Glucocorticoids (GC) are widely used anti-inflammatory agents known to suppress T cell activation by interfering with the TCR activation cascade. The attenuation of early TCR signaling events by these compounds has been recently attributed to a selective displacement of key signaling proteins from membrane lipid rafts. In this study, we demonstrate that GC displace the acyl-bound adaptor proteins linker for activation of T cells and phosphoprotein associated with glycosphingolipid-enriched microdomains from lipid rafts of murine T cell hybridomas, possibly by inhibiting their palmitoylation status. Analysis of the lipid content of the membrane rafts revealed that GC treatment led to a significant decrease in palmitic acid content. Moreover, we found an overall decrease in the proportion of raft-associated saturated fatty acids. These changes were consistent with a decrease in fluorescence anisotropy of isolated lipid rafts, indicating an increase in their fluidity. These findings identify the mechanisms underlying the complex inhibitory effects of glucocorticoids on early TCR signaling and suggest that some of the inhibitory properties of GC on T cell responses may be related to their ability to affect the membrane lipid composition and the palmitoylation status of important signaling molecules.
Glucocorticoids (GC)3 are small lipophilic molecules that regulate a large number of physiological processes, including immune responses. Despite the broad therapeutic use of these compounds as anti-inflammatory agents, their precise molecular mechanism of action has been difficult to identify, partly due to the multiplicity of the biological effects they exert on a large set of lymphoid cells (1). Upon binding to cytoplasmic receptors, the hormone-receptor complexes migrate to the nucleus where they regulate gene transcription both positively and negatively (2, 3). GCs are thought to exert their anti-inflammatory actions by negatively regulating the expression of numerous genes coding for proinflammatory cytokines and receptors (4). Inhibition of gene transcription has been shown to result from the ability of GC to interfere with the activity of numerous transcription factors, either by binding to negative regulatory elements in the promotor region or through protein/protein interactions, impeding the ability of these factors to positively direct gene transcription (5). In addition, to affect the distal signaling steps of gene transcription, several studies have uncovered the ability of GC to inhibit membrane proximal steps of the signaling cascade in immune cells (see Ref. 6 for review). Indeed, GC have been shown to affect the response of both T and B cells by inhibiting the early increase in intracellular calcium concentration induced upon Ag-receptor ligation, a key event promoting downstream activation of novel gene transcription (7, 8). We have recently confirmed these observations and shown that GC treatment inhibits the early tyrosine phosphorylation of several membrane-associated substrates including the immunoreceptor tyrosine-based activation motif-bearing, TCR-associated ζ-chain, the transmembrane adaptor linker for activation of T cells (LAT) molecule and the ζ-associated protein of 70 kDa tyrosine kinase. Further analysis of GC-treated T cell hybridomas revealed that hypophosphorylation of membrane-associated substrates was not due to reduced tyrosine kinase expression and/or activity, but was rather correlated with an altered membrane compartmentalization of these signaling enzymes and their substrates (9).
Numerous observations have recently uncovered the important role of membrane domains in signal transduction (10, 11, 12, 13). In particular, TCR signaling has been shown to occur in plasma membrane domains referred to as rafts (14, 15, 16, 17, 18, 19, 20, 21, 22, 23). These lipid microdomains, also referred to as detergent-insoluble glycolipid-enriched membranes, glycolipid-enriched membranes, or detergent-resistant membranes, consist primarily of cholesterol and glycosphingolipids and are enriched in proteins and lipids containing saturated fatty acid residues (10, 24, 25). Of particular relevance, numerous signaling proteins, including both protein-tyrosine kinases of the Src family, (p56Lck and p59Fyn) and transmembrane adapter proteins such as LAT and phosphoprotein associated with glycosphingolipid-enriched microdomains (PAG) are preferentially located in these membrane domains in T lymphocytes (17, 20). The confinement of signaling molecules in these membrane subdomains appears to be required for efficient signal transduction, as suggested by experiments in which disruption of membrane rafts using cholesterol-binding compounds led to a profound inhibition of calcium responses to receptor ligation (14, 15). The present work was undertaken to better understand the mechanism by which glucocorticoids affect the early steps of TCR signaling. Using a model T cell line, we demonstrate in this work that GC affect both protein and lipid raft composition. Analysis of the membrane lipid composition revealed a significant decrease in palmitic acid content. This correlated with an altered palmitoylation status of the adapter proteins LAT and PAG, probably affecting their localization within membrane rafts. Taken together, these data suggest that GC affect T cell signal transduction by altering the protein content of membrane rafts, impeding the adequate clustering of tyrosine kinases with their substrates.
