Ethics statement: All the treatments of animals (mice) in this research followed the guidelines of the Institutional Animal Care and Ethics Committee at Eotvos Lorand University that operated in accordance with permissions 22.1/828/003/2007 issued by the Central Agricultural Office, Hungary and all animal work was approved by the appropriate committee.
Naive CD4+ T helper (Th) lymphocytes maturate in the thymus and then emigrate to lymph nodes, spleen, and periphery where upon meeting with foreign antigens become activated and differentiate to polarized effector Th cell subpopulations such as Th1, Th2, or Th17. The CD4+ T cell subpopulations were first identified in mouse, then in human (1–3), as Th1 and Th2 subsets, which control the sensitive balance of cellular and humoral immune response. They differ in their cytokine production profile following antigen stimulation, namely IFNγ or IL-4 secretion by Th1 and Th2 cells, respectively (2). Th1 cells are engaged with immunity to intracellular microorganisms, whereas Th2 cells for the immune response to extracellular pathogens including helminths (4). Abnormal activation of Th1 cells is coupled with some organ-specific diseases while overbalanced Th2 activity is often responsible for allergic inflammatory diseases and may promote or amplify autoimmune diseases, as well (5, 6). The direction of naive T cells' polarization depends on multiple factors such as the type of infectious agent/route of infection, the nature of the antigen presenting cell (APC) and most of all the dominant local cytokine milieu (7–11). The fundamental differences in the membrane organization and signaling properties of differentiated Th1 and Th2 cells, however, still remained mostly unclear.
Many recent studies focused therefore on the molecular background of T cell polarization. As one of the key elements, the NFAT transcription factor (6, 12) was directly implicated in activation of several cytokine genes; its activation was proposed to direct the cytokine gene activation during TCR-mediated activation in a Ca2+-dependent manner (13–15). Thus, the upstream signaling events generating/regulating an elevated cytosolic Ca2+ level can all quantitatively define the response of CD4+ Th cells to antigens. The antigen receptors, the signal transducing and amplifying molecules, or the ion-selective channel proteins are all embedded in the lipid bilayer of T cell membrane. The membrane's lipid composition or their compartmentation into cholesterol-rich membrane microdomains were already implicated to control some of the above molecules and functions (16, 17).
Among others, the IFNγ-receptor was found colocalized with TCR in the membrane rafts during activation of T cells, but this colocalization is disrupted by IL-4 signaling (18), suggesting that the T cell polarization involves an active “cross-inhibition” by the other lineage direction. Th2 cells were reported e.g. to be more dependent on CD4 involvement in signaling and using ITK rather than other TEC kinases to amplify Ca2+-flux through activation of PLCγ (19). The Trpm4 sodium channel, mediating membrane depolarization (20, 21), increased expression of Ca2+-activated K+ channels (KCa) in Th1 cells and a faster clearance of cytosolic calcium from Th2 cells (22) were also implicated in the differential Ca2+-response of Th1 and Th2 cells. A recent study reported that plasma membrane lipid order may also be a key differentiation factor leading to polarized T cell signaling/function (23).
Thus, the aim of the present work was to reveal the differences in the plasma membrane composition/organization of Th1 and Th2 cells and the coupling between the plasma membrane organization and the downstream transcriptional regulation of polarized cytokine production by Th cells. Recently we have developed murine Th hybridoma cell lines by cell fusion technology for studying Th cell differentiation. This model system, consisting of three selected hybridomas with Th0, Th1 and Th2 effector properties, was already successfully used to show adenosine A2A receptor-mediated effects on Th cell differentiation, effector functions and cell death (24, 25).
