TWIK-related Acid-sensitive K+ Channel 1 (TASK1) and TASK3 Critically Influence T Lymphocyte Effector Functions*

Two major K+ channels are expressed in T cells, (i) the voltage-dependent KV1.3 channel and (ii) the Ca2+-activated K+ channel KCa 3.1 (IKCa channel). Both critically influence T cell effector functions in vitro and animal models in vivo. Here we identify and characterize TWIK-related acid-sensitive potassium channel 1 (TASK1) and TASK3 as an important third K+ conductance on T lymphocytes. T lymphocytes constitutively express TASK1 and -3 protein. Application of semi-selective TASK blockers resulted in a significant reduction of cytokine production and cell proliferation. Interference with TASK channels on CD3+ T cells revealed a dose-dependent reduction (∼40%) of an outward current in patch clamp recordings indicative of TASK channels, a finding confirmed by computational modeling. In vivo relevance of our findings was addressed in an experimental model of multiple sclerosis, adoptive transfer experimental autoimmune encephalomyelitis. Pretreatment of myelin basic protein-specific encephalitogenic T lymphocytes with TASK modulators was associated with significant amelioration of the disease course in Lewis rats. These data introduce K2P channels as novel potassium conductance on T lymphocytes critically influencing T cell effector function and identify a possible molecular target for immunomodulation in T cell-mediated autoimmune disorders.


Two major K ؉ channels are expressed in T cells, (i) the voltagedependent K V 1.3 channel and (ii) the Ca 2؉ -activated K ؉ channel KCa 3.1 (IKCa channel). Both critically influence T cell effector functions in vitro and animal models in vivo.
Here we identify and characterize TWIK-related acid-sensitive potassium channel 1 (TASK1) and TASK3 as an important third K ؉ conductance on T lymphocytes. T lymphocytes constitutively express TASK1 and -3 protein. Application of semi-selective TASK blockers resulted in a significant reduction of cytokine production and cell proliferation. Interference with TASK channels on CD3 ؉ T cells revealed a dose-dependent reduction (ϳ40%) of an outward current in patch clamp recordings indicative of TASK channels, a finding confirmed by computational modeling. In vivo relevance of our findings was addressed in an experimental model of multiple sclerosis, adoptive transfer experimental autoimmune encephalomyelitis. Pre-treatment of myelin basic protein-specific encephalitogenic T lymphocytes with TASK modulators was associated with significant amelioration of the disease course in Lewis rats. These data introduce K 2 P channels as novel potassium conductance on T lymphocytes critically influencing T cell effector function and identify a possible molecular target for immunomodulation in T cell-mediated autoimmune disorders.
The last decade has revealed much knowledge about the intracellular events accompanied by T lymphocyte activation following recognition of antigens bound to major histocompatibility complexes. K ϩ selective ion channels in T cells and their role in immune responses have been discussed for decades, since the discovery that non-selective K ϩ blockers could inhibit T cell proliferation in vitro (1)(2)(3). The role of K ϩ channels in the activation of T cells is pivotal, because opening the channels hyperpolarizes the membrane potential, which in turn increases the influx of Ca 2ϩ via Ca 2ϩ release-activated Ca 2ϩ channels (CRAC) 4 (4). Signaling cascades after T cell receptor stimulation involve phospholipase C-␥ (5)-mediated cleavage of phosphatidylinositol 4,5-bisphosphate into diacylglycerol (DAG) and inositol 1,4,5-trisphosphate, which results in a transient elevation of intracellular calcium concentration ([Ca 2ϩ ] i ) triggered by inositol 1,4,5-trisphosphate binding to inositol 1,4,5-trisphosphate receptors (5).
Intracellular Ca 2ϩ release triggers activation of CRAC channels resulting in longer lasting (ϳ1 h) elevated [Ca 2ϩ ] i mandatory for further transcription-dependent steps of T cell activation (4,6). With the combined use of patch clamp electrophysiology and molecular biology, the voltage-dependent K ϩ channel 1.3 (K V 1.3 channel) and the intermediate conductance Ca 2ϩ -activated channel (IKCa channel; IK Ca 1 or K Ca 3.1; which acquires its Ca 2ϩ dependence from constitutively bound calmodulin) were found to be the dominating K ϩ channel types expressed in T cells (7)(8)(9)(10). K V 1.3 is activated by membrane depolarization, whereas K Ca 3.1 channel opening is triggered by a rise in Ca 2ϩ ions (5). Selective blockade of K V 1.3 leads to suppression of T cell effector function, e.g. decreased cytokine release and suppression of proliferation (11,12). Underlining this finding, in vivo blockade of K V 1.3 has been shown to mediate beneficial effects in experimental autoimmune encephalomyelitis (EAE), a rodent model for multiple sclerosis (12,13). The importance of K ϩ channels has been further validated by research on human T lymphocytes, which revealed a differential expression patterns of K V 1.3 and K Ca 3.1 on specific T cell subtypes (14). Interestingly, "chronically activated" T cells induced by repeated antigenic challenges of the cells selectively and stably up-regulate K V 1.3 channels (15). Furthermore, these cells could be defined as CCR 7 Ϫ CD45RA Ϫ effector memory T lymphocytes (16), which are implicated in different autoimmune disorders (15). In agreement with that K V 1.3 has been shown to be highly expressed in human myelin-reactive T cells of patients with multiple sclerosis (17) as well as in autoreactive T cells from patients with other autoimmune diseases such as type-1 diabetes or rheumatoid arthritis (18). Thus K ϩ channels, especially K V 1.3, have emerged as attractive targets for highly selective immunosuppressive strategies applicable for example in the therapy of T cellmediated autoimmune disorders (19).