Materials and Methods
Cell culture, reagents, and Abs
The pigeon cytochrome c/IEk-specific 3B4.15 murine hybridoma cell line was described elsewhere (8). The human Jurkat E6-1 was obtained from American Type Culture Collection (Manassas, VA). Dexamethasone (DEX; soluble in ethanol), filipin, and methyl-β-cyclodextrin (MCD) were purchased from Sigma-Aldrich (St. Louis, MO). Cells were cultured in RPMI 1640 (Life Technologies, Grand Island, NY) supplemented with 5% FCS, nonessential amino acids, 2 mM L-glutamine, penicillin-streptomycin, and 5 × 10−5 M 2-ME. Primary and secondary Abs used in this study were specific for the following markers: LAT (rabbit serum 3023; kindly provided by Dr. L. E. Samelson, National Institutes of Health, Bethesda, MD); TCRβ (hamster mAb H57–957; BD Biosciences, Franklin Lakes, NJ); CD90.1 (mouse mAb OX7; kindly provided by Dr. P. Draber, Academy of Sciences of the Czech Republic, Prague, Czech Republic); CD45 (rat mAb MB23G2; American Type Culture Collection); PAG (rabbit serum, kindly provided by Dr. V. Horejsí, Academy of Sciences of the Czech Republic); phospholipase C-γ1 (rabbit serum; Santa Cruz Biotechnology, Santa Cruz, CA); CD3ζ (hamster mAb H146–968; kindly provided by Dr. R. T. Kubo, La Jolla Institute for Allergy and Immunology, San Diego, CA); HRP-coupled protein A (Sigma-Aldrich); HRP-coupled sheep anti-mouse IgG (Abcam, Cambridge, U.K.); and HRP-coupled goat anti-rat IgG (Southern Biotechnology Associates, Birmingham, AL).
Hybridomas were washed twice in FACS buffer (PBS containing 0.1% BSA and 0.01% (w/v) NaN3) and incubated for 30 min with FITC-labeled anti-CD3ε mAb (clone 145–2C11; Ref. 26), FITC-labeled anti-Thy1.2 mAb (clone 53-2.1; BD PharMingen, San Diego, CA), biotin-conjugated anti-TCRβ mAb (clone H57–957; BD Pharmingen), biotin-conjugated anti-CD45 (clone I3/2.3; Southern Biotechnology Associates); and biotin-conjugated cholera-toxin B (Sigma-Aldrich). Cells were washed twice and incubated for 15 min with Cy-Chrome-streptavidin (BD PharMingen). For filipin staining, cells were washed twice in PBS and then fixed for 20 min at room temperature with freshly prepared paraformaldehyde (4% in PBS (w/v)). After washing and quenching with PBS-glycine (1.5 mg/ml) for 10 min at room temperature, cells were incubated with filipin (0.1 mg/ml in PBS) for 15 min. Cells were then washed and analysed on a FACSort using CellQuest software (BD PharMingen).
Isolation of lipid rafts
Lipid rafts were isolated from 3B4.15 T cell hybridomas as described (9, 27, 28). Briefly, T cell hybridomas (5 × 107/ml) were lysed on ice in 1 ml of MNE buffer (Mes 50 mM, pH 6.5, 150 mM NaCl, 5 mM EDTA) containing 1% Triton X-100 w/v, 1 mM Na3VO4, 1 mM PMSF, and 1 mM NaF. Lysates were gently sonicated (five bursts of 5 sec at 5W) and cleared at 5000 rpm for 10 min at 4°C. The supernatant (1 ml) was mixed with an equivalent volume of 80% sucrose made with MNE buffer and transferred to an ultracentrifuge tube (5 ml). This solution was carefully overlaid with 2 ml of 35% sucrose followed by 1 ml of 5% sucrose (both prepared in MNE buffer) and the tubes were placed in a cooled AH650 Sorvall rotor (Newtown, CT). The gradients were then ultracentrifuged at 39,000 rpm for 16 h at 4°C. After centrifugation, 0.5-ml fractions were collected from the top of the gradient. These aliquots were mixed with equal volume of 2× SDS buffer and analyzed by SDS-PAGE. Fractions were then subjected to SDS-PAGE under reducing conditions except for CD90 and CD45 that required nonreducing conditions for adequate detection. Blots were revealed using the appropriate primary and HRP-secondary reagents and were developed by chemiluminescence (ECL; Amersham Biosciences, Little Chalfont, U.K.).