Here we compared the membrane expression patterns of raftophilic T cell accessory proteins and lipids in the polarized T cells using flow and image cytometry. Our data suggest that plasma membrane compartmentation of the TCR/CD3 antigen receptor complex and Kv1.3 ion channels, as well as the differential expression of raftophilic accessory proteins (e.g. CD4, CD48, CD59) together may lead to the observed differences in the calcium response of polarized Th cells. This in turn may cause different residence time for the nuclearly translocated NFAT, as observed here. In addition, the NFAT residence time directly depended on membrane cholesterol level (raft integrity). Since the in vivo Th1/Th2 balance is also strongly influenced by their death/survival equilibrium, the sensitivity of the polarized Th cell lines to TCR-mediated, activation induced cell death (AICD) was also investigated. Polarized cells, especially Th1, were more sensitive than their Th0 precursor. Our results suggest that the difference in raftophilic lipid and protein composition of the plasma membrane, the differential compartmentation of the TCR/CD3 receptor complex and Kv1.3 ion channels, coupled to altered NFAT activation through Ca2+-signals, may be an important axis determining polarized T cell effector function.
MATERIALS AND METHODS
Generation of the Th Hybridoma Cell Lines
Inguinal and popliteal lymph nodes were isolated from Balb/c mice 11 days after immunization with FITC-KLH in complete Freunds Adjuvant. Red blood cell-depleted cell suspension was reactivated with 1 μg/ml FITC-LPS (Sigma, St. Louis, MO, USA) and 25 μg/ml FITC-KLH in the presence of 20 ng/ml mouse rec-IFNγ (R&D, Minneapolis, MN,USA) and 5 μg/ml anti-IL-4 (11B11 rat monoclonal IgG, prepared in our lab). Three days later CD4+ T-cells, enriched by depleting CD8+ and FcγR+ cells with panning technique using 53.6 and 2.4G2 rat monoclonal antibodies (clones from ATCC, Middlesex, UK), were fused with BW 1100.129.237 TCRαβ− thymoma cells in the presence of polyethylene glycol (Sigma) then cultured in HAT completed DMEM selective medium (Sigma). Th0 (2.13), Th1 (7.6) and Th2 (7.5) hybridoma clones were selected on the basis of their activation-induced cytokine production. Their CD3, CD4, and CD28 expression and cytokine producing capacity were checked regularly.
Th1 and Th2 Directed Activation of Freshly Prepared CD4+ Mouse Spleen T Cells
Balb/c spleen cells, enriched in CD4+ T cells using CD8 and FcγR specific antibodies in panning technique (Supporting Information Fig. 1), were activated with 1 μg/ml anti-CD3 (145-2c11) and mitomycin C (Sigma) treated LK-35 mouse B cells in the presence of 5 ng/ml mouse rec-IFNγ, 5 ng/ml mouse rec-IL-12, and 10 μg/ml anti-IL-4 or 10 ng/ml mouse rec-IL-4 and 10 μg/ml anti-IFNγ (R4-6A2), respectively. On the fourth day of culture cells were washed and recultured without anti-CD3 and any added cytokines but in the presence of 10% mouse IL-2 (IL-2-Sp2 (26) cell culture supernatant) and the anti-cytokine antibodies as before. Expression of cell surface molecules and activation induced cytokine production were detected 3 days later.
Measurement of Protein and Lipid Expression of the Hybridomas
Cell surface protein expression of the hybridomas was measured by flow cytometry (Supporting Information Fig. 1) using FACSCalibur or FACSAria III cytometers (Becton Dickinson, Franklin Lakes, NJ USA). Cells were stained directly with monoclonal anti-CD3-FITC, anti-CD4-FITC (Immunotools, Friesoythe, Germany), or with anti-cholesterol mouse IgG3 (AC8) followed by goat anti-mouse IgG-FITC (Southern, Birmingham, Alabama, USA). CD48-FITC (Santa Cruz Biotechnology, CA, USA) and CD59 mAbs (Santa Cruz Biotechnology) were used to stain the corresponding antigen proteins. Alexa-488 (or 647) conjugated choleratoxin-B (Molecular Probes-Invitrogen, Carlsbad, CA, USA) was used for the detection of GM1 ganglioside expression.