Two-pore-domain K ϩ channels (K 2 P, KCNK), often described as "background" or "leak" channels, play a pivotal role in setting the resting membrane potential and modulating neuronal excitability (20 -23). K 2 P channels consist of four transmembrane domains arranged in tandem building two pores. Importantly, their action is mostly time and voltage independent (21). TASK1 and TASK3 channels, two functional members of the K 2 P channel family, can be regulated by a diversity of stimuli (extracellular acidification, G q proteins, and muscarine (20,(22)(23)(24)), and exhibit insensitivity to "classical" potassium channel blockers (e.g. tetraethylammonium, 4-aminopyridine). Based on their specific electrophysiological properties, pharmacological profile, and functional impact on the resting membrane potential they might represent a relevant ionic conductance influencing basic T cell function. We therefore questioned whether K 2 P channels are expressed on T lymphocytes and dissect their functional role on the level of electrophysiological properties and T cell effector function, both in vitro and in an animal model of multiple sclerosis in vivo.
Isolation and Culture of Human T Cells-Human T cells were purified from peripheral blood samples of healthy donors (n ϭ 49). Peripheral blood mononuclear cells were prepared by density centrifugation (Lymphoprep, Axis-Shield, Oslo, Norway) according to the manufacturer's instructions. This was followed by magnetic cell sorting (MACS, CD3 Microbeads, Miltenyi Biotec, Karlsruhe, Germany) for CD3. Purity of CD3 ϩ T lymphocytes was Ͼ98% as assessed by flow cytometry. Cells were maintained in RPMI 1640 containing 10% human AB-serum, 25 mM HEPES, 1% glutamine, and 1% antibiotics.
Assessment of T Cell Function, Pharmacological Blockades-4 ϫ 10 6 freshly isolated CD3 ϩ T lymphocytes were seeded in 1 ml of T cell medium per well. CD3/CD28 dynabeads (Dynal Biotech) were added at a T cell to bead ratio of 1:1. Channel blockers in different concentrations were applied in parallel to bead stimulation (K V 1.3 blockers, 10 and 100 nM charybdotoxin, and 1, 10, and 100 nM Psora-4; TASK1/3 blocker, 30, 100, and 250 M bupivacaine; TASK1 inhibitor, 3, 30, and 100 M anandamide; TASK3 blockers, 50 M, 500 M, and 1 mM spermine, and 100 nM, 1 M, 10 M ruthenium red). The solvent solution in the final experimental solution did not exceed 1%. Application of the solvent alone (1%) did not influence the analyzed parameters. Stimulations were all done in duplicates. After 24 h incubation at 37°C and 5% CO 2 , cells were centrifuged and subjected to further analysis by flow cytometry (stainings for annexin V and propidium iodide). In parallel, supernatants were assessed for IFN␥ or IL2 protein levels by enzyme-linked immunosorbent assay (R&D Systems, Wiesbaden, Germany) according to the manufacturer's instructions.
Flow Cytometry for TASK Channel Expression on T Lymphocytes-Flow cytometry acquisition was done by standard methods. For antibody staining cells were resuspended in FACS buffer (PBS containing 0.1% bovine serum albumin and 0.1% NaN 3 ), directly labeled antibodies were added. For annexin V/propidium iodide assays, cells were resuspended in annexin binding buffer (10 mM HEPES, 150 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , and 0.18 mM CaCl 2 ) and stained with annexin-FLUOS (Roche) and propidium iodide (Calbiochem). For evaluation of TASK1/3 antibody specificity the following antibodies were used: rabbit anti-TASK1 (number P0981, Sigma), goat anti-TASK1 (number SC-32067), rabbit anti-TASK3 (number AB5721, Chemicon, Ochsenhausen, Germany), and goat anti-TASK3 (number SC-11320, Santa Cruz, Santa Cruz, CA) followed by the appropriate secondary antibodies (goat anti-rabbit fluorescein isothiocyanate or donkey anti-goat Cy3; Dianova, Hamburg, Germany) using standard protocols for extracellular (see above) and intracellular staining (permeabilization buffer and fixation buffer from ebioscience, San Diego, CA). Data acquisition was done using a FACS Calibur system (BD Biosciences). Results were analyzed using CellQuest Pro Software (BD Biosciences).