Fatty acid and cholesterol analysis
3B4.15 T cell hybridomas and Jurkat cells were incubated overnight with DEX (1 μM). Rafts from 5 × 107 cells (for 3B4.15 hybridomas only) were isolated as described in the above section. For bulk membranes, cells were incubated for 10 min on ice with hypotonic buffer (10 mM HEPES, pH 7.4, 42 mM KCl, 5 mM MgCl2) and passed five times through a 27G needle. Membranes were pelleted from postnuclear supernatants as described (29, 30). Rafts and bulk membranes were diluted in Tris buffer (20 mM Tris-HCl, pH 8.2, 140 mM NaCl), pelleted (100,000 × g, 30 min, 4°C), extracted, and transesterified to methyl esters by a one-step reaction modified from Ref. 31 as described in detail in Ref. 30 . Briefly, freeze-dried samples were dissolved in methanol (Merck)/benzene (Riedel-de Haën, Seelze, Germany), 4:1 (v/v), including 100 ng of heptadecanoic acid as internal standard, followed by methanolysis. Fatty acid methyl esters were separated by gas chromatography on a DB-23 (J & W Scientific, Folsom, CA) with fatty acid methyl esters of the highest available quality (Sigma-Aldrich) used as standards. The lipid extracts prepared for fatty acid analysis were also used for detection of cholesterol by gas chromatography and electron-impact ionization mass spectroscopy (EI-MS, m/z 386, M+). Lipid extracts were directly injected in a Hewlett-Packard (Boise, ID) GC-MS 5973 system equipped with a CP Sil5 25 m × 0.2 mm × 0.12 μm capillary column (Chrompack, Middleburg, Netherlands) as described in Ref. 30 . Heptadecanoic acid (m/z 284, M+) was detected by EI-MS as internal standard in parallel with cholesterol. Because heptadecanoic acid served as internal standard for both fatty acid and cholesterol quantitation, the amount of cholesterol was expressed as the mole percent of the sum of all fatty acids, which was used as an estimate of the total amount of the lipid phase of the membrane. The significance of differences in fatty acid and cholesterol content between rafts and bulk membranes of control and DEX-treated 3B4.15 T cell hybridomas and Jurkat T cells was calculated by paired Student’s t test and p < 0.05 was considered statistically significant.
Analysis of protein palmitoylation
Cell labeling was carried out as described previously (32, 33). Briefly, prior to labeling, T cell hybridomas (5 × 107 cells/condition) were cultured overnight at 37°C in DEX (1 μM) or ethanol-supplemented media and then labeled for 4 h with 25–50 μCi/ml 16-125I-labeled iodohexadecanoic acid or 1 mCi/ml (9, 10) [3H]palmitic acid (Amersham Biosciences) in DMEM containing 2% dialyzed FBS and 0.5% defatted BSA (Sigma-Aldrich). Cells were harvested and lysed in radioimmunoprecipitation buffer containing 50 mM Tris (pH 8.0), 500 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 1 mM Na3VO4, 5 mM NaF, and 1 mM PMSF. Cell lysates were immunoprecipitated with anti-LAT or anti-PAG polyclonal Abs and immunocomplexes were analyzed by SDS-PAGE,following by phosphorimaging or Western blotting with anti-LAT or anti-PAG Abs. For tritium-labeled proteins, gels were fixed in acetic acid(10%)/methanol (30% (v/v)), incubated in enhancer (Amplify; Amersham Biosciences), dried, and exposed to Ray MaxTritium films (ICN Pharmaceuticals, Costa Mesa, CA) for 9 months at −80°C.
