1Changes in the membrane potential (Vm) of human T lymphocytes upon K+ channel block were inferred from alterations in K+ current reversal potential in cell-attached patches. It was found that a high concentration of charybdotoxin (100 nm, CTX), which blocks both voltage-gated (K(V)) and Ca2+-activated (K(Ca)) potassium channels in these cells, depolarizes Vm of lymphocytes only partially. Subsequent whole-cell measurements of the same cells showed that 39 ± 25% of the voltage-gated current remains in the presence of CTX.
2The CTX-resistant current reverses at potentials between −80 and −90 mV, indicating that it is K+ selective. The current is activated at more depolarized potentials compared with the unblocked −IK(V) current with a threshold between −40 and −20 mV and a half-maximal activation at +50 mV. Inactivation during prolonged depolarization is slow. Steady-state inactivation is half-maximal at -45 mV and complete at potentials > − 20 mV. The CTX-resistant IK(V) is completely blocked by nifedipine and is not sensitive to dendrotoxin.
3The effect of nifedipine on the Vm of lymphocytes varies between cells depending on the contribution of the nifedipine-sensitive current to whole-cell IK(V) Combined application of CTX and nifedipine completely depolarizes Vm.
4The extent to which T cell receptor-evoked Ca2+ signals of resting cells are inhibited by K+ channel blockers correlates with the magnitude of the depolarization induced by the drugs. Complete suppression of the response is achieved only by combined block of the CTX-sensitive and -insensitive IK(V). The enhanced Ca2+ response of activated cells, which express increased numbers of K(Ca) channels, is in addition subject to modulation by blockers which prevent the hyperpolarization during the Ca2+ rise mediated by these channels.
Stimulation of the T cell receptor (TCR) of T lymphocytes induces an increase of the intracellular calcium concentration ([Ca2+]i), which is an essential step in the pathway towards T cell activation, ultimately leading to proliferation, differentiation or cell death (for review see Crabtree & Clipstone, 1994). The Ca2+ signal consists of two components: (1) the activation of phospholipase C following receptor stimulation evokes an inositol 1,4,5-trisphosphate (IP3)-mediated Ca2+ release from intracellular stores (Imboden & Stobo, 1985) and (2) by a mechanism termed capacitative Ca2+ entry (Putney, 1990), a signal from depleted stores subsequently activates a transmembrane Ca2+ conductance (Zweifach & Lewis, 1993), closely resembling ICRAC (Ca2+ release-activated Ca2+ current) in mast cells (Hoth & Penner, 1992). ICRAC exhibits voltage-independent gating but is inwardly rectifying resulting in a strong dependence of the Ca2+ influx on the membrane potential, Im, with hyper-polarization promoting Ca2+ influx.
The above observations suggest that, in addition to CTX-sensitive K+ channels, CTX-insensitive conductances contribute to Vm of lymphocytes. Lee, Levy & Deutsch (1992) showed that in human T cells the residual single channel events, which can be observed in the whole-cell configuration after inactivation of the n-type IK(V), were CTX resistant. The channel was activated at more depolarized potentials compared with the inactivating current.
In the present report, the CTX-resistant current of lymphocytes and its role in lymphocyte physiology are described in more detail. The effects of CTX on Vm of human T lymphocytes, measured from the reversal of K+ currents through cell-attached patches, are related to the extent of subsequent whole-cell K+ current block in the same cell. It is shown that under the present recording conditions a considerable component of the voltage-gated current is CTX resistant. This current was biophysically and pharmacologically further characterized and its contribution to Vm was assessed. In addition, the effects of block of the various K+ conductances on TCR-evoked Ca2+ signalling are shown.
Human peripheral blood lymphocytes were isolated from heparinized blood of healthy volunteers by centrifugation on a Ficoll–Hypaque density gradient (Pharmacia, Uppsala, Sweden). Adherence for 60 min to plastic culture flasks in the presence of 10 % fetal calf serum was used to eliminate most of the monocytes. The resulting cell populations contained 75–80% T lymphocytes as assessed by fluorescent labelling. Isolated cells were cultured at a density of 0.5 × 106 cells ml−1 in RPMI 1640 (Gibco, Paisley, UK) supplemented with 10% heat-inactivated fetal calf serum (Jacques Boy, Paris, France), 1 mn glutamine, 5 mM pyruvate, 100 μ ml−1 penicillin and 100 μg ml1 streptomycin (Gibco) and incubated at 37 °C in a humidified atmosphere containing 5% C02. To obtain activated T lymphocytes, 10 μml− phytohaemagglutinin-P (PHA-P, Difco, Detroit, MI, USA) was added for 24 h to the culture medium. Cells were kept in culture for up to 12 days and washed every 1–3 days.