T Cell Proliferation Assay in the Presence and Absence of TASK Channel Modulators-Human T cells were labeled with 4 M CFSE (Molecular Probes) for 10 min in the dark. RPMI containing 15% fetal calf serum was added for 20 min to stop labeling and was followed by three washing steps with RPMI, 10% fetal calf serum. CFSE-labeled cells were suspended in T cell medium and 1 ϫ 10 5 cells per well were cultured at 37°C and 5% CO 2 in the presence of CD3/CD28 beads and ion channel modulators as described above. After 3 days of proliferation CFSE-labeled cells were washed and resuspended in FACS buffer. Proliferation assays were performed in duplicate and analyzed by flow cytometry calculating the responder frequency (dividing the number of dividing cells by total number of cells).
Immunocytochemistry of Human and Rat T Cells-Immunocytochemical stainings were performed on human T cells and rat myelin basic protein (MBP)-specific T cells. Cells were placed on coverslips coated with poly-L-lysine (Sigma) and fixed with 4% paraformaldehyde. Subsequently, cells were blocked with PBS containing 10% horse serum (PAA Laboratories, Cölbe, Germany), 2% bovine serum albumin, and 0.3% Triton X-100 overnight. Next, the primary antibodies (rabbit anti-TASK1, Sigma; rabbit anti-TASK3, Chemicon; rabbit anti-K V 1.3, Chemicon) were added and incubated for 1 h. Cells were washed with PBS containing 0.3% Triton X-100 and incubated with secondary antibodies (Cy3-conjugated rabbit anti-goat, 1:100, Dianova) for another hour. Counterstaining of cell nuclei was performed using DAPI (0.5 g/ml, Merck). Pictures were collected by immunofluorescence microscopy (Axiophot, Zeiss, Jena, Germany). Negative controls without the primary antibody revealed no positive signals (data not shown).
Western Blot of T Cell Lysates-Whole cell lysates from isolated purified T cells were used for Western blot analysis. In brief, cells were washed with ice-cold PBS, resuspended in lysis buffer (PBS containing 1% Triton X-100 and protease inhibitor mixture (Roche Diagnostics, Mannheim, Germany), and solubilized by sonification on ice. Cell lysates were centrifuged and protein content in the clarified supernatant was measured by Bradford reaction. Samples (50 g/lane for TASK1 and 100 g/lane for TASK3) were subjected to 10% SDS-PAGE, followed by transfer to nitrocellulose membranes. Whole mouse brain (C57/Bl6) was used as a positive control for TASK channel expression (23). Protein transfer was visualized by Ponceau S staining and membranes were then blocked with PBS containing 0.05% Tween 20 and 5% dry milk. The membranes were probed with rabbit anti-TASK1 (raised against the C-terminal part of the channel, 1:200; Sigma) or rabbit anti-TASK3 (polyclonal antibody against the C-terminal part of TASK3, 1:200; Chemicon), respectively. The secondary antibody was horseradish peroxidase-conjugated donkey anti-rabbit (1:3,000; Amersham Biosciences). The antibody reaction was detected by enhanced chemiluminescence reaction (ECL, Amersham Biosciences).
Adoptive Transfer Experimental Autoimmune Encephalomyelitis-Female Lewis rats (150 -160 g) were kept at standard conditions with free access to food and water. All animal experiments were approved by local authorities and conducted according to the German law of animal protection. Adoptive transfer experimental allergic encephalomyelitis was induced using a specific rat MBP T cell line (MBP-TC), which was generated as described previously (27). The MBP-TCs used for this experiment were freshly restimulated for 3 days with antigen presenting cells and guinea pig MBP protein (in one Petri dish: 3 ϫ 10 6 MBP-TCs ϩ 150 ϫ 10 6 antigen presenting cells ϩ guinea pig MBP 10 g/ml in 10 ml of restimulation medium, which consists of RPMI 1640 containing 1% glutamine, 1% antibiotics, 1% rat serum, and 0.4% ␤-mercaptoethanol). After 24 and 48 h cells were split and fed RPMI 1640 containing 1% glutamine, 1% antibiotics, 5% fetal calf serum, and 7.5% supernatants of concanavalin A-activated spleen cells. 24 h before cell transfer, channel blockers (10 nM Psora-4, 30 M anandamide, and 100 nM charybdotoxin) were added directly into the cell cultures. After separation from the antigen presenting cells by Ficoll gradients, 6 ϫ 10 6 MBP-TCs were transferred into four groups of Lewis rats intravenously (each n ϭ 8 -9) or plated for another 24 h for evaluation of IFN␥ production. The animals were weighed and scored every day by a blinded investigator. Disease severity was assessed following a well established 6-grade scoring system: 0 ϭ healthy; 1 ϭ weight loss, limp tip of tail; 2 ϭ limp tail, mild paresis; 3 ϭ moderate paraparesis, ataxia; 4 ϭ tetraparesis; 5 ϭ moribund; 6 ϭ death (28).