Preparation of rafts for fluorescence anisotropy measurements
Rafts from control or DEX-treated 3B4.15 cells (5 × 107 cells/condition) were isolated as described (34). Cells were washed twice in buffered saline solution (BSS: 135 mM NaCl, 5 mM KCl, 1 mM MgCl2, 1.8 mM CaCl2, 5.6 mM glucose, 20 mM HEPES, pH 7.4). Cells in BSS (4 × 107 cells/sample) were lysed by mixing 1:1 with 2× ice-cold lysis buffer containing 0.1% Triton X-100 (v/v) for 10 min. The cell lysates were then diluted with an equal volume of 80% sucrose (w/v) in HEPES/saline buffer (25 mM HEPES, 150 mM NaCl, 2 mM EDTA, pH 7.5) at 4°C. Twenty milliliters of this lysate were added to each polycarbonate centrifuge tube (Beckman Coulter, Palo Alto, CA). This solution was carefully overlaid with 5 ml of 35% sucrose followed by 5 ml of 5% sucrose (both prepared in HEPES/saline buffer) and the tubes were placed in a cooled SW28 rotor. The gradients were then ultracentrifuged at 100,000 × g (20,000 rpm) for 16 h at 4°C. After centrifugation, the upper band located at the interface between the 5 and 35% sucrose layers was collected and contained the lipid rafts. The purification of the raft fractions was assessed by Western blotting using the CD90 and CD45 markers (as shown, see Fig. 2⇓A). The material was collected and diluted 2- to 3-fold in PBS/2 mM EDTA and centrifuged at 27,000 × g (20,000 rpm) in a 50Ti rotor (Beckman Coulter) at 6°C for 1 h. The pelleted rafts were resuspended in 1 ml of BSS.
Rafts (4 × 107 cells/ml) resuspended in 1 ml of BSS were labeled with the fluorescent probe 2-[3-(diphenylhexatrienyl)propanoyl]-1-hexadecanoyl-sn-glycero-3-phosphocholine (DPH-PC; Molecular Probes, Eugene, OR) by adding the probe in methanol at a molar ratio of <1:100 probe-phospholipid and incubating at 37°C for 5 min (34). Labeled rafts were centrifuged in a SW60Ti rotor (Beckman Coulter) at 27,000 × g (16,500 rpm) for 45 min at 4°C and resuspended in BSS. 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) and 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC) multilamellar liposomes were described previously (35, 36).
Fluorescence anisotropy measurements
Fluorescence anisotropy measurements were performed in a SLM 8000C spectrofluorometer with Glan-Thompson polarizers placed in T-geometry. Excitation was performed at 360 nm and emission was recorded at 430 nm (34). For each experiment, 1 ml of sample in BSS was placed in a 10 × 10 × 45 mm acrylic cuvette (Kartell, Noviglio, Italy), and the mixture was stirred continuously (small magnetic bar) in a thermostatic chamber. Polarization measurement was made by simultaneously measuring the vertical and horizontal components of the polarized emission. Correction for the background polarization due to sample turbidity was made using unlabeled rafts or liposomes. The ratio of intensities in the vertically and horizontally polarized detectors was measured with vertically and horizontally polarized excitation, giving respectively the Rvert and Rhoriz ratios. Polarization (P) was calculated as P = (Rcorr − 1)/(Rcorr + 1) where R = Rvert/Rhoriz. Anisotropy was derived from P using r = 2P/(3 − P).