Shortly before the experiments, cells were plated on a poly-l-lysine-coated coverslip mounted in a recording chamber (Warner Instruments, Hamden, CT, USA), which allowed complete solution changes within 5 s, and washed with external solution (mM: 160 NaCl, 4.5 KCl, 2 CaCl2, 1 MgCl2, 5 Hepes, adjusted to pH 7.4 with NaOH and 330 mosmol l−1 with glucose). In some experiments a high K+ bath solution was used consisting of (mm): 160 KCl, 2 CaCl2, 1 MgCl2, 10 Hepes, pH 7.4, 330 mosmol l−1
Intracellular [Ca2+] was measured at the single cell level as described previously (Verheugen et al. 1997). Briefly, cells were loaded with 2 μm fura-2 AM in culture medium for 15 min at 30 °C before plating them on the poly-l-lysine-coated coverslip. Images of the emission at 510 nm during alternating excitation at 340 and 380 nm were collected using an epifluorescence inverted microscope (Diaphot, Nikon, Chamfigny-sur-Marne, France) with a × 40 objective lens (Fluor, Nikon) and an intensified CCD camera (Hamamatsu, Paris, France). An image processor (Vprobe, ETM Systems, Irvine, CA, USA) averaged and stored background-subtracted 340 nm/380 nm ratio images on-line until analysis with the Vprobe software. Sixteen consecutive images (33 ms image−1) were averaged at each wavelength, and one ratio was stored every 10 s. Data were analysed off-line at the single cell level by averaging the non-zero 340/380 ratio values within a rectangle positioned over each cell in the field. The ratio (R) was converted into [Ca2+]i using the formula:
with Rmax, Rmin, and Kappt determined as described in Verheugen et al. 1997. All Ca2+ imaging experiments were done at room temperature (20–26 °C).
Currents were recorded in the cell-attached and whole-cell patch-clamp configurations. Pipettes (Clark; GC150 borosilicate glass) with a resistance of 4–6 MΩ were filled with a solution containing (mm): 120 KCl, 8.72 CaCl2, 1 MgCl2, 5 Hepes, 10 EGTA, adjusted with ∼40 KOH to pH 7.3 (final K+ concentration of 160 mm) and glucose to 320 mosmol l−1, unless stated otherwise. The free Ca2+ concentration of this pipette solution was calculated to be ∼1 μm (Partiseti, Korn & Choquet, 1993), which was confirmed by comparing the fluorescence of the pipette solution, to which 20 μm of the impermeant form of fura-2 was added, with that of calibration solutions buffered to different concentrations of free Ca2+ (World Precision Instruments, Sarasota, FL, USA). The high K+ pipette solution with 1 μm free Ca2+ was chosen in order to be able to measure both cell-attached and whole-cell currents from the same cell: a [Ca2+]i of 1 μm obtained after establishing the whole-cell configuration, is the optimum concentration for the activation of K(Ca) channels (Verheugen et al. 1995) and does not affect the K(V) channel (J. A. H. Verheugen, unpublished observations; see also Fig. 7). Data were collected using an Axopatch-1D amplifier controlled by pCLAMP 6 software (Axon Instruments). Currents were low-pass filtered at 2 kHz. The junction potential was zeroed before establishing the seal.
The membrane potential of T lymphocytes was determined from the reversal potential of K+ currents through the cell-attached patch (Verheugen et al. 1995): with a high K+ solution in the pipette the equilibrium potential for potassium (EK) across the patch is ∼0 mV and K+ currents will reverse when the pipette potential cancels Vm out. Voltage ramps (from Vh= -100 to +200 mV at 0.6 mV ms−1; note that for cell-attached recording Vh=−Vpipette) were applied to activate voltage-gated K+ channels and to establish their reversal potential (see Fig. 1A). For analysis of currents evoked by ramp stimulation a correction was made for a leak component by linear extrapolation of the closed level below the threshold for activation of the voltage-gated current (dotted lines in Fig. 1A). In between stimulations the patch was held at −60 mV (i.e. the patch potential was 60 mV hyperpolarized with respect to Vm) to remove K(V) channel inactivation. In addition to voltage-gated channels, Ca2+-activated K+ channels are present in most patches of activated T lymphocytes (see Fig. 2A), and their activity was used to confirm the fura-2 measurements of [Ca2+]i (which is proportional to the K(Ca) channel open probability) and the accuracy of the determination of the reversal potential of IK(v) (which intersects with the open level of the K(Ca) channel during the ramp). Note that the K(Ca) and K(V) channels in the cell-attached patch are shielded by the pipette from agents added to the bath.