Measurement of Cytokine Production of the Th Hybridomas
Hybridoma cells were stimulated with 5 μg/ml anti-CD3 (145-2c11) or 5 μg/ml anti-CD3 and LK-35 antigen presenting cells (2.5 × 105 cell/ml) or 5 μg/ml Concanavalin A (Sigma) for 18 hr. Supernatant cytokine contents of activated and control cell cultures compared to standard dilutions of the appropriate cytokines were detected in sandwich ELISA using high-binding ELISA plates (Costar, Lowell, USA), anti-IL-4 (11B11), anti-IFNγ (R4–6A2) or anti-GM-CSF (BD Biosciences, Bedford, MA, USA) as capturing and biotin conjugated anti-IL-4, anti-IFNγ (Immunotools), or anti-GM-CSF (BD Biosciences) as detecting antibodies. IL-2 production was measured by a bioassay based on the IL-2 dependent CTLL-2 (ATCC) survival detected by MTT (Sigma) assay system.
Flow Cytometric Analysis of Calcium Response
Typically 5 × 106 cells were loaded with Fluo3-AM calcium indicator (Molecular Probes-Invitrogen) in 1 ml Hank's A buffer pH 7.4 (143 mM sodium chloride, 1 mM sodium sulphate, 5 mM potassium chloride, 1 mM sodium hydrogen phosphate, 0.5 mM magnesium chloride, 1 mM calcium chloride, 5 mM glucose, 10 mM Hepes) for 60 min at 37°C then for another 30 min in 10 ml volume. Cells were washed twice and measured in Hank's A buffer. Propidium iodide was used to gate out dead cells. The calcium response profiles were analyzed by a method and software described earlier (27).
Confocal Laser Scanning Microscopy
For GM1 ganglioside and CD3 colocalization cells were stained with choleratoxin-B (Molecular Probes-Invitrogen) and monoclonal anti-CD3-FITC (Immunotools) for 15 min on ice. For NFAT1 translocation cells were activated with 2.5, 5, or 15 μg/ml anti-CD3 for 15, 30 or 60 min at 37°C then following fixation/permeabilization cells were stained with mouse monoclonal anti-NFAT1 (Abcam, Cambridge, UK) or anti-NFAT2 (Abcam) and goat anti-mouse IgG1 Alexa-488 (Molecular Probes-Invitrogen) as described earlier (28). Nucleus was counterstained with 10 μM Draq5 (Enzo, Farmingdale, USA) for 3 min at 37°C. Images were taken with Olympus FluoView 500 Laser Scanning Confocal Microscope (60× oil immersion objective; N.A.: 1.25). Percentage of nuclearly translocated NFAT was calculated from the intensities of nuclear NFAT fluorescence defined by Draq5 counterstaining and the total cellular NFAT fluorescence, respectively. Colocalization analysis was performed from ≥100 cells/sample by ImageJ software (http://rsbweb.nih.gov/ij) using the appropriate colocalization plug-in. The Pearson's colocalization index (CI) was used to get a reliable estimate on the extent of protein colocalization; CI values close to zero indicate no or a very low degree whereas CI ≥ 0.5 reflects a high degree of colocalization and CI = 1 value would correspond to a full overlap between the two colors in each pixel of the image (29).
Detection of Apoptosis
Cells activated with 15 μg/ml anti-CD3 at 37°C for 4 hr in Hank's A buffer (pH 7.4) were tested for mitochondrial membrane potential using 100 nM DiOC6(3) fluorescent potential sensitive dye (Molecular Probes-Invitrogen). DiOC6(3) fluorescence was measured by flow cytometry 30 min after its addition to the cells. Cells activated in the same way for 8 hr were tested for caspase activation using 50 μM D2-R110 (Molecular Probes-Invitrogen) fluorogenic caspase substrate (incubated with the cells for 30 min at 37°C). After washing, the cellular D2-R110 fluorescence was measured by flow cytometry. DNA fragmentation was analyzed with propidium iodide staining following a hypotonic extraction after 12 hr treatment with anti-CD3. For comparison, cells were also stimulated for cell death in other ways with 50 μM C2-ceramide for 6 hr (30) or with 10 μg/ml Concanavalin A for 24 hr. After wash, samples were incubated with 300 μl hypotonic fluorochrome solution (0.1% TritonX, 0.1% sodium citrate, 50 μg/ml propidium iodide) for 4 hr at 4°C and the subdiploid percentage was determined by flow cytometry.