DEX treatment does not affect expression of several membrane-associated molecules
Using a model T cell line, we have recently demonstrated that inhibition by the synthetic glucocorticoid analog DEX of the early phosphorylation events in response to TCR stimulation correlated with an altered submembrane localization of important signaling molecules (9). This finding led us to postulate a possible effect of GC on both protein and lipid composition of membrane rafts. To exclude the possibility that GC affected the total membrane-associated expression of important signaling molecules and receptors, the phenotype of a murine T cell hybridoma was analyzed by flow cytometry following incubation in DEX-supplemented media (Fig. 1⇓). Expression of CD127 (IL-7Rα), a receptor known to be up-regulated by GCs (37) was used as positive control in this experiment. DEX (1 μM, overnight incubation) failed to significantly affect the expression of cell surface-associated molecules belonging to the TCR complex (CD3ε and TCRβ, see Fig. 1⇓). Similarly, expression of molecules associated with protein kinase activity (CD4, data not shown), or phosphatase activity (CD45, see Fig. 1⇓) was not affected by DEX. Notably, the expression of molecules associated with cholesterol-enriched membrane domains was not affected by GCs. The GPI-linked protein Thy1 (CD90) is a very abundant protein associated with the outer leaflet of the raft domains of the plasma membrane of murine T cells (17). Similarly, the GM1 ganglioside is an integral component of these microdomains which can be detected with the B subunit of cholera toxin. DEX treatment did not significantly affect the expression of these raft-associated molecules. Finally, cell-associated free cholesterol was labeled using the cholesterol-binding drug filipin (38). As expected, cholesterol depletion of the cell membrane by MCD resulted in the dose-dependent reduction in filipin staining (see Fig. 1⇓). In marked contrast, DEX treatment failed to down modulate filipin staining, suggesting that DEX did not affect cholesterol levels in T cell hybridomas.
DEX treatment displaces LAT and PAG from lipid rafts and inhibits their palmitoylation
Although DEX did not affect expression levels of several signaling molecules, it affected their submembrane distribution, as illustrated in Fig. 2⇓. T cell hybridomas were solubilized in Triton X-100, and lysates were subjected to overnight ultracentrifugation over a sucrose gradient. Equal volumes of all collected fractions were analyzed by immunoblotting using Abs to known components of the membrane and to the adaptor molecules PAG and LAT. The raft-containing fractions (fractions 2–4) were shown to be enriched in GPI-linked CD90 molecules, while the expression of molecules known to be excluded from rafts (such as CD45) or known to be in the cytoplasm of resting cells (PLC-γ1, data not shown) was only detected in the detergent-soluble fractions (fractions 6–9). Note that in this experiment, DEX led to a slight increase in raft-associated CD90 expression, in keeping with the flow cytometry data shown in Fig. 1⇑. In agreement with our previous observations, a selective loss of PAG and LAT expression from the rafts was observed in DEX-treated cells (see Fig. 2⇓A). A densitometric analysis on the immunoblots revealed a significant reduction of the relative amount of LAT and PAG recovered in the raft-containing fractions (fractions 1–4) from DEX-treated cells (from 45 to 18% for LAT and from 30 to 8% for PAG) while the amount of the same proteins increased in fractions 5–9 containing detergent-soluble material from DEX-treated cells (from 55 to 82% for LAT and from 70 to 92% for PAG). In accordance with previous observations (27), the p36 isoform of LAT was predominantly found in raft fractions whereas both p36 and p38 isoforms were found in the Triton-soluble fractions. Moreover, and in keeping with our previous work (9), a time course experiment showed that the DEX-induced displacement of PAG and LAT required a long incubation time (16 h) and needed de novo protein synthesis (data not shown).
Recently, the role of posttranslational modification of signaling proteins for adequate membrane localization has been documented. In particular, the addition of saturated fatty acids (generally palmitic acid) to transmembrane proteins has been shown to be required for adequate targeting to the lipid rafts. Site-directed mutagenesis studies have revealed the important role of palmitoylation at two cysteine residues near the transmembrane region (Cys26 and Cys29) of LAT in targeting this adaptor molecule to the rafts (27, 39). The PAG/Cbp adaptor protein has also been shown to be palmitoylated (40, 41). Based on these studies, we hypothesized that displacement of LAT and PAG from lipid rafts could be a consequence of altered protein acylation. Palmitoylation levels of LAT and PAG were assessed following incubation of hybridomas with [3H]palmitate or with the iodinated palmitate analog 16-iodohexadecanoic acid. Cell extracts were immunoprecipitated with anti-LAT and anti-PAG Abs and immunocomplexes were analysed by Western blotting to assess protein content and by phosphorimaging (for 125I labeling) or fluorography (3H labeling). As depicted in Fig. 3⇓, DEX significantly decreased the amount of radioactive palmitate incorporated in both immunocomplexes, suggesting that it significantly affected the acylation status and/or turnover of these two important adaptor proteins. Western blot experiments indicated that equivalent amounts of proteins were immunoprecipitated from control and DEX-treated cells.