Whole-cell IK(v) currents were evoked by depolarizing voltage steps or voltage ramps from a Vh of -80 mV, unless stated otherwise; the voltage-independent IK(Ca) was quantified as the slope of the current evoked by the voltage ramp below the threshold for activation of IK(V) (Grissmer et al. 1993; Partiseti et al. 1993). No series resistance compensation or leak subtraction was used. In contrast with the cell-attached recordings, the leak component of the whole-cell current was negligible compared with the macroscopic K+ currents. An interval of 30–60 s was maintained between subsequent stimulations to allow recovery from current inactivation. Data were analysed using Axograph 3 software (Axon Instruments). Tune constants of inactivation were determined by fitting whole-cell current traces with a single exponential decay function plus a steady level for IK(Ca) and non-inactivating IK(V) components. Using functions with multiple exponents did not yield better fits.
Data are expressed as means ±s.d. Results were compared using Student's two-tailed t test.
Charybdotoxin (CTX, synthetic peptide) and α-dendrotoxin (DTX, native peptide) were obtained from Latoxan (Rosans, France); nifedipine was from Sigma. All blockers were used at saturating concentrations (CTX at 100 nm; DTX at 100 nm; nifedipine at 100 μm), which was considered to be the case when a 5 times lower dose had the same effect. Fura-2 AM was purchased from Sigma.
To determine the effect of CTX on Vm and [Ca2]i of lymphocytes, fura-2-loaded T cells were patch clamped in the cell-attached configuration with a high K+ solution in the pipette. Vm was monitored from the reversal potential of IK(V) through the patch (Verheugen et al. 1995), which was activated every 10 s by applying a depolarizing voltage ramp (Fig. 1A). The fura signal was recorded simultaneously.
Figure 1B shows that the resting potential of lymphocytes averages −60mV (−56.2 ± 2.7 mV; n= 5 cells) with typical fluctuations of 20–30 mV, in agreement with previous results (Verheugen et al. 1995; Verheugen & Vijverberg, 1995). CTX, at an essentially saturating concentration of 100 nm, reversibly depolarized the membrane potential to ∼−35 mV in this cell. Although the extent of the CTX-induced depolarization varied between cells, from 3 to 22 mV (12.2 ± 8.4; n= 5 cells), Vm in the presence of CTX never dropped below an average level of −30 mV (-44.1 ± 9.8 mV; P < 0.01). The basal [Ca2+]i level was not affected by CTX (Fig. 1B).
In lymphocytes with elevated [Ca2+]i, evoked by physiological (TCR) or pharmacological (e.g. ionomycin) stimulation or by mechanical disturbance by the patch pipette (as was the case in the cell shown in Fig. 2), Ca2+-activated K(Ca) channels can be observed in addition to IK(v). In the cell-attached current traces K(Ca) channel activity is seen as the discrete open events at Vh=−60 mV (Fig. 2A). The K(Ca) channel activity results in a hyper-polarizion of Vm: in cells with [Ca2+]i close to 1 μm which is the optimum concentration for K(Ca) channel activation (Verheugen et al. 1995), Vm was in the order of −78.0 ± 3.5 mV (n= 5 cells). In addition, compared with the strong Vm fluctuations seen at low [Ca2+]i that are most probably the result of the periodic opening and closing of K(V) channels, Vm is held at a more steady level by the K(Ca) activity. Upon addition of CTX to the bath Vm depolarizes to −50.6 ± 6.1 mV (n= 5 cells; P < 0.001). In addition, [Ca2+]i gradually decreased in the presence of CTX (Fig. 2B). In contrast, when the bath solution was replaced with a high K+ solution a profound depolarization occurred to values between −20 and 0 mV, which was accompanied by a rapid decline in [Ca2+]i (Fig. 2B).
Following the cell-attached measurements, the whole-cell configuration was established to assess the effects of CTX on the whole-cell currents of the same cell (Figs 1C and 2C). With a pipette solution buffered to 1 μm free Ca2+, both IK(ca) and IK(V) can be measured after a 3–7 min period to allow complete perfusion of the intracellular compartment with the pipette solution. IK(V) currents were evoked by applying depolarizing voltage steps or voltage ramps. IK(Ca) is measured from the slope of the current below the threshold for activation of IK(V) (Grissmer et al. 1993; Partiseti et al. 1993). In the resting cell of Fig. 1, the zero-slope of the control current indicates the absence or a very low density of K(Ca) channels, whereas in the activated cell of Fig. 2 a clear K(Ca) component is present in the control current. IK(V) evoked by voltage steps has slow inactivation kinetics (time constant of inactivation (τinact) = 575 ± 257 ms; n= 33; Fig. 2C, inset) compared with IK(V) recorded with a KF–EGTA pipette solution, which is more commonly used in studies of the K(V) channel of lymphocytes (τinact≈ 150 ms; e.g. Cahalan et al. 1985; Verheugen, Oortgiesen & Vijverberg, 1994a; see Fig. 5C). IK(V) during a voltage ramp displays a typical sigmoidal onset and a progressively increasing deviation from a linear current–voltage relationship as a consequence of current inactivation (Figs 1C and 2C).
