anti-pLAT (1:1000, Tyr 226; Upstate Biotechnology), anti-p42/44 MAPK 1/2 (1:1000; Santa Cruz Biotechnology, Santa Cruz, CA), anti-pAkt (1:500, Santa Cruz), or anti-p38 (1:2000, Santa Cruz) (19). Membranes were placed on a rotary end-over-end mixer and incubated overnight at 4 °C under constant rotary movement and then washed. The corresponding secondary antibody conjugated to horseradish peroxidase (Chemicon International, Temecula, CA) was added (1:2000) for 1 h. Membranes were then washed, and the proteins indicated enhanced chemiluminescence (Amersham). Densitometric analyses were performed with an image analyzer (Chemigenius2 Bio Imaging System; Syngene, Frederick, MD).
Measurement of fluorescence anisotropy in T cells
Cell membrane anisotropy (r) was determined by using fluorescent probe 1,6-diphenyl-1,3,5-hexatriene dissolved in tetrahydrofuran (4 mmol/L; Sigma-Aldrich) as previously described (19). Fluorescence was recorded on a spectrofluorimeter coupled to a Vextra polarizer (Hitachi F-4500; Oriental Motor Co Ltd, Tokyo, Japan). The lipid probe was excited by vertically polarized light at 360 nm, and the emitted light was recorded at 430 nm through a polarizer orientated parallel and perpendicular to the direction of polarization of the excitation beam. Fluidity (f) was derived from the inverse of anisotropy (f = 1/r).
Laser scanning confocal microscopy
T cells (1 x 106 lymphocytes) in PBS (1 mL) were treated with 10 µg Alexa 594-conjugated cholera toxin B subunit (Molecular Probes, Eugene, OR) for 15 min at 4 °C, washed, and resuspended in the same medium. Cells were allowed to adhere to coverslips coated with poly-L-lysine, and fluorescence was recorded along the z axis with laser scanning confocal microscopy (Thermo Noran, Middleton, WI) as previously described (19). Image processing and surface quantification of pixel intensities were performed with the use of the National Institutes of Health IMAGE freeware (Internet: http://rsb.info.nih.gov/nih-image/).
T cell cholesterol uptake and cholesterol efflux
T cells were incubated with [3H]cholesterol (1 µg/mL; Sigma-Aldrich) in RPMI-1640 containing 1% fetal bovine serum. In certain experiments, cells were washed with PBS and then lysed with 4 mol NaOH/L for radioactivity count. In the remainder of the experiments, T cells were intensively washed and reincubated in RPMI-1640 containing bovine serum albumin. Cholesterol released into the medium as well as the remaining cellular [3H]cholesterol was measured. Methyl-ß-cyclodextrin (MBCD) (0.5 mmol/L, Sigma-Aldrich) was used to enhance cholesterol exchange between the cell membrane and the medium.
Statistical analysis
Data are expressed as means ± SEMs, unless stated otherwise. For plasma metabolite and hormone concentrations, data collected at baseline and during the course of HL infusion were averaged. In ex vivo experiments involving a time course experiment or multiple stimulation conditions, a two-factor analysis of variance (ANOVA) for repeated measures was performed for time-by-treatment analysis. When a significant time-by-treatment interaction was shown, subsequent analyses were performed by using Scheffe's post hoc test to adjust for multiple comparisons. Otherwise, paired Student's t test was performed to assess the effect of HL infusion compared with baseline. Significance was set at P < 0.05.
RESULTS
Plasma metabolite and hormone concentrations
By design, plasma glucose and growth hormone concentrations did not change from baseline (Table 1), whereas plasma insulin increased during the euglycemic hyperinsulinemic clamp period (P < 0.05). Plasma triacylglycerols were higher than baseline concentrations during the clamp period with HL infusion (P < 0.05). HL infusion during the euglycemic hyperinsulinemic clamp did not affect total plasma FFA concentrations but tended to reduce plasma palmitate and oleate concentrations and led to an elevation in linoleate concentrations of 2.5-fold (P < 0.05). Plasma glycerol concentration was also significantly increased by 2.3-fold during HL infusion (P < 0.05).