Numeric Model of T Cell Outward Currents-A single compartment T cell model including currents I Kv1.3 , I TASK , and I C was developed within the NEURON Simulation Environment (29). The length and diameter of the compartment were set to 10 m each, resulting in a total membrane area of 314 m 2 . By assuming the following ion distribution (mM): K ϩ in ϭ 135, K ϩ out ϭ 3.5, Ca 2ϩ in ϭ 6.5 ϫ 10 Ϫ4 (30), Ca 2ϩ out ϭ 0.13, we derived a potassium reversal potential (Ek) of about Ϫ94 mV and a calcium reversal potential (E Ca ) of about 98 mV following the Nernst equation. A specific membrane capacitance of 1 F/cm 2 was applied and all simulations were executed at 25°C.
The delayed rectifier current I Kv1.3 was realized by adapting the membrane mechanism I KD from a Purkinje cell model (31). Shifting the half-maximum activation and inactivation to Ϫ28.8 and Ϫ57.8 mV, respectively, led to a current that matches the slowly inactivating current component seen in human T cells (32,33). The maximum specific conductance gkbar was set to 0.0023 S/cm 2 to reach a maximum current of about 200 pA at a membrane potential of ϩ40 mV. The Ik current amplitude was calculated according to the following equation: Ik ϭ gkbar ϫ m ϫ h ϫ (V M Ϫ Ek); m, activation state variable; h, inactivation state variable; and V M , membrane voltage. Including the potassium leak current I TASK into the model was accomplished by using the corresponding membrane mechanism as described earlier (34). The described kinetic of that current thereby remained unaffected, whereas the maximum specific conductance was adapted to reach a maximum current of about 200 pA at a membrane potential of ϩ40 mV. The calcium-dependent potassium current I C originates from Arthur Houweling's NEURON demo. Kinetics of the current were described previously (35) and remained unchanged during all simulations. Setting the maximum specific conductance gkbar to 0.00034 S/cm 2 resulted in an approximate 115 pA at a membrane potential of ϩ40 mV. The following equation describes this non-inactivating current: Ik ϭ gkbar ϫ m ϫ (V M Ϫ Ek), where Ik, current amplitude; m, activation state variable; and V M , membrane voltage.
Statistical Analysis-All results are presented as mean Ϯ S.E. Statistical analysis was performed using the Student's t test modified for small samples as described previously (36). Statistical significance was set at p Ͻ 0.05, which is indicated by double asterisks.

TASK1 and TASK3 Channels Are Expressed on Human T
Lymphocytes-In a first set of experiments purified human CD3 ϩ T lymphocytes (n ϭ 6 donors) were assessed by immu-TASK on T cells MAY 23, 2008 • VOLUME 283 • NUMBER 21 nocytochemistry for the expression of K 2 P channels TASK1 and TASK3. Voltage-gated K V 1.3 channel was stained as a positive control (18). Immunoreactivity for TASK1, TASK3, and K V 1.3 was clearly detectable (Fig. 1, A-C). Counterstaining of cell nuclei with DAPI indicated a membrane-bound distribution pattern for all channels, whereas incubation with the secondary antibody alone showed no signals (data not shown). Protein expression of TASK1 and TASK3 by human T cells was corroborated by Western blot analysis (n ϭ 6 donors). The band recognized by the anti-TASK1 antibody was ϳ50 kDa (Fig. 1D, lanes 4-6) as published earlier (37). In human lympho-cytes as well as in our positive control (whole mouse brain, lane 3) a second band was visible at 65 kDa (Fig. 1D, second arrow), not visible in specificity controls (Fig. 1D, lanes 1 and 2). Using an anti-TASK3 antibody a distinct band at 60 kDa was visible, both in the mouse brain and human T lymphocytes (Fig. 1E, lanes [2][3][4][5]. To verify antibody specificity, different antibodies for TASK1 and TASK3 were also assessed by flow cytometry. Antibodies from Sigma (TASK1) and Santa Cruz (TASK3) recognizing intracellular epitopes of these channels showed specific binding only in intracellular FACS staining with permeabilized cells (Fig. 1F, right side) and not in extracellular staining proto- cols (Fig. 1F, left side). The same specificity was found for an anti-TASK1 antibody (intracellular target) by Santa Cruz, whereas anti-TASK3 (extracellular; Chemicon) revealed posi-tive signals in intracellular and extracellular staining protocols (data not shown). Taken together, these results point to expression of K 2 P channels TASK1 and -3 on human CD3 ϩ T lymphocytes.