After equilibration of the cell interior with the pipette solution, 100 nm CTX was applied to the bath. Figures 1C and 2C show the steady-state effects of CTX that were attained within 1 min and which were fully reversible upon wash-out of the drug. CTX completely blocks IK(Ca), as is apparent from the disappearance of any current below the IK(V) threshold (Fig. 2C). In contrast, the voltage-gated current is reduced in the presence of CTX but not eliminated, leaving a CTX-resistant component intact (Figs 1C and 2C). The CTX-resistant current has a more linear current–voltage relationship during ramp stimulation compared with the unblocked current, corresponding to a much slower current inactivation (Fig. 2C, inset). The reversal potential of the CTX-resistant current was close to −80 mV (see below), indicating a K+-selective conductance. Thus, the modest effect of CTX on the membrane potential is reflected in its incomplete block of the whole-cell IK(V).
The amplitude of the voltage-dependent current remaining in the presence of 100 nm CTX varied widely between cells. The peak current evoked by a voltage step to +60 mV amounted to 219 ± 195 pA (range, 0–759 pA; n= 24). The CTX-resistant current typically represented ∼40% of total IK(V) (39 ± 25%; n= 24) but cells with a completely CTX-sensitive or with a completely CTX-resistant IK(V) were also observed. A clear relationship between the activation state of the cells and the size of the CTX-resistant current or its relative contribution to total IK(V) was not apparent.
Several lines of evidence argue against the possibility that the remaining current in the presence of CTX is merely due to a reduced potency of the toxin or an artefact arising from a membrane area that is inaccessible for the toxin, e.g. as a consequence of its contact with the coverslip. First, the CTX concentration used is saturating since 20 nm CTX blocked IK(V) to the same extent as l00 nm (data not shown). Furthermore, the observation that in some cells 100 nm CTX blocked IK(V) completely, whereas in some others IK(V) was entirely CTX resistant, is consistent with a variable contribution of distinct components to the total current. In addition, the K(Ca) conductance, which has a similar sensitivity to CTX as the n-type IK(V) (Sands et al. 1989; Grissmer et al. 1993) is invariably profoundly blocked by 100 nm CTX. As described below, the two IK(V) components have clearly distinct pharmacological profiles and different biophysical properties, which can also be observed in the absence of the toxin in those cells which have only one current component (e.g. Fig. 3B).
Characterization of CTX-insensitive IK(V)
The CTX-resistant and CTX-sensitive whole-cell currents are compared in Fig. 3A. The CTX-sensitive components were obtained by subtracting the current that remains in the presence of 100 nm CTX from the control current. When comparing the two IK(V) components, the slower inactivation rate of the CTX-resistant IK(V) is the most conspicuous difference. In addition, the inward tail current upon repolarization to −80 mV disappears in the presence of CTX. The voltage dependence of the two current components is strikingly different. In Fig. 3C (upper panel), the conductance, calculated from the ramp currents, is plotted against the voltage. These data show that the threshold for activation of the CTX-resistant current upon voltage ramp stimulations is at a 5–15 mV more positive potential. Maximal activation of the CTX-resistant current is at extremely positive potentials of > +80 mV, whereas the CTX-sensitive current is maximally activated at approximately 0 mV (see also Fig. 4A).
In cells that lack a CTX-sensitive IK(V) component, the same characteristics of the CTX-resistant current are also seen in the absence of CTX (Fig. 3B and C, lower panel). This argues strongly against the possibility that the current remaining in the presence of CTX is due to unblocking of CTX at depolarized potentials as described for large conductance K(Ca) channels (MacKinnon & Miller, 1988).
Figure 4 gives an overview of the voltage dependence of activation (Fig. 4A), deactivation (Fig. 4B) and steady-state inactivation (Fig. 4C) of the CTX-resistant current. Current–voltage (I–V) relationships of the control current and the CTX-resistant current were obtained by applying depolarizing voltage steps from −60 to +100 mV in 20 mV increments from a holding potential of −80 mV (Fig. 4A). Whereas the unblocked current is activated with a threshold at -50 to -40 mV, the CTX-resistant current has an activation threshold between −40 and −20 mV. Plotting the peak current vs. voltage shows that the reduction of IK(V) by CTX occurred over the entire voltage range. The voltage dependence of control and CTX-resistant IK(V) was determined by plotting normalized peak conductance vs. potential. Whereas the unblocked current is half-maximal at a potential between −40 and −20mV, half-maximal activation of the CTX-resistant IK(V) was at approximately +50 mV.