Patch Clamp Measurements and Data Analysis
Whole-cell patch-clamp measurements and data analysis were performed as described in Ref.31. The standard bath solution (S-ECS) was (in mM): 145 NaCl, 5 KCl, 1 MgCl2, 2.5 CaCl2, 5.5 glucose, 10 HEPES (pH 7.35, 305 mOsm). The pipette solution was (in mM): 140 KF, 11 K2EGTA, 1 CaCl2, 2 MgCl2 and 10 HEPES (pH 7.20, ∼295 mOsm). TEA (tetraethyl-ammonium) containing bath solution composed of (in mM): 10 TEACl, 135 NaCl, 5 KCl, 1 MgCl2, 2.5 CaCl2, 5.5 glucose, 10 HEPES (pH 7.35, 305 mOsm). Margatoxin (MgTx) (Alomone Labs., Jerusalem, Israel) was dissolved in S-ECS supplemented with 0.1 mg/ml BSA (Sigma) to suppress nonspecific binding of the toxins to the wall of the tubes and to the Petri dish.
Comparison of multiple groups was accomplished using one way analysis of variance (ANOVA) or ANOVA on Ranks and all pairwise multiple comparison (Dunn's Method) was used to isolate the groups that differ from others at P < 0.05. R version of 2.13.0 Patched (2011-04-21 r55576) statistical software was used for the further significance calculations. The R Foundation for Statistical Computing is accessible free at ISBN 3-900051-07-0, Platform: i386-pc-mingw32/i386 (32-bit).
Generation and Functional Properties of Polarized Effector Th Hybridoma Cell Lines: Validation of a Polarized Th Cell Model
Polarized Th hybridomas were generated as described in the Materials and Methods. Three established cell clones (Th0:2.13; Th1:7.6; Th2:7.5) were selected for further examinations on the basis of their activation induced cytokine production and gene expression profile. Figures 1A–1C shows their IFNγ and IL-4 production as characteristic Th1 and Th2 type cytokine responses to activation stimuli, respectively. Th cells respond via TCR optimally if they also receive a costimulatory (second) signal, therefore the effector response of Th clones was tested to various stimuli, such as TCR crosslinking by anti-CD3 mAb; anti-CD3 + costimulating (metabolically inactivated) LK-35 B cells; or the polyclonal Concanavalin-A.
The 7.6 (Th1) cells activated by either way were unable to produce IL-4 while they produced IFNγ, depending on the inducing stimuli (Fig. 1B). The 7.5 (Th2) cells produced IL-4 but not IFNγ at all (Fig. 1C). The precursor 2.13 (Th0) cells produced IL-2 and GM-CSF comparable with the 7.6 or 7.5 clones (data not shown), but their selective cytokine (IL-4 or IFNγ) secretion was negligible and detectable only in case of anti-CD3 + LK-35 induced activation (Fig. 1A). The Th1 and Th2 cells showed a differential T-bet and GATA-3 expression (data not shown), as expected on the basis of the literature (6).
Recently increased Orai1, STIM and KCa expression was reported in polarized Th cells upon their activation, and Th1 cells were shown to have increased KCa activity, whereas Th2 cells a high calcium clearance rate (22, 32). Since expression and compartmentation of Kv1.3 ion channels may have a strong impact on the calcium signal of T cells (31, 33), we tested the expression and functional properties of these channels in our polarized Th cells. The type of Kv ion channels was identified using specific antagonists. TEA (10 mM) inhibited and slowed the C-inactivation of whole-cell potassium current in Th2 cells (Fig. 1D), while 50 pM MgTx effectively reduced the outward K+ current (peak current in the presence of MgTx is approximately 30% of the control: Fig. 1E). These are hallmarks of the interactions of TEA (34) or MgTx inhibitors with Kv1.3 channels (35), respectively. Similar results were obtained for Th1 cell lines (data not shown). These data indicate that both the Th1- and Th2- hybridomas express Kv1.3 potassium channels.