Electrophysiological Recordings Reveal the Contribution of TASK to the Potassium Outward Current in Human T Cells-
In the next set of experiments whole cell patch clamp recordings of purified CD3 ϩ human T lymphocytes were used to evaluate the contribution of TASK channels to the potassium outward current of T cells (n ϭ 31 donors; multiple single cell recordings of each donor). Stepping the membrane potential from Ϫ80 to ϩ40 mV (Fig. 4A, left panel, inset) under voltage-clamp conditions evoked an outward current of 424 pA (range: 213-944 pA). Application of the voltage protocol at 30-s intervals indicated a stable outward current over time and current run-down could by analyzed as 6 Ϯ 2% (Fig. 4, A, Fig. 4C, right panel). These results for the first time demonstrate a contribution of currents through K 2 P channels to outward currents in human T cells. Next, blocker-sensitive currents were analyzed after graphical subtraction of currents in the presence and absence of channel modulators. Expectedly, the Psora-4-sensitive current component (the current after application of Psora-4 was subtracted from the control current, Fig. 4A, right panel, black trace minus gray trace) revealed a rapid current onset upon stimulation and a slow inactivation kinetic indicative of the delayed rectifying current K V 1.3 (Fig. 4D, left panel). However, the anandamide-sensitive current component (the current after application of anandamide was subtracted from the control current, Fig. 4B, left panel, FIGURE 3. Modulation of TASK inhibits proliferation of human T lymphocytes. Human CD3 T lymphocytes were labeled with CFSE and cultured in the presence of CD3/CD28 beads (A) and various K ϩ channel modulators (B-D). Application of the potassium channel modulators Psora-4 (P, 10 nM, n ϭ 5), anandamide (A, 30 M, n ϭ 5), and ruthenium red (R, 100 nM, n ϭ 5) indicated a marked reduction of T cell proliferation. Exemplary histograms are shown. See corresponding Table 2.  black trace minus gray trace) displayed a fast current onset with nearly no inactivation over the stimulation protocol, a typical feature of voltage-independent TASK channels (Fig. 4D, right   panel). These findings were further corroborated by the electrophysiological fingerprint of charybdotoxin (50 nM), a second blocker of K V 1.3 (and KCa3.1), on human T lymphocytes (64.5 Ϯ 6.4% reduction, n ϭ 4, p ϭ 0.0002; Fig. 5A, inset  and right panel). The charybdotoxin-resistant currents (Fig. 5A, left panel) strongly resemble the properties of pure TASK currents, which provides further support for the coexistence of voltage-dependent and leak potassium currents on T cells. TASK3 modulation by ruthenium red (1 and 10 M) induced a significant current reduction of 40.9 Ϯ 8.5 and 52.6 Ϯ 4%, respectively (n ϭ 4, p ϭ 0.04/ 0.001; Fig. 5B). Taken together this data therefore clearly indicate the functional relevance of TASK channels on outward currents in human T lymphocytes and demonstrate the functional coexpression of K V 1.3 and TASK channels on these cells.
A Computational Model to Dissect the Contribution of Three K Channels (K V 1.3, KCa, and TASK) for K ϩ Outward Currents-In the next experimental step we used a numeric model of human T cells to analyze the current components contributing to the resting membrane potential in human T cells. The model included the voltage-dependent K V 1.3 (I Kv1.3 ) channel (Fig. 6A), whereas inactivation and activation parameters were calculated according to the literature (31,39). Based on these settings the currents evoked by a depolarizing voltage protocol (500 ms, from Ϫ50 to ϩ50 mV, decrement 10 mV; Fig.  6A, inset) showed slow inactivation kinetics and characteristics as described for the net outward current of human T cells (16,33). Next, the model was extended by the inclusion of a TASK current (I TASK ) as described earlier (34). TASK current alone resulted in a nearly voltage-independent model response to depolarizing voltage steps (Fig. 6B). Finally, we supplemented the model by including a Ca 2ϩ -dependent K ϩ current (I C , Fig.  6C) (35). Based on our pharmacological results we assumed the contribution of all three conductances to the net outward current in human T cells and calculated 40% I Kv1.3 , 40% I TASK , and 20% I C . The composed current evoked by a depolarizing voltage step to ϩ40 mV displayed typical features (e.g. slow inactivation) as recorded for human T lymphocytes (Fig. 6D, left panel). Mimicking Psora-4 actions (inhibitor of K V 1.3 channels as demonstrated by electrophysiological recordings) in the model reduced the K V 1.3 component to 0% resulting in a nearly voltage independent current response evoked by a depolarizing voltage step (Figs. 4A, right panel, gray trace, and 6D, middle  panel). This finding further corroborates our electrophysiological data above, suggesting a marked contribution of TASK channels to the outward current of human T cells. In a last step we modeled anandamide actions (TASK channel inhibitor as demonstrated by electrophysiological recordings) by reducing I TASK to 0%. The remaining current component represents Kv1.3 channels (Figs. 4B, left panel, gray trace, and 6D, right panel). Recapitulating these findings from a numerical cell model support the results from electrophysiological recordings showing the functional coexpression of TASK and K V 1.3 channels in human T lymphocytes.