The reversal potential of the CTX-resistant current was between −80 and −90 mV as determined from the tail currents upon repolarization to different potentials after a 100 ms conditioning pulse to +60 mV (Fig. 4B). Under the present experimental conditions, the only ionic species with such a negative Nernst potential is K+, indicating that the CTX-insensitive voltage-gated conductance is K+ selective. The deactivation rate upon repolarization changed in proportion with the potential, accelerating at more negative potentials. At potentials of −80 mV and below, deactivation of the CTX-resistant current was faster compared with the unblocked IK(V), resulting in a decrease of the tail current in the presence of CTX (Figs 3 and 4A).
Inactivation of the CTX-resistant current was incomplete during prolonged depolarizations of many seconds. Nevertheless, the current showed clear steady-state inactivation when a depolarized holding potential was maintained for ∼30 s (Fig. 4C): at Vh=−45 mV, half of the channels are inactivated, and steady-state inactivation is essentially complete at Vh=−20 mV. In comparison, steady-state inactivation of the CTX-sensitive n-type K(V) channel occurs at more hyperpolarized potentials with a mid-point at −60 to −70 mV and complete inactivation at potentials of −30 mV (Cahalan et al. 1985).
Pharmacology of the CTX-resistant K(V) channel
The sensitivity of the CTX-resistant current to dendrotoxin (DTX) and nifedipine, drugs which selectively block certain subclasses of K+ channels (Grissmer et al. 1994), was assessed. Whereas the unblocked IK(V) of lymphocytes showed some sensitivity to 100 nm DTX (19 ± 13% block; n= 6; data not shown), the CTX-resistant component was completely insensitive to this drug (Fig. 5A). In contrast, whereas 100 μm nifedipine also only partially blocked total IK(V) (by 63 + 25%; n= 8; Fig. 5B), this drug fully blocked the CTX-resistant current. Therefore, a virtually complete block of the voltage-gated current was obtained in the presence of both CTX and nifedipine (Fig. 5A and B).
Although the effects of CTX and nifedipine were additive, they were not mutually exclusive, i.e. the CTX-sensitive component also displayed some sensitivity to nifedipine. In cells with a 100% CTX-sensitive JK(V) current, 100 μm nifedipine blocked 40% of the current (n= 2). In addition to its effect on the voltage-gated conductances, nifedipine at 100 μM partially blocked the K(Ca) conductance (Fig. 5B). The effects of CTX and nifedipine on the K+ currents are largely reversible, except that inactivation kinetics of IK(V) after wash-out of nifedipine remain usually somewhat faster than before application of the drug (Fig. 5A, see also Fig. 7).
In many of the previous studies of whole-cell voltage-gated K+ currents in lymphocytes, pipette solutions contained KF and EGTA (and a very low free [Ca2+]i, which tend to stabilize the seal (e.g. Cahalan et al. 1985; Verheugen et al. I994a). Therefore, the sensitivity of IK(V) to CTX and nifedipine was also investigated under conditions similar to these previous studies (Fig. 5C). In agreement with these studies, the control IK(V) had a faster inactivation rate compared with the IK(V) seen with pipette solution buffered to 1 μm free [Ca2+]i CTX (100 nm) largely blocked IK(V) but a CTX-resistant component remained consistently, amounting to 23 ± 12% (n= 1) of the control current, which appears somewhat smaller than the resistant component seen with a 1 μm free [Ca2+]i solution in the pipette (P < 0.05). Again, the CTX-resistant component was sensitive to nifedipine and IK(V) was fully blocked in the simultaneous presence of the two drugs (Fig. 5C).
Effects of nifedipine on Vm and [Ca2+]i of lymphocytes
To assess the contribution of the nifedipine-sensitive K+ conductance to the Vm of lymphocytes, similar experiments to those described above for CTX were performed with nifedipine, i.e. Vm and [Ca2+]i were monitored, from cell-attached currents and the fura signal, respectively, in the absence and presence of nifedipine before recording whole-cell currents from the same cell. As illustrated in Figs 6 and 7, the effects of nifedipine on Vm and [Ca2+]i depended on the extent of its voltage-gated current block, which varied from cell to cell.