DEX treatment alters raft lipid fatty acyl composition
To test whether glucocorticoids could affect protein acylation by altering the membrane lipid environment, we isolated bulk membranes and rafts from DEX and control-treated 3B4.15 T cell hybridomas and analyzed their fatty acyl composition (Table I⇓). A subline of the human Jurkat T cell line lacking GCR expression was used as negative control and was found to display a similar membrane fatty acid composition as the murine 3B4.15 cell line. As expected from previous reports, rafts from control cells were highly enriched in the saturated palmitic (C16:0) and stearic (C18:0) acids, representing ∼70% of the total fatty acid population in this fraction. In contrast, monounsaturated fatty acids like palmitoleic (C16:1 (n-7)) and oleic acids (C18:1 (n-9)) were considerably less abundant in the detergent-resistant fraction when compared to bulk membranes, representing <10% of total fatty acid compared to 30% of the fatty acid population of total membrane extracts. Polyunsaturated fatty acids like arachidonic acid (C20:4 (n-6)) were also reduced in rafts compared to bulk membranes. These data confirm that the raft-containing fraction is highly enriched in saturated fatty acids.
Treatment of the GCR-expressing T cell line with DEX led to a significant reduction in palmitic acid content, both in rafts and in the total membrane fraction. This loss in palmitic acid was paralleled by increased levels of stearic (C18:0) and oleic (C18:1 (n-9)) acid. Analysis of fatty acid content revealed that DEX treatment decreased the overall proportion of saturated fatty acids and increased the amount of monounsaturated fatty acids present in both lipid rafts and bulk membranes. Moreover, DEX treatment induced a small rise (from 4 to 6%) in the proportion of polyunsaturated fatty acids present in rafts. DEX did not affect the cholesterol content of the raft fraction, in agreement with the flow cytometry analysis presented in Fig. 1⇑. Finally, DEX failed to affect the fatty acyl composition of the GCR-negative Jurkat cell line. In conclusion, the major effect of DEX on membrane fatty acid composition is the reduction in palmitic acid present in both the rafts and the total membrane fractions.
DEX treatment reduces membrane fluidity
The high concentrations of saturated acyl chains and cholesterol found in rafts have been shown to reduce the membrane fluidity of these membrane domains when compared to the detergent-soluble membrane fraction. Fluorescent anisotropy can be used to evaluate the fluidity of biological membranes (Ref. 34 and references therein). Using two different probes (DPH-PC and N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)-1,2-dihexadecanoyl-sn-glycero-3-phosphoethanolamine), Gidwani et al. (34) have recently shown that the fluidity of both model (liposomes) and biological membranes is sensitive to cholesterol-dependent ordering. To evaluate whether the DEX-induced change in fatty acid composition led to a change in raft membrane fluidity, these cholesterol-enriched domains were isolated from control and DEX-treated hybridomas. Liposomes containing phospholipids of a distinct nature were used to set internal standards. Liposomes containing the disaturated DPPC showed very high anisotropy values and a transition temperature around 42°C (Fig. 4⇓). In contrast, liposomes containing the monunsaturated DOPC showed low anisotropy values and no detectable transition in this temperature range. Rafts isolated from control 3B4.15 T cell hybridomas by sucrose gradients and labeled with DPH-PC showed high anisotropy values. Rafts isolated from DEX-treated cells showed significantly lower anisotropy values. The mean anisotropy value at 37°C from six independent experiments was 0.228 ± 0.004 for control cells and 0.203 ± 0.002 for DEX-treated cells (p < 0.005, n = 6). Experiments performed on Jurkat T cells showed no significant effects on cells insensitive to glucocorticoids (data not shown). These observations are in agreement with the findings reported in Table I⇑, and indicate that the modification in the fatty acid composition induced by DEX correlates with an increase in membrane fluidity.