View this table:
TABLE 1. Plasma concentration of metabolites and hormones at baseline and during euglycemic hyperinsulinemic clamp with intravenous infusion of heparin plus a lipid emulsion (HL)1
Change in T cell membrane anisotropy and membrane raft distribution
A shift in size and granularity of T cells was observed during HL infusion compared with baseline [forward size scatter: 193 ± 22 and 127 ± 25 arbitrary units, respectively (P < 0.05); side size scatter: 312 ± 41 and 207 ± 13 arbitrary units, respectively (P < 0.05); n = 4 subjects] (Figure 1). As shown in Figure 2, fluorescence anisotropy in T cells significantly increased during HL perfusion, compared with baseline (0.267 ± 0.013 and 0.218 ± 0.003, respectively; P < 0.05, Scheffe test). This reduction in membrane fluidity (f = 1/r) was apparent in T cells maintained in culture up to 24 h. After 48 h of incubation in culture media, T cell fluorescent anisotropy returned to that observed before HL infusion (NS, Scheffe test) (Figure 2). On confocal imaging, we found an altered lipid raft distribution in T cells during HL infusion (Figure 3B) compared with baseline (Figure 3A).
FIGURE 1.. Flow cytometry–determined T cell size and granularity before (A) and 2 h after (B) intravenous infusion of heparin plus a polyunsaturated fatty acid–rich lipid emulsion (HL) with the use of forward scatter (FSC) and side scatter (SSC), respectively. As shown by the upward and rightward shifts, T cell size and granularity increased after HL infusion (P < 0.05, paired Student's t test). n = 4 subjects.
FIGURE 2.. Mean (±SD) anisotropy (r) of T cells before (
FIGURE 3.. T cell membrane raft organization before (A) and 2 h after (B) intravenous infusion of heparin plus a polyunsaturated fatty acid–rich lipid emulsion in a single T cell from the same donor. The T cells were labeled with 594-Alexa conjugated cholera toxin B subunit and fluorescence was monitored by confocal microscopy. The images are from 1 representative experiment that was repeated in 3 different subjects, with similar results.
T lymphocyte proliferation and apoptosis
CD3- and CD28-induced proliferation of T cells sampled during HL infusion was reduced by 50% compared with baseline after 48 h of incubation (P < 0.05) and decreased proliferation was still present up to 96 h in culture (P < 0.05; Figure 4B). T cell proliferation was unaffected after 2 h of ex vivo incubation with octreotide (data not shown). The basal rate of apoptosis was not significantly changed after 2 h of HL infusion compared with baseline (4.7 ± 1.8% and 1.3 ± 0.6%, respectively). However, after 24 h in culture, T lymphocytes displayed a significant increase in the apoptotic rate after HL infusion compared with baseline (21.4 ± 4.7% and 4.1 ± 1.5%, respectively; P < 0.05). The percentage of necrotic cells determined by propidium iodide staining did not change after HL infusion compared with baseline.
FIGURE 4.. Mean (±SD) thymidine incorporation in T cells at baseline (
T cell receptor and CD28 expression and signaling
TCR (Figure 5A) and CD28 (Figure 5B) expression were significantly unchanged by HL infusion compared with baseline (TCR: 1089 ± 143 and 1245 ± 128 mean fluorescence intensity, respectively; CD28: 69 ± 14 and 58 ± 8 mean fluorescence intensity, respectively; n = 3 subjects). Moreover, the culturing of cells for extended periods of time did not affect CD28 expression significantly (data not shown). Total tyrosine phosphorylation (Figure 6A) after anti-CD3 and anti-CD28 stimulation was markedly reduced during HL infusion compared with baseline. As shown in Figure 6, B and C, enhanced LAT phosphorylation on activation was almost completely abolished after HL infusion compared with baseline. However, HL infusion had no effect on p42/44 MAPK and p38 tyrosine phosphorylation. We also found that CD28 signaling through PKB/Akt was markedly impaired in T cells after HL infusion compared with baseline (Figure 6 D and E).
FIGURE 5.. CD28 expression in T cells labeled with anti–T cell receptor (TCR; A) or anti-CD28 monoclonal antibody (clone 28.2) and secondary antibody conjugated to fluorescein isothiocyanate (FITC; B) before (baseline) and during intravenous infusion of heparin plus a polyunsaturated fatty acid–rich lipid emulsion (HL). After being washed 3 times, T lymphocytes were resuspended in phosphate-buffered saline before flow cytometric analysis (FACScalibur; Becton Dickinson, Montreal, Canada). No significant changes from baseline were detected in TCR or CD28 expression during HL infusion. The histograms represent the results of one experiment.