The biophysical properties of Kv1.3 channels can provide information on their expression level and functionality in T cells. Notably, activation time constant (τa), describing the kinetics of Kv1.3 channel opening, was found significantly higher in Th1 than Th2 hybridomas (τa,Th1 = 0.88 ± 0.1 ms (n = 8), τa,Th2 = 0.46 ± 0.01 ms (n = 8), ANOVA, P < 0.05). The inactivation time constant (τi) at +40 mV depolarizing potential, however, was the same in Th1 and Th2 cells (τI,Th1 = 192.6 ± 10.1 ms (n = 7), τi,Th2 = 163.5 ± 10.1 ms (n = 4), ANOVA P = 0.214).
The current density (CUD, ratio of peak current at +40 mV and the whole-cell capacitance of the cell) was also determined. This is an appropriate estimate of the number of functional Kv1.3 channels per unit membrane area, i.e. “ion channel density”. CUD for Th2 cells was higher than in Th1 cells (CUDTh2 = 808 ± 183 pA/pF and CUDTh1 = 377 ± 33 pA/pF; Fig. 1F; P < 0.05).
Based on these properties, these stable Th-effector cell lines can be considered as appropriate models for studying factors controlling polarized Th cell response.
The TCR/CD3-Mediated Ca2+-Response Is Different in Th1 and Th2 Effector Cells
Earlier electrophysiological studies reported some differences in the Ca2+-response of distinct Th subpopulations (22). Therefore, here we analyzed the anti-CD3 induced calcium signal quantitatively in our Th1 and Th2 cells by flow cytometry, using an analysis software developed recently (27), with special attention to the kinetics of the signal. In good accordance with earlier data, the Th1 cells responded with a more intense and sustained calcium signal than Th2 cells (Figs. 2A and 2B) to the same TCR stimulation. In addition, TCR-stimulation increased raft-association of Kv1.3 channels, positively controlling the Ca2+ signal, in both polarized Th cells (Fig. 2H).
The Nuclear Residence Times of NFAT1/2 Are Shorter in Th2 than in Th1 Cells and the NFAT Translocation Is Membrane Cholesterol-Dependent
During T cell activation, Ca2+-release into the cytoplasm activates through calmodulin the calcineurin phosphatase, which dephosphorylates the NFAT family transcription factors (NFATc1, NFATc2). This leads to their translocation into the nucleus and activation of cytokine gene transcription. The choice of IFNγ or IL-4 production might depend on both the duration and the amount of NFAT location in the nucleus which can be different in the Th subpopulations. Therefore next we investigated how NFAT1 and NFAT2 translocate into the nucleus after TCR stimulation with 2.5–15 μg/ml doses of anti-CD3 mAb. Their translocation showed saturation above 5 μg/ml anti-CD3 dose, as detected 30 min after stimulation. The polarized Th cells did not show significant difference in these saturation curves (Figs. 2C and 2D). The time dependence (from 15 to 60 min of TCR stimuli), however, clearly shows that both NFAT1 and NFAT2 have significantly shorter nuclear residence time in Th2 than in Th1 cells. In Th2 cells NFATs started to be rephosphorylated and shuttled back to the cytoplasm 30 min after anti-CD3 stimulation, while in Th1 cells the majority of NFAT1 still remained in the nucleus even after 60 min activation (Figs. 2E and 2F). This is nicely demonstrated by the representative confocal images of Th1 and Th2 cells recorded at different times of TCR stimulation (Fig. 2I).
Depletion of membrane cholesterol (a stabilizing factor of lipid rafts) by MBCD before TCR stimulation resulted in impaired NFAT nuclear translocation in both Th cell subsets (Fig. 2G). This suggests a direct coupling between the actual membrane composition/organization and the calcium signal-dependent NFAT activation.