GCs represent potent immunosuppressive drugs capable of inhibiting the activity and effector function of most classes of immune cells. Although the prevailing concept is that GC modulate the expression of many proinflammatory genes at the transcriptional level, a growing body of data indicate that these compounds also affect a membrane proximal step of the signaling cascade initiated by Ag and cytokine receptors (see Ref. 6 for review). A detailed analysis of the early signal transduction events occurring in a model T cell line revealed an important role of membrane subdomains in mediating the immunomodulatory effects of GCs. Indeed, these studies revealed that attenuation of T cell signaling was not a consequence of reduced expression and/or functional activity of important kinases and/or substrates, but was rather related to a selective displacement of these signaling proteins from lipid rafts. Inhibition of TCR signaling requires long-term exposure to GC (at least 6 h), GCR binding and de novo protein synthesis, suggesting that this novel inhibitory property of GC is a consequence of novel gene transcription initiated by the hormone-receptor complex and not a mere consequence of GC insertion into the plasma membrane. It is noteworthy that the ability of GC to alter membrane localization of signaling molecules could be evidenced in resting T cell hybrids, i.e., prior to TCR stimulation.
The aim of the present study was to further characterize the mechanism by which GC interfere with membrane compartmentalization of signaling molecules in T lymphocytes. Based on the literature, at least three possible mechanisms could be envisioned to explain these observations: 1) altered protein acylation, leading to dispersion of resident raft proteins into the detergent soluble membrane fraction and/or 2) alterations in membrane lipid composition such as increased proportion of membrane-associated polyunsaturated fatty acids, as suggested by the early TCR signaling defect observed in T cell lines enriched with polyunsaturated fatty acids, and/or 3) the reduction of membrane cholesterol content (14, 15, 30, 32, 33, 42, 43). Our observations strongly suggest that GC affect TCR signaling by altering protein acylation status or turnover of raft-associated proteins. Numerous observations have revealed the key role of lipid acylation in TCR-mediated signaling. In particular, studies performed using mutant proteins and specific inhibitors of palmitoylation have demonstrated the important role of protein acylation in membrane targeting, raft association and function of several important signaling proteins (27, 44, 45, 46). The addition of myristate and palmitate to the Src family kinases Fyn and Lck is responsible for their efficient membrane localization and signaling function (47). Similarly, the adaptor protein LAT is palmitoylated and LAT acylation is necessary for raft targeting (27, 48). Mutation of the palmitoylation sites within LAT prevents its partitioning to rafts and results in impaired TCR signaling (27) as evidenced by decreased calcium mobilization, activation of the ras pathway, CD69 up-regulation and NFAT-mediated gene transcription (27, 39). Moreover, even moderate displacement of LAT from lipid rafts similar to that induced by DEX in our experiments was recently shown to significantly inhibit T cell signaling (43). The adaptor molecules LAT and PAG were chosen herein due to their high expression levels in the cell line used in this study. In addition, tyrosine phosphorylation of LAT and PAG is sensitive to GC-mediated inhibition (see Ref. 9 and data not shown). DEX was found to inhibit the palmitoylation status of these molecules, which likely accounts for their altered membrane compartmentalization (see Fig. 2⇑B). Of note, GC affected the membrane localization of other signaling molecules (including the src family kinases, see Ref. 9), but labeling of these proteins with radioactive palmitate or palmitate analogs was insufficient for adequate characterization (data not shown).
Protein palmitoylation involves the posttranslational attachment of palmitate to cysteine residues of proteins. Detailed understanding of this reversible modification has been limited by the lack of knowledge of the transferases and thioesterases that are involved in this important regulatory step. However, several studies indicate that the palmitoyltransferase activity is membrane-bound upon subcellular fractionation (Refs. 49, 50, 51 ; Ref. 51 also available online at http://stke.sciencemag.org/cgi/content/full/sigtrans;2000/63/re1), suggesting that the palmitate moiety that is transferred to proteins belongs to the pool of membrane-associated fatty acids. DEX was shown to cause a significant reduction in the membrane-associated palmitate (see Table I⇑). Although the effect of GC on metabolism varies with species, tissue type and state of differentiation, their ability to affect lipid biosynthesis, through for example modulation of fatty acid synthase levels (a key enzyme in the de novo fatty acid biosynthesis) has been previously described (52). Despite numerous efforts, we have not been able to identify a potential target for GC in lipid biosynthesis at the transcriptional level. For example, mRNA levels for the fatty acid synthase and the acetyl-CoA carboxylase, which also participates in the biosynthesis pathway leading to palmitate were not affected upon GC exposure (data not shown). In any event, our observations indicate that DEX treatment leads to altered palmitate membrane raft content. Therefore, it is tempting to speculate that reduced palmitate would alter the posttranslational modification of multiple signaling molecules, leading to reduced palmitoylation and subsequent dispersion of these proteins in the detergent soluble membrane fraction.