FIGURE 6.. T cell receptor–and CD28-mediated signal transduction in T cells before (baseline) and during intravenous infusion of heparin plus a polyunsaturated fatty acid–rich lipid emulsion (HL). Freshly isolated T cells were stimulated with anti-CD3 (clone Leu-4) in combination with anti-CD28 (clone 28.2). A: Western blot showing the reduction of total tyrosine phosphorylation in response to CD3 + CD28 costimulation in T cells isolated during HL infusion compared with baseline. Loading controls with the use of anti-LAT (linker of activated T cells) antibody are shown. ZAP, zeta chain–associated protein of 70 kDa. B: Western blot showing the marked impairment of LAT phosphorylation on CD3 + CD28 costimulation in T cells isolated during HL infusion compared with baseline, whereas p38 and ERK (extracellular signal-regulated kinase 1, or p42/44 mitogen–activated protein kinase) activation was only mildly impaired by HL infusion. Loading controls with the use of anti-LAT antibody are shown. C: Densitometric analysis of LAT phosphorylation on CD3 + CD28 costimulation compared with no stimulation (+ and –, respectively) in T cells at baseline () and during HL infusion (). n = 4 subjects. There was a significant difference between HL infusion and baseline (P < 0.001) and a significant interaction between treatment and stimulation conditions (P < 0.001) by two-factor ANOVA for repeated measures. *Significantly different from HL infusion, P < 0.05 (Scheffe post hoc test to adjust for multiple comparisons). AU, arbitrary units. D: Western blot showing the impairment of protein kinase B (Akt) phosphorylation 1 and 5 min after CD3 + CD28 costimulation in T cells isolated during HL infusion compared with baseline. Loading controls with the use of anti-Akt antibody are shown. E: Densitometric analysis of Akt phosphorylation before (0 min) and 1 and 5 min after CD3 + CD28 costimulation in T cells isolated at baseline () and during HL infusion (). n = 4 subjects. There was a significant difference between HL infusion and baseline (P < 0.001) and a significant interaction between time and treatment (P < 0.001) by two-factor ANOVA for repeated measures. *Significantly different from HL infusion, P < 0.05 (Scheffe post hoc test to adjust for multiple comparisons).
T cell cholesterol uptake and efflux
Cholesterol uptake was significantly reduced in T lymphocytes during HL infusion in comparison with baseline (35232 ± 2137 and 48312 ± 3128 cpm, respectively; P < 0.05). Incubation of T cells with MBCD significantly increased cellular cholesterol uptake at baseline (76172 ± 3498 cpm), whereas it had no effect on T cells isolated during HL infusion (41837 ± 2694 cpm). After 24 h of incubation with [3H]cholesterol, T lymphocytes isolated during HL infusion retained a smaller fraction of cellular cholesterol compared with T cells isolated at baseline (68.7 ± 6.9 and 93.0 ± 2.1% respectively; P < 0.05). In contrast, after 48 h in culture, T cells isolated during HL infusion retained a greater fraction of cellular cholesterol than did T cells isolated at baseline (29.8 ± 3.4% and 7.2 ± 1.6%, respectively; P < 0.05). In vitro cholesterol retention of T cells isolated during HL infusion returned toward baseline values 24 h after incubation with MBCD (94.2 ± 3.2% and 93.0 ± 2.1% in HL with MBCD compared with baseline, respectively) and 48 h after incubation with MBCD (4.2 ± 1.9% and 7.2 ± 1.6% in HL with MBCD compared with baseline, respectively).
DISCUSSION
In the present study, we provide evidence of an acute suppressive effect of enhanced intravascular triacylglycerol lipolysis on peripheral T cell activation and proliferation in vivo in healthy human subjects after HL infusion. Because Intralipid contains mainly linoleic acid, intravascular lipolysis of this lipid emulsion resulted in a significant increase in plasma linoleate, whereas a decrease in plasma palmitate and oleate was observed, which was due to inhibition of intracellular lipolysis induced by hyperinsulinemia (15). Observed changes in plasma FFA composition associated with a significant reduction in cell membrane fluidity and alteration of cholesterol exchange of T cells with the extra cellular environment. Moreover, we also showed for the first time that a 2-h increase in intravascular lipolysis of triacylglycerols rich in PUFA resulted in a marked inhibition of TCR- and CD28-mediated signal transduction. The decreased LAT phosphorylation and PKB/Akt activation on CD3 and CD28 stimulation was associated with irreversible impairment of T cell proliferation and enhanced T cell apoptosis. It is of note that the observed reduction in CD3 and CD28-induced T cell signaling and proliferation with HL infusion was not attributable to a change in expression of TCR or CD28. CD28 costimulatory signaling is essential for T cell proliferation because it induces the coalescence of lipid rafts by the reorganization of the cytoskeleton (22) and provides the metabolic signal necessary to sustain T lymphocyte proliferation (23). However, we found that some intermediate signaling events, including p42/44 MAPK and p38, were much less affected by HL infusion. This indicates that alternative pathways may also contribute to MAPK activation in vivo in T cells. The results of the present study support and extend the earlier observations of Stulnig et al (13), who used a similar in vivo experimental design and showed an alteration in the calcium metabolism of T cells.