Differential Membrane Compartmentation of the TCR/CD3 Complex in the Polarized Th Cells
Since lateral distribution and membrane compartmentation of the TCR/CD3 complex by lipid rafts may influence the initial signaling intensity in T cells, next we investigated these properties in polarized Th subpopulations. Using images of fluorescently labeled GM1/GM3 gangliosides and TCR/CD3 complex, the Pearson's coefficients were calculated from double-stained cells to quantify the extent of their colocalization. The TCRs became more raft-localized upon activation in all cell types (Figs. 3A–3G). Their raft-association, however, was significantly weaker in Th2 cells (Fig. 3G). Most importantly, while TCR stimulation in Th1 or Th0 cells leads to a typically polarized (cap-like) cell surface distribution of T cell receptors, in Th2 cells no such polarized distribution was observed (Figs. 3A–3F).
Cytokine Induced Th Cell Polarization Results in Altered Expression of Membrane Gangliosides, Cholesterol, and Raftophilic Accessory Proteins
Next we investigated whether differentiation of polarized T cells changes the expression patterns of key molecular components of T cell lipid rafts, namely the GM1 gangliosides, cholesterol, the CD4 coreceptor, as well as the CD48 and CD59 GPI-anchored accessory proteins. While there was no significant difference between GM1 ganglioside levels of Th1 and Th2 cells, the undifferentiated cells (Th0) expressed much less gangliosides in their membrane (Fig. 3H). The membrane cholesterol level was significantly lower in Th2 than in Th1 cells (Fig. 3I), which may result in a lower number of stable rafts in Th2 cells. The CD4 expression was slight, but the CD3 density significantly higher in Th2 than in Th1 cells. The raftophilic CD59 was expressed dominantly in Th1 cells, while CD48 in Th2 cells (Figs. 4A and 4C).
To avoid the phenotypic differences that are mostly due to the hybridomas, the same expression patterns were investigated in in vitro polarized primary splenic CD4+ T cells, as well. The freshly prepared, CD4+ spleen T cells were activated and cultured in polarizing cytokine milieu as detailed in Materials and Methods. After 7 days, the expression profiles were determined by flow cytometry. The expression differences between cells skewed into Th1 or Th2 direction (Figs. 4B and 4D) were highly similar to those obtained with polarized Th hybridomas. These data together suggest that the cytokine-driven differentiation of Th cells may result in selectively altered expression pattern of several raftophilic lipids (GM1 and cholesterol) and proteins (CD48, CD59) which are known to control T cell activation signaling.
Polarized Th Cells Show Differential Sensitivity to TCR Mediated Cell Death (AICD) but not to other TCR/Fas-Independent Stimuli
The sensitivity of 7.6 (Th1), 7.5 (Th2) polarized Th cells, and 2.13 (Th0) undifferentiated Th cells to activation induced cell death (AICD) was investigated finally using three different markers of apoptosis. Proapoptotic anti-CD3 stimuli (15 μg/ml dose) resulted in significantly higher number of apoptotic cells in Th1 than in Th0 or Th2 hybridomas as shown either by the decreased mitochondrial membrane potential detected after 4 hr (Figs. 5A and 5B); by the caspase activation detected after 8 hr (Figs. 5C and 5D) or by the DNA fragmentation detected after 12 hr (Figs. 5E and 5F). Notably, the undifferentiated Th0 cells were the least sensitive to AICD. In contrast, other death-inducing stimuli, like 6 hr treatment with C2-ceramide, activating the mitochondrial death-pathway, (Fig. 5G), or 24 hr ConA stimulation (Fig. 5H) resulted in the same extent of DNA fragmentation in Th1 and Th2 cells. Histograms on Figures 5A, 5C, and 5E are representative illustrations of the flow cytometric data evaluation strategies.
The cytokines polarizing naive Th cells to Th1 or Th2 direction, as well as the key transcriptional factors involved in these processes have been identified so far (6). Much less is known, however, about the key membrane and signaling factors controlling polarized Th cell functions and fate, although they may be potential targets for modulation of Th1/Th2 balance in pathological situations. Since an altered Th1/Th2 balance may often be coupled to serious immunological disorders, such as organ specific diseases, inflammatory allergic symptoms or promotion of autoimmune diseases (5, 6), there is an urgent need for finding key signal-therapy targets.