Selective Blockade of T Lymphocyte TASK Channels Ameliorates Experimental Autoimmune Encephalomyelitis, a Model of Multiple Sclerosis-Guided by our data on the functional impact of TASK on T cell activation in vitro and our patch clamp results, we challenged the question whether selective modulation of T lymphocyte TASK channels modulates T cell-mediated inflammatory disorders in vivo. We therefore chose adoptive transfer experimental autoimmune encephalomyelitis, a well established animal model of multiple sclerosis in Lewis rats. To evaluate the effect of TASK modulation on T cells, we incubated encephalitogenic MBP-specific T lymphocytes with the K ϩ channel modulators (10 nM Psora-4, 100 nM charybdotoxin, and 30 M anandamide) prior to adoptive transfer in rats. Expression of K V 1.3 as well as TASK1 could be demonstrated on MBP-specific rat T lymphocytes by immunocytochemistry, similar to our results observed in human T lymphocytes (Fig. 7A). Intravenous transfer of restimulated MBP-specific (6 ϫ 10 6 cells/animal) T cells elicited a typical EAE with symptom onset at 2-3 days, disease maximum around day 5, and recovery from days 6 -10 (onset, 3.3 Ϯ 0.3 days, n ϭ 8; disease maximum, score 4.7 Ϯ 0.5, n ϭ 8; decline to baseline, 10.4 Ϯ 0.5 days, n ϭ 8; Fig. 7, C-E). Incubation of encephalitogenic T cells with the Kv1.3 blocker Psora-4 24 h prior to transfer led to a slightly delayed onset of disease (not significant, ns, p ϭ 0.09), amelioration of the disease maximum (score 3.0 Ϯ 0.6, n ϭ 8, p ϭ 0.03), and faster recovery of symptoms (p ϭ 0.03; Fig. 7C). Different results were seen when using charybdotoxin that showed no significant effect on the investigated parameters (onset, p ϭ 0.25; disease maximum, score 4.8 Ϯ 0.8, p ϭ 0.74, n ϭ 8; decline to baseline, p ϭ 0.57; Fig. 7D). Selective blockade of TASK channels with anandamide ameliorated adoptive transfer experimental allergic encephalomyelitis. Although disease onset was delayed in the presence of anandamide (ns, p ϭ 0.39), the maximum disease score was reduced significantly (2.8 Ϯ 0.3, n ϭ 9, p ϭ 0.02). Furthermore, some animals recovered earlier (p ϭ 0.03; Fig. 7E). In agreement with these results from EAE, MBP-specific T cells showed a significantly reduced IFN␥ production in vitro when pretreated with Psora-4 (0.71 Ϯ 0.08, p ϭ 0.02, n ϭ 5) and anandamide (0.66 Ϯ 0.09, p ϭ 0.01, n ϭ 5), whereas charybdotoxin failed to have a lasting effect on T cells (0.84 Ϯ 0.15, p ϭ 0.32, n ϭ 5). This could be due to different binding affinities of the used substances (Fig. 7B). Of note, the application of channel modulators in vivo was not associated with obvious side effects. Taken together, results obtained from the animal model demonstrate the pathophysiological relevance of TASK channel modulation in a T cellmediated disorder.

DISCUSSION
Potassium-selective ion channels play a key role in modulating effector functions of T lymphocytes. Based on convincing data in vitro and in animal models of autoimmune disorders in vivo, their selective pharmacological intervention has been proposed as a potential therapeutic strategy for T cell-mediated autoimmune disorders (12,16). Thus far, research reports have focused on 2 major K ϩ channels expressed in T cells, (i) the voltage-dependent K V 1.3 channel and (ii) the Ca 2ϩ -activated K channel KCa3.1. Pharmacological interventions of both types inhibit T effector functions in vitro and ameliorate disease in animal models of T cell-mediated autoimmunity in vivo (17).