One possible mechanism for the HL-mediated reduction in T cell activation and proliferation could be a shift in lipid raft properties due to cis unsaturated fatty acid uptake. Lipid rafts are found in large quantities in the plasma membranes of T cells and tend to include acylated membrane proteins such as LAT and p56Lck (24). In vitro exposure of Jurkat T cells to n–3 PUFAs results in PUFA uptake within lipid raft phospholipids as well as dissociation of LAT and p56Lck from lipid rafts and impaired phosphorylation of LAT (9, 11, 12). A diet rich in n–3 PUFAs was also associated with PUFA uptake by plasma membrane phospholipids and a reduction in the sphingomyelin content of T cell lipid rafts in mice (10). We can speculate that the enrichment of lipid raft phospholipids with cis unsaturated linoleyl chains from intravascular lipolysis of Intralipid may hinder the interaction of LAT with TCR and CD28 on stimulation of T cells, thereby preventing LAT phosphorylation and scaffolding of the protein complex as required for optimal intracellular signaling.
T cell membrane fluidity and the properties of lipid rafts are dependent on the cholesterol content of the cell membrane (25). We observed a marked reduction in T cell cholesterol uptake and efflux after in vivo HL infusion. Such a reduction in membrane cholesterol flux may contribute to an alteration of lipid raft properties in T cells and may be implicated in the reduction of T cell proliferation observed after HL infusion. The underlying mechanisms leading to the observed changes in cholesterol metabolism require further investigation.
Other factors may also contribute to our findings. Intravascular lipolysis of triacylglycerols may be responsible, at least in part, for changes in circulating lipoproteins such as HDLs, which, in turn, may affect lipid oxidation and thus immunomodulation. HL infusion in the present study is associated with an elevation in plasma glycerol and triacylglycerol concentrations. However, incubation of T cells with serum from Intralipid- or glycerol-infused subjects, without heparin stimulation of intravascular lipolysis, did not alter T cell activation in a previous study (13). The use of octreotide in the pancreatic clamp protocol is an unlikely explanation for our results, because in vitro incubation of peripheral T cells with octreotide did not affect cell proliferation on stimulation. A change in the level of other less prevalent fatty acids not measured in the plasma, such as stearic acid, cannot be ruled out in the present study. However, linoleate, palmitate, and oleate account for 80% of fatty acids contained in the lipid emulsion and for 75% of the total plasma FFA concentration during HL infusion. Therefore, the magnitude of change in the plasma concentration of these fatty acids, if present, would be expected to be much smaller than that of linoleic acid in our in vivo protocol.
In clinical trials, the administration of an intravenous lipid emulsion with total parenteral nutrition has been associated with reduced T cell proliferation, increased rate of infection, more prolonged requirement of mechanical ventilation, and the need for more prolonged intensive care and hospitalization (26, 27). Interestingly, IL-2 secretion and proliferation of peripheral human lymphocytes were significantly inhibited by in vitro incubation with Intralipid as opposed to an olive oil–based lipid emulsion relatively poor in PUFAs (28), which, again, suggests that the immunosuppressive effects of Intralipid may be caused by its high PUFA content. Our results provide further insight into the underlying mechanisms of these earlier findings and perhaps into the understanding of immune system modulation in pathophysiologic states that affect postprandial plasma lipid metabolism, such as obesity, insulin resistance, and type 2 diabetes (29). Peripheral mononuclear cell proliferation was reduced in obese subjects, with or without type 2 diabetes, and improved on weight loss induced by a hypocaloric diet (30, 31). A significant relation between peripheral T cell anisotropy and plasma triacylglycerol concentrations in humans was shown previously (32). The findings of the present study support a possible mechanistic link between abnormal postprandial plasma lipid and lipoprotein metabolism and dysfunction of cell-mediated acquired immunity in insulin-resistant states.
In conclusion, we have shown for the first time that enhanced in vivo intravascular lipolysis of a PUFA-rich lipid emulsion leads to modification of peripheral T cell plasma membrane fluidity, lipid raft organization, and impaired TCR- and CD28-mediated activation and proliferation in humans. Our results are clinically relevant to nutritional strategies in both intensive and postoperative care settings. Understanding the full physiologic and pathophysiologic implications of our results in normal and disordered postprandial lipid metabolism requires further investigation.
ACKNOWLEDGMENTS
AL helped execute the ex vivo experiments, analyze the data, and write the manuscript. AG and CF helped execute the ex vivo experiments. FF helped execute the in vivo experiments and some of the ex vivo protocols. ND helped execute the ex vivo experiments and analyze the data. ACC helped with the study design, execute the in vivo experiments, analyze the data, and write the manuscript. TF helped with the study design, execute the ex vivo experiments, analyze the data, and write the manuscript. None of the authors had a conflict of interest to report.
REFERENCES