An example of a cell in which nifedipine had little or no effect is given in Fig. 6. In this cell, Vm under control conditions depended on the level of [Ca2+]i as described above: whereas at low [Ca2+]iVm fluctuated strongly around an average value of about –50 mV, during a spontaneous transient elevation of [Ca2+]i K(Ca) channel activity stabilized Vm at approximately -75 mV (Fig. 6B). Nifedipine was applied both at periods of high and of low [Ca2+]i In neither situation did the drug have any effect on Vm or on [Ca2+]i (Fig. 6B). Subsequent whole-cell recording showed that a considerable IK(V) current could be evoked in the presence of nifedipine (Fig. 6C).
In contrast, strong effects of nifedipine were observed in the cell shown in Fig. 7. In this cell, nifedipine (applied after elevation of [Ca2+]i with ionomycin) induced a complete depolarization to values close to 0 mV and a rapid decrease of [Ca2+]1 (Fig. 1B). In agreement with the strong effects on Vm, a complete block of IK(V) by nifedipine was observed in subsequent whole-cell measurements of this cell (Fig. 7C).
In two other cells nifedipine had intermediate effects, depolarizing Vm from control levels of −43 ± 7 and −55 ± 4 mV to, respectively, −22 ± 10 and -38 ± 7 mV in the presence of the drug, corresponding to a partial block of the whole-cell IK(V) in these cells (data not shown).
Combined application of CTX and nifedipine depolarizes Vm in every cell to values between −20 and 0 mV (data not shown) in agreement with the complete block of IK(V) under these conditions (see Fig. 5).
Consequences of selective K+ channel block for TCR-evoked Ca2+ signaling
The Ca2+ signal evoked by T cell receptor stimulation depends critically on the level of Vm (Gelfand, Cheung & Grinstein, 1984; Oettgen, Terhorst, Cantley & Rosoff, 1985; Gelfand, Cheung, Mills & Grinstein, 1987; Hess, Oortgiesen & Cahalan, 1993). Since the present results revealed the contribution of distinct K+ conductances to the Vm of lymphocytes, the modulation of PHA-evoked Ca2+ responses by selective K+ channel antagonists was studied. In Fig. 8, Ca2+ responses upon PHA stimulation are shown under control conditions and in the presence of CTX and nifedipine, either applied separately or in combination. For comparison, the effect on the Ca2+ signal of high extracellular [K+], which completely depolarizes Vm, is shown. Since the expression of K+ channels, in particular the CTX-sensitive K(Ca) channel, is strongly upregulated following T cell activation, these experiments were performed both in resting and activated T cells. Figure 8A shows representative responses of individual cells under the different conditions as well as the averages of 40–100 cells in the same experiment. The average Ca2+ levels under the various conditions in the period between 600 and 1500 s after PHA application are compared in Fig. 8B.
Compared with the modest [Ca2+]i elevations upon PHA stimulation in resting cells a strongly enhanced Ca2+ response was seen on secondary TCR stimulation in activated cells (Fig. 8A, top panels) in agreement with previous results (Verheugen et al. 1997). The effects of the K+ channel antagonists on Ca2+ signalling varied with the activation state of the cell (Fig. 8A, middle panels): CTX and nifedipine applied alone hardly affected the Ca2+ response in resting cells. However, the upregulation of the Ca2+ response in activated cells was largely prevented by CTX and, to a lesser extent, by nifedipine. On the other hand, nifedipine and CTX applied together almost completely abolished the Ca2+ response in every cell, regardless of their activation state. An extensive depolarization induced by high extracellular K+ was only slightly more effective in suppressing the Ca2+ response (Fig. 8, lower panels).
Upon washout of K+ channel blockers, an overshoot of [Ca2+]i is generally observed (Fig. 8A). This is most probably the result of a strongly activated ICRAC pathway as a consequence of a continuous signal from empty Ca2+ stores that were unable to be refilled with Ca2+ from the extracellular environment because of membrane depolarizations (Hess et al. 1993). Therefore, the extent of the overshoot could reflect the degree of depolarization. This is in agreement with the large increase in [Ca2+]1 upon return to normal bath solution after high K+ (Fig. 8A, lower panels). While the overshoot is almost as pronounced after washout of CTX + nifedipine, it is less pronounced after nifedipine alone and basically absent following CTX treatment.
Taken together these data indicate that in resting cells the extent to which the Ca2+ signal is inhibited by K+ channel blockers correlates with the magnitude of the depolarization induced by the drugs. In activated cells, on the other hand, CTX and nifedipine had stronger inhibitory effects on the Ca2+ signal even when applied alone, despite the modest depolarization as judged from the [Ca2+]i overshoot. This suggests that in activated cells the primary inhibitory effect of K+ channel block on Ca2+ signalling arises from the prevention of hyperpolarization rather than from depolarization. This hyperpolarization during the [Ca2+]i rise is mediated by K(Ca) channels which are blocked completely by CTX and partially by nifedipine (see Fig. 5).