Using a Th hybridoma model system, we demonstrate here that membrane organization of the TCR/CD3 complex, lipid rafts, and the T cell calcium signaling machinery are key factors in the polarized effector Th cell function. Earlier studies reported on increased plasma membrane raft/ganglioside level and de novo sphingolipid biosynthesis during the transition from naive to effector T cells (36). Similar difference, i.e. increased GM1 level in the cytokine-differentiated Th1 and Th2 cells was found here compared to the undifferentiated Th0 cells. Th1 cells, in contrast to the nearly equal GM1 levels, had higher membrane cholesterol level than Th2 cells. This likely results in a lower number of rafts in the plasma membrane of Th2 cells. Different membrane lipid order was also reported in Th1 and Th2 cell populations (23), moreover pharmacological manipulation of the membrane order could alter T cell function. Condensation of the plasma membrane and accumulation of raft lipids at the site of antigen-specific activation through TCR/CD3 complexes was also reported in T cells (37). Moreover, manipulation of the T cell lipidome by polyunsaturated fatty acids impaired formation of TCR signaling foci. Accumulation into lipid raft enriched immunological synapses of proteins involved in calcium influx (e.g. Orai1 and STIM) or maintaining the driving force for Ca2+ entry (e.g. Kv1.3 voltage gated K+ channels) was also shown upon Th cell activation (31, 33, 38), consistent with the present data. The functional (gating) properties of the latter ion channels was found to be cholesterol-sensitive (16).
The data presented here are consistent with the more intense and sustained Ca2+-response of Th1 cells shown here and by other groups. Our results emphasize that a differential plasma membrane receptor- and ion channel-compartmentation by lipid raft microdomains is a notable factor that can control activation signaling and in turn the cytokine profile of Th cells. Our data provide several pieces of new and strong evidence for this, such as the enhanced lipid raft-localization of TCRs and their spatial polarization in Th1 but not in Th2 cells. The differently organized “signal transduction modules” (39) of Th1 and Th2 cells, respectively, may result in their altered synapse formation with APCs. Indeed the synapse of Th2 cells was found less organized with a failure to form the classic ‘bulls-eye’ patterning of central TCR (40), in accordance with our above finding.
We also show here that the cytokine-driven differentiation/polarization of Th cells results in altered expression levels of some raftophilic accessory membrane proteins involved in T cell activation. Skewing cells into Th1 direction by cytokines enhanced CD59 expression while Th2 cells display higher CD48 expression in their membrane raft domains than Th1 cells. These characteristic differences were detectable already after one week culturing in the appropriate polarizing cytokine milieu. The different expression levels of these raftophilic GPI-anchored membrane proteins, similar to other cells (41, 42), may also have significance in the effector functions of the Th cell subsets. CD48 (Blast-1) is a CD2-SLAM family member while CD59 (protectin) is a membrane inhibitor of complement-reactive lysis blocking MAC (membrane attack complex) assembly and has been shown to modulate immune response in a complement independent manner, as well (43). Both molecules are broadly expressed in man and mice too and are shown to be involved in T-cell activation as costimulatory ligands and also as costimulatory receptors. In one hand anti-CD59 increased calcium signal and IL-2 secretion (44), on the other hand virus specific T-cell responses were significantly enhanced in CD59 deficient mice (45). The blockade of CD59 enhanced tumor specific T-cell response (46), however the mechanism and the possible counterreceptor(s) are still unclear. Coligation of Fas (CD95) with CD59 inhibits apoptotic signal whereas CD28 recruitment amplifies it suggesting that CD28 and CD59 colocalize in different types of membrane microdomains (47). Stimulation of CD48 may induce rearrangement of signaling factors in lipid rafts (48). CD48 but not CD59 was shown being recruited to the immobilized TCR/CD3 complex (49). Benefits of manipulating through CD59 were described in virus or tumor-specific response (50) while CD48 was proposed as a potential target in the therapy of Th2 mediated diseases (51, 52). The observed cell-specific enrichment of these GPI-proteins in lipid rafts may influence not only the T cell activation process, but the extent of cell death, as well, therefore they are also notable players of the polarized cellular immune response.