Altered localization of proteins in this setting could also be an indirect consequence of altered lipid raft composition. The GC-induced decline in the proportion of saturated fatty acids was accompanied by a significant rise in unsaturated fatty acids. Lipid rafts contain predominantly saturated fatty acids (mainly myristic and palmitic acids) that are very important to maintain these microdomains into a liquid-ordered phase (12, 20, 53). GC did not lead to altered cholesterol content of lipid rafts (see Fig. 1⇑ and Table I⇑) nor total cell-associated cholesterol, the latter measured by an in vitro enzymatic assay (data not shown). Nevertheless, the membrane fluidity of the rafts appeared to be increased in response to DEX treatment. Thus, changes in membrane lipid composition and hence its biophysical properties could represent an additional mechanism by which GC affect the partitioning of palmitoylated proteins into rafts.
In conclusion, our data provide evidence that GC could mediate some of their immunomodulatory effects by altering the raft lipid composition. Because raft localization of various signaling proteins is crucial for TCR signaling, small acylation modifications or small changes in the membrane fatty acyl content could mediate some of the inhibitory effects of GC and attenuate early signaling events. The precise identification of the molecular targets of DEX in this experimental setting will require a detailed analysis of the genes whose expression is modulated upon GC treatment, a study presently underway in our laboratory. In any event, the reversible nature of cysteine acylation makes the palmitoylation-depalmitoylation cycle a possible target for controlling protein function and/or compartmentalization, and the present work represents the first example of hormonal regulation of this regulatory step.
We would like to thank Drs. L. E. Samelson, R. T. Kubo, P. Draper, and V. Horejsi for providing reagents; Drs. Deborah Brown, Thomas Harder, Susan Pierce, Luc Berthiaume, Jonathan Ashwell, Hannes Stockinger, and Anthony Magee for interesting discussions and comments on the manuscript; Dr. Erika Baus for careful review of the manuscript; and Thomas Sigmund for skilful technical assistance in lipid analyses. We are also very grateful to Dr. David Holowka for critical discussions on the anisotropy experiments.
↵1 This work was funded by the Belgian Program in Interuniversity Poles of Attraction initiated by the Belgian State Prime Minister’s Office, Science Policy Programming, by a Research Concerted Action of the Communauté Française de Belgique (to F.V.L., J.U., and O.L.), and by a grant from the Fonds National de la Recherche Scientifique/Télévie, Belgium (to F.V.L.). This work was also funded by the Austrian Science Foundation (P13507-Med) and the International Cooperative Program of the Austrian Federal Ministry for Education, Science, and Culture (both to T.M.S.), and by National Institutes of Health Grant GM75966 (to M.D.R.).
↵2 Address correspondence and reprint requests to Dr. François Van Laethem at the current address: Experimental Immunology Branch, National Cancer Institute, Building 10, Room 3N119, Bethesda, MD 20892. E-mail address:
↵3 Abbreviations used in this paper: GC, glucocorticoid; LAT, linker for activation of T cells; PAG, phosphoprotein associated with glycosphingolipid-enriched microdomains; DEX, dexamethasone; MCD, methyl-β-cyclodextrin; PLC, phospholipase C; BSS, buffered saline solution; DPH-PC, 2-[3-(diphenylhexatrienyl)propanoyl]-1-hexadecanoyl-sn-glycero-3-phosphocholine; DOPC, 1,2-dioleoyl-sn-glycero-3-phosphocholine; DPPC, 1,2-dipalmitoyl-sn-glycero-3-phosphocholine.
- Received September 9, 2002.
- Accepted December 20, 2002.
- Copyright © 2003 by The American Association of Immunologists