The present results show that at least three distinct potassium conductances contribute to the membrane potential of human T lymphocytes: in addition to the previously described voltage-gated n-type channel and the Ca2+-activated IK-type channel, which are both charybdotoxin sensitive, a CTX-resistant voltage-gated K+ channel is present, constituting, under the present recording conditions, on average 40% of the whole-cell voltage-gated current. The voltage dependences of both activation and (steady-state) inactivation of the CTX-resistant current are shifted to more depolarized potentials compared with the retype channel. The current has slow inactivation kinetics compared with the CTX-sensitive current.
The existence of a CTX-resistant K+ channel in human T lymphocytes has been proposed before, based on the observation that CTX only partially inhibits the regulatory volume decrease (RVD) induced by hyposmotic solutions (Grinstein & Smith, 1990). Moreover, some electro-physiological data on a CTX-resistant channel in T cells have been reported (Lee et al. 1992). Nevertheless, expression of the Kv1.3 gene in Xenopus oocytes (Grissmer et al. 1990; Cai et al. 1992; Attali et al. 1992) or the murine T cell line CTLL-2 (Deutsch & Chen, 1993) induces a voltage-gated current with properties closely resembling those of the native n-type current in human lymphocytes, suggesting that the channels underlying the lymphocyte current are the product of a single gene. However, in these studies the whole-cell IK(V) currents of lymphocytes were recorded with pipette solutions containing fluoride or with a very low free [Ca2+]. Under these conditions, properties of IK(V) change in the period immediately after establishment of the whole-cell configuration (Cahalan et al. 1985; Lee et al. 1992). For example, the voltage dependence shifts to more negative values and the rate of inactivation becomes faster. In addition, there is a change from incomplete to complete inactivation. In contrast, voltage-gated currents recorded under conditions which avoid extensive intracellular dialysis, e.g. cell-attached (Verheugen et al. 1995) or perforated-patch measurements (Oleson, DeFelice & Donahoe, 1993; but see also Chung & Schlichter, 1993) have kinetic properties which resemble those of whole-cell −k(v) immediately after break-in. Furthermore, with the pipette solution used in the present study, which does not contain fluoride and has a free [Ca2+] of 1 μm, the whole-cell -K(v) retains its initial kinetic properties of slow and incomplete inactivation and the only change in whole-cell currents in the first period after break-in is a gradual activation of the K(Ca) channels, in parallel with the perfusion of the cell with the 1 μm free Ca2+ solution (J. A. H. Verheugen, unpublished observations). With the latter pipette solution the CTX-resistant current constitutes in general a larger fraction of total IK(V) compared with the current recorded with a KE solution (present results). Taken together, these observations suggest that a component of the voltage-gated current of human T lymphocytes, corresponding to the CTX-resistant current described in the present paper, is inhibited by intracellular dialysis with a fluoride–EGTA solution. The observation that CTX depolarizes Vm to a lesser extent at elevated [Ca2+]i (Figs 1 and 2) further suggests that the CTX-resistant K(V) conductance of lymphocytes is somehow stimulated by [Ca2+]i
Frequently, voltage-gated currents in cell-attached patches consist of two distinct components with a slightly different voltage dependence (Verheugen et al. 1995; see also Fig. 1). It is conceivable that these components represent the CTX-sensitive and CTX-resistant IK(V) current, respectively.
Whereas the CTX-sensitive current is partially blocked both by nifedipine and dendrotoxin, the CTX-resistant current is blocked completely by nifedipine and is completely insensitive to dendrotoxin (Fig. 5). This pharmacological profile differs from that found in murine thymocytes, where IK(V) could be subdivided into a CTX-sensitive and a dendrotoxin-sensitive component (Freedman, Fleischmann, Punt, Gaulton, Hashimoto & Kotlikoff, 1995). Whereas the Kv1.1 gene was shown to underlie the CTX-resistant current in murine thymocytes (Freedman et al. 1995), the molecular nature of the CTX-resistant current in human cells remains to be determined. The pharmacological properties of this current as well as kinetics of inactivation and voltage dependence are reminiscent of the current induced by expression of the Kv3.1 gene (Grissmer et al. 1994). Another possible candidate could be the IsK protein, which was cloned from cDNA from lymphocytes and which induced a non-inactivating, CTX-insensitive K+ current when expressed in Xenopus oocytes but which could not be recorded directly in lymphocytes (Attali et al. 1992). It was subsequently shown that IsK acts as a K+ channel subunit rather than forming a complete channel itself (Barhanin, Lesage, Guillemare, Fink, Lazdunski & Romey, 1996; Sanguinetti et al. 1996). The IsK-induced K+ current is increased when [Ca2+]i is raised (Attali et al. 1992; Barhanin et al. 1996), which would be consistent with a current which can be recorded with an elevated [Ca2+] in the pipette but which is suppressed under low [Ca2+]i conditions.