In recent studies the scaffolding protein Dlgh1 (Disc large homolog 1) was shown to link the TCR mediated cell activation to the NFAT1 via p38, and its knock-down or reduced expression can abrogate T cell function/proliferation (53, 54). Kv1.3 ion channels significantly contribute to the sustained phase of the Ca2+-signal upon T cell activation. Although the Kv1.3 current density was found slightly greater in Th2 than in Th1 cells, the magnitude of the Ca2+-signal is larger in the latter cell type. As shown for SLE (Systemic Lupus Erythematosus) T-cells, the Ca2+-signal is more sustained and NFAT activity is higher compared to control T cells, even though the biophysical properties and expression level of Kv1.3 were the same (55, 56). Furthermore, the cease or altered kinetics of the redistribution of Kv1.3 channels during engagement of T cells in immunological synapse can also modify Ca2+-responses. In addition, a scaffolding protein of Kv1.3 (Dlgh1/CARMA-1), which can bind and couple various protein kinases to the channel, such as p56lck and PKA, was demonstrated to promote NFAT activation (54, 57, 58). Based on these, we suppose that the control of Kv1.3 current in Th1 cell lineage may be due to the modification of the channel by protein kinases, rather than targeting to distinct lipid compartments, although it cannot be excluded, either. In addition, it is worth to note that the channel opening proved to be faster in Th1 than in Th2 cells. Other major sources of the differential calcium responses of Th1 and Th2 cells can be the enhanced expression and activity of Ca2+-dependent K+-channels (KCa) and the more rapid Ca2+-extrusion (clearance) by Th2 cells, as suggested earlier (22). A differential regulation of Th1 and Th2 cell Ca2+-signaling by Trpm4 sodium channel mediating membrane depolarization (21) may also be accounted for the differential calcium response of polarized Th cells, such as the accumulation and upregulation of STIM and Orai1 calcium channels in the immunological synapse upon T cell activation (32).
The plasma membrane-related changes converted to altered Ca2+ signals may in turn affect the nuclear translocation and residence time of NFAT family proteins, key transcriptional regulators of cytokine gene activation in lymphocytes (12, 15). NFAT2/NFATc1 and NFAT1/NFATc2 were proposed to positively regulate activation of cytokine genes coupled to Th2 and Th1 responses, respectively (59). Our data clearly demonstrate that 30 min after activation the extents of NFAT1 and NFAT2 nuclear translocation do not differ in Th1 and Th2 cells, but the nuclear residence time is much shorter in Th2 cells for both of them. In addition, we demonstrate herein that nuclear residence of NFAT1/2 is significantly reduced by depleting membrane cholesterol—a stabilizer of raft microdomains—suggesting a coupling between the membrane microdomain organization and cytokine gene activation/production of polarized Th cells (Fig. 6).
Concerning the different cell death sensitivity of the polarized Th cells, we show that the extent of cell death nicely correlates with the strength of TCR stimulation and the downstream signaling in rank order. In contrast, the three polarized Th cells did not show difference in cell death when it was induced through a TCR/Fas-independent pathway. These data are consistent with a proposed model of differential control of Th1 vs. Th2 cell death governed by a CD95-caspase line (Th1 cells) or a Grb-dependent line (Th2 cells) (60).
In conclusion, our results suggest that the differences in membrane microdomain composition/organization, in TCR/CD3 and ion channel membrane compartmentation—in connection to the difference in the Ca2+ signal profiles and NFAT nuclear residence times—are characteristic properties of differentiated Th1 and Th2 cells, respectively. These properties may underlie their differential activation of cytokine genes, i.e. the polarized Th cell response. Another important aspect of this question, namely how the polarizing cytokine signals affect the genes and the protein expression levels of the molecular components of this signal axis, still remained unanswered and needs further investigations.
The authors greatfully thank the skillful technical assistance to Erzsebet T. Veress and Márta Pásztor, the help in preparation of hybridomas to Helga Latos and the valuable discussions to Dr. Andrea Balogh.