TASK Channel Expression on Human T Lymphocytes: Implications for T Cell Effector Functions-Our report identifies and characterizes the K 2 P channels TASK1 and -3 as a third K ϩ conductance relevant for T lymphocyte activation and function. TASK channel expression was demonstrated by immunocytochemical stainings as well as Western blotting techniques. Functionally TASK channels contribute to the membrane resting potential of T cells and critically affect T cell receptor-mediated effector functions in vitro. This was evidenced by experiments measuring cytokine secretion and cell proliferation in the presence of two pharmacological TASK inhibitors and whole cell patch clamp recordings of human CD3 ϩ T cells. Computational modeling corroborated and extended electrophysiological dissection of the role of TASK channels among three relevant K ϩ conductances on T cells (K V 1.3, KCa3.1, and TASK) and support a hypothetical model, how TASK channels contribute to Ca 2ϩ entry and stabilization of the membrane potential. Finally, pathophysiological relevance of our findings was demonstrated by testing T cell selective pharmacological blockade of TASK channels in experimental autoimmune encephalomyelitis, an animal model of multiple sclerosis: TASK inhibition was associated with slightly delayed disease onset, significant amelioration of disease severity, and significant earlier recovery after adoptive transfer of pretreated encephalitogenic cells.
Role of Potassium Currents in Ca 2ϩ -dependent T Cell Activation-Interaction of the T cell receptor-CD3 complex with the antigen-loaded major histocompatibility complex molecules initiates intracellular signaling via the phospholipase C␥ pathway, resulting in an initial release of Ca 2ϩ from intracellular stores, increased intracellular Ca 2ϩ , activation of protein kinase C, and augmented expression of K V 1.3 and IK channels in the cell membrane. Ca 2ϩ re-enters the cell through CRAC channels, causing membrane depolarization and further raising intracellular Ca 2ϩ levels. The resulting K ϩ efflux hyperpolarizes the membrane and maximizes the driving force for continued Ca 2ϩ influx via CRAC channels. T cell response to antigenic stimulation is highly modulated by the shape and nature of this calcium signal indicating the importance of differences in this signaling pattern (16).
The ionic conductances thus far described to be responsible for calcium signaling in T lymphocytes are the voltage-gated potassium channel (K V 1.3), the calcium-activated potassium channel (IKCa1), and CRAC (3,7,11,14,40,41). Interestingly, a broad spectrum of antigenic and mitogenic stimuli of T cells up-regulate expression of IK channels and also K V 1.3 channels in vitro (8,42). The increased expression of the K ϩ channels significantly augments the hyperpolarizing capacity of the activated T cells (compared with naïve cells), accelerates the Ca 2ϩ influx, and secures a sustained high level of intracellular Ca 2ϩ , which is necessary for full-blown T cell proliferation and cytokine production. Accordingly, in the presence of K ϩ blockers, the capability of the T cells to maintain a negative membrane potential and a long-lasting Ca 2ϩ signal is reduced, which is the rationale for their putative efficacy in attenuating T cell-mediated immune responses.
Role of K 2 P Channels in T Cell Physiology: Electrophysiology and Computational Modeling-What might be the role of our newly described K 2 P channels in T cells in this complex interplay? K 2 P represent a unique family of potassium channels and are mainly responsible for setting the membrane potential in a number of different cell types (20,34,(43)(44)(45)(46). Given the insensitivity of TASK channels against classical potassium channel blockers, these channels show a response against a unique panel of inhibitors including bupivacaine (23,(47)(48)(49)(50)(51), anandamide (35,52), spermine (53), and ruthenium red (53,54). Pharmacological interference with TASK channels attenuates T cell proliferation and cytokine secretion, similar to the interference with K V 1.3 using well established channel blocking compounds such as Psora-4 (38) and charybdotoxin (55). Notably, the inhibitory effect on T cell activity was comparable between the K V 1.3 silencing substances Psora-4 and charybdotoxin (note here that charybdotoxin also acts on IKCa1 channels) and the TASK channel inhibitors strongly suggesting an important functional impact of the leak channels on T-cell function.