Consequences of potassium channel block for the membrane potential and Ca2+ signalling
In the present study, the reversal potential of the voltage-gated IK(V) current in cell-attached patches was used to infer the level of the membrane potential of the cell and to assess the extent of the depolarizations induced by K+ channel block. The cell-attached measurements showed that the resting Vm of T cells fluctuates around -60 mV (present results; Verheugen & Vijverberg, 1995), which is close to the activation threshold of the CTX-sensitive K(V) channel. The conspicuous fluctuations of Vm are consistent with the opening and closing of single K(V) channels. At elevated [Ca2+]i e.g. during the Ca2+ response upon TCR stimulation, these fluctuations disappear and Vm settles at a more hyper-polarized level as a consequence of the activation of K(Ca) channels (Figs 2, 6 and 7). When the CTX-sensitive K(V) and K(Ca) channels are blocked, Vm shifts to a moderately depolarized level of −35 to −50 mV (Figs 1 and 2), i.e. the threshold of the CTX-resistant K(V) channel, which is activated at slightly more depolarized potentials compared with the CTX-sensitive K(V) channel. Block of the CTX-resistant IK(V) by nifedipine depolarizes Vm to a variable extent, depending on the relative contribution of this current to the total IK(V) (Figs 6 and 7).
The results of Fig. 8 show that the inhibition of the TCR-evoked Ca2+ response by K+ channel blockers is proportional to their effect on the membrane potential. The moderate depolarization by CTX in resting T lymphocytes does not alter appreciably the Ca2+ signalling of these cells. In contrast, more profound depolarization by a combination of CTX and nifedipine prevents even the small influx required for the modest Ca2+ responses of resting cells. In contrast, CTX strongly reduces the enhanced Ca2+ response in activated cells. This suggests that the increased Ca2+ influx depends critically on the additional hyper polarization, mediated by K(Ca) channels which are expressed at elevated levels in these cells, and which are blocked completely by CTX and partially by nifedipine (Fig. 5).
Nifedipine is traditionally used as a blocker of voltage-gated Ca2+ channels. However, it is unlikely that the effects of nifedipine on Ca2+ signalling in T lymphocytes can be attributed to block of Ca2+ channels since human T lymphocytes do not seem to express voltage-gated Ca2+ channels (for review see Lewis & Cahalan, 1995) and the voltage-independent Ca2+ conductance, ICRAC, of these cells is not sensitive to classical inhibitors of voltage-gated Ca2+ channels (Fasolato, Innocenti & Pozzan, 1994). Synergistic effects of cyclosporin A and calcium antagonists such as nifedipine, diltiazem and verapamil in inhibiting lymphocyte proliferation (Marx, Weber, Merkel, Meyer-zum-Buschenfelde & Kohler, 1990) are probably a consequence of the sensitivity of lymphocyte K+ channels to these agents (Lewis & Cahalan, 1995; present results).
Suppression of immune functions by potassium channel block has been the subject of numerous previous studies (reviewed in Lewis & Cahalan, 1995). It is generally assumed that effects of channel block on T cell functioning arise from reduced Ca2+ signalling as a consequence of membrane depolarization, which subsequently hampers interleukin-2 production. Whereas the inhibitory effects of specific K+ channel block, e.g. by CTX, are in general quite modest (e.g. Price et al. 1989; Gelfand & Or, 1991; Freedman et al. 1992; Verheugen et al. 1997), less specific blockers such as tetraethylammonium, 4-aminopyridine or quinine can completely suppress lymphocyte functioning (Chandy et al. 1984). This difference can be explained by the contribution of a CTX-resistant K+ channel, characterized in the present study, to the membrane potential of human T lymphocytes. Onr observations that CTX affects primarily Ca2+ signalling in activated T cells and that additional channel block can bring about complete suppression of the Ca2+ response in both resting and activated cells shows the possibility of selectively modulating subsets of cells, thus supporting the notion that K+ channel antagonists might be used as selective immunosuppressants.
We thank Drs Alain Trautmann and Pjotr Bregestovski for their comments on the manuscript. This work was supported by a research grant from the Human Frontiers Science Organisation to J. Verheugen (LT46/95) and by Direction de la Recherche et de la Technologie (DRET; DGA g94-101).