Contribution of TASK currents to the net outward current of human T lymphocytes was assessed using whole cell patch clamp recordings. A well established protocol stepping the membrane potential from Ϫ80 to ϩ40 mV over 500 ms every 30 s elicited a marked outward current with a rapid onset and a slow inactivation (single exponential decay, ϳ 195 ms) as described earlier (16). Expectedly, application of K V 1.3 channel modulators resulted in a marked current reduction, whereas current kinetics remained unchanged (16). The endocannabinoid anandamide, however, a well known inhibitor of TASK channels (35,52), induced a comparable inhibition of the outward current, but in contrast to K V 1.3 blockers anandamide significantly altered current kinetics. Blocker-sensitive currents obtained by graphical subtraction of currents after application of the inhibitor from currents under control conditions showed nearly no voltage dependence in the anandamide-sensitive component, therefore suggesting the contribution of the TASK leak current (23,34,35). These findings were corroborated by results from a single-compartment computational model including currents I Kv1.3 , I TASK , and I C . Furthermore, functional expression of K 2 P channels could be shown on isolated human CD3 ϩ T lymphocytes and all of the recorded CD3 ϩ T cells showed currents indicative of TASK channels. It is currently unknown how (or if) expression of these channels is regulated (e.g. under inflammatory conditions). Furthermore, the investigation of expression levels in different T cell subtypes (naïve, Th 17 encephalitogenic T cells, regulatory T cells, CD4 ϩ , CD8 ϩ , T central memory cells (T cm ), and T effector memory cells (T em )) is warranted to determine whether TASK channels correlate to specific T cell functions (similar to K V 1.3 channels). However, based on the abundant expression of the K 2 P channel in a number of different cell types (20,34,(43)(44)(45)(46) it might be speculated that these channels are expressed throughout different T cell subtypes and that influence of TASK on K ϩ conductances represents a "baseline" feature common to all T cells. Our findings imply the possibility that earlier results concerning the net outward current in human T lymphocytes could be contaminated by a TASK channel component, not known or extrapolated at that time. However, the exact role of each component contributing to the outward current in T cell can only be estimated: based on our pharmacological and modeling results we suggest a nearly equal contribution of currents through TASK channels (ϳ40%) and K V 1.3 channels (ϳ40%) to the outward current in human CD3 ϩ T cells, whereas other potassium conductances (e.g. IKCa1) contribute to the remaining ϳ20%. Taken together, our data support the following hypothetical model for K 2 P channels in the cascade of TCR-mediated T cell activation, as summarized in the schematic in Fig. 8.
Selective Blockade of T Lymphocyte TASK Channels Ameliorates Experimental Autoimmune Encephalomyelitis, a Model of Multiple Sclerosis-To prove the pathophysiological relevance of T lymphocyte TASK channels we investigated their modulation in experimental autoimmune encephalomyelitis (27). Selective blockade of T lymphocyte K ϩ channels was achieved by preincubation of encephalitogenic MBP lines prior to transfer with various blockers. This approach was used to dissect the relevance on T cells, which would have been contaminated when applying the channel blockers systemically. In accordance with the literature, K V 1.3 blockade by Psora-4 ameliorated EAE: animals showed a delayed disease onset (nonsignificant), less severe disease course (significant), and a tendency toward This up-regulation of Ca 2ϩ is transferred to a sustained intracellular Ca 2ϩ elevation by activation of CRAC, resulting in membrane depolarization. Furthermore, the increased Ca 2ϩ concentration leads to long lasting changes in transcription processes and T cell proliferation. To maintain the driving force for Ca 2ϩ entry through the plasma membrane, the Ca 2ϩ -induced depolarization is counterbalanced by potassium channels. Our data suggests that ϳ40% of the K ϩ outward current in human CD3 ϩ T lymphocytes is mediated by K V 1.3, ϳ20% by IKCa1, and ϳ40% by K 2 P channels, a third and novel K ϩ conductance in human T cells. Based on their pharmacological profile (e.g. current inhibition by O 2 depletion or extracellular acidification) we hypothesize that TASK channels influence T cell actions under (patho-)physiological conditions (adapted from Ref. 42).
faster recovery (Fig. 7). Importantly, TASK channel modulation via anandamide resulted in very comparable results indicating that TASK-mediated effects on T lymphocytes are of pathophysiological relevance in a model of T cell-mediated autoimmunity. It is therefore tempting to speculate that attenuation of T cell function by selective pharmacological interference with TASK channels may translate into clinical benefits in T cell-mediated (auto)immune disorders. However, based on the broad expression pattern of TASK channels pharmacological treatment of EAE animals can only partially be linked to TASK channels expressed on immune cells because the effects might be contaminated through effects of TASK channels expressed, e.g. on neurons (13,23). As a note of caution, long term effects of ion channel blockers inhibiting TASK channels are not known. Therefore further work, including the detailed characterization of TASK1-and TASK3-knock out mice is clearly warranted.
In summary our study introduces TASK channels on T lymphocytes as critical components for the maintenance of resting membrane potential and T cell functions. Accordingly, channel modulation results in a marked reduction of the outward current in human T cells accompanied by significantly altered effector functions. T cell selective pharmacological blockade in an animal model of human multiple sclerosis serves as a proof of concept concerning the (patho)physiological relevance of these channels. Further studies are warranted to delineate the potential of TASK channels as a target and TASK channel blockers as potential drug candidates in T cell-mediated autoimmune diseases such as multiple sclerosis.