Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Fluoxetine has become an important antidepressant drug during the past two decades because it was shown to be effective in the treatment of depression without inducing serious side effects (Stark et al., 1985; Fuller and Wong, 1990; Wong et al., 1995). Its mechanism of action is to inhibit the reuptake of serotonin into the synaptic cleft [selective serotonin reuptake inhibitor (SSRI)] (Wong et al., 1974). This inhibition is thought to underlie the therapeutic effects of this drug. However, in addition to this pharmacological property, this drug has been reported to influence a variety of ion channels, including voltage-activated K⁺ channels. The inhibition of ionic currents mediated through voltage-activated K⁺ and Na⁺ channels by fluoxetine was first demonstrated by Rae et al. (1995) in rabbit corneal and human lens epithelium. This inhibitory effect of fluoxetine on K⁺ currents was observed only in perforated-patch recordings, not in inside-out patch recordings. A similar mechanism has been proposed to explain the inhibitory effect of fluoxetine on delayed rectifier K⁺ channels of jejunal smooth muscle cells (Farrugia, 1996). Therefore, it is likely that fluoxetine acts on K⁺ channels through diffusable cytoplasmic factors, suggesting an indirect action. However, our previous study showed that inhibitory effects of fluoxetine on K⁺ currents in PC12 cells did not appear to be mediated through protein kinases or G proteins (Hahn et al., 1999). Recent studies suggest direct interaction of fluoxetine with ion channels. In a study examining the cloned nicotinic acetylcholine receptor, fluoxetine was proposed to produce a block by interacting with open channels (García-Colunga et al., 1997). Tytgat et al. (1997) has shown that the inhibition by fluoxetine of the Kv1.1 channel could be described by blockade of the open state of the channel. However, it should be noted that therapeutic doses of fluoxetine result in a plasma concentration closer to 1 µM (Altamura et al., 1994). All of these findings have little relevance in understanding the clinical effects of fluoxetine because much higher concentrations are necessary to inhibit K⁺ channels than are needed for an SSRI effect. Although there are at least two pathways by which fluoxetine could exert such an effect, the mechanism of block of K⁺ channels is not clearly understood. Thus, we decided to explore what interactions fluoxetine might have with Kv1.3 as well as with corresponding K⁺ currents found in human T lymphocytes. Here we report that fluoxetine at clinically relevant concentrations blocks Shaker-type K⁺ channel, Kv1.3, which plays an important role in the activation of T lymphocytes.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Cell Culture. We used the stable Chinese hamster ovary (CHO) cell line expressing Kv1.3 that has been described in detail elsewhere (Hahn et al., 1996). Cells were cultured in Iscove's modified Dulbecco's medium (GIBCO BRL, Grand Island, NY) supplemented with 10% FBS, 0.1 mM hypoxanthine, 0.01 mM thymidine, and 500 µg/ml G418 (GIBCO BRL) in a humidified 5% CO₂ incubator at 37°C. The cultures were passed every 3 to 5 days, using brief trypsin treatment. T lymphocytes were purified from peripheral blood cells from healthy volunteers. Peripheral blood mononuclear cells were isolated by density gradient centrifugation on Ficoll-Hypaque (Sigma Chemical Co., St. Louis, MO). After 1 h of adherence to nylon wool (Robbin Scientific, Sunnyvale, CA) in RPMI medium (GIBCO BRL), nonadherent cells were collected. Cells were maintained in RPMI containing 10% FCS. Cells used for electrophysiological experiments were seeded onto glass coverslips (diameter, 12 mm; Fisher Scientific, Pittsburgh, PA) in a Petri dish for 24 to 48 h before use. On each experimental day, coverslips with attached cells were transferred to a continually perfused recording chamber (RC-13; Warner Instrument Corporation, Hamden, CT).

Electrophysiology. Voltage-clamp recordings were performed using the whole-cell and excised inside-out configurations of the patch-clamp technique (Hamill et al., 1981) at room temperature with an Axopatch 200B amplifier (Axon Instruments, Foster City, CA). Micropipettes were fabricated from PG10165-4 glass capillary tubes (World Precision Instruments, Sarasota, FL) and had a resistance of 2 to 4 M when filled with pipette solution. Capacitative currents were compensated with the analog compensation. Linear leak currents were not compensated. Series resistance was approximately 5 to 10 M, and series resistance compensation (70-80%) was used in whole-cell recordings if the current exceeded 1 nA. Currents were filtered at 2 kHz (four-pole Bessel filter) and sampled at 5 kHz. Pulse generation, data acquisition, and analysis were performed with an IBM pentium computer, using the pClamp 6.03 software (Axon Instruments). For whole-cell recordings, the electrodes were filled with a solution containing 140 mM KCl, 1 mM CaCl₂, 1 mM MgCl₂, 10 mM HEPES, and 10 mM EGTA (pH 7.3 with KOH). This solution served as the bath solution for inside-out patches. The bath solution for whole-cell recordings contained 140 mM NaCl, 5 mM KCl, 1.3 mM CaCl₂, 1 mM MgCl₂, 20 mM HEPES, and 10 mM glucose (pH 7.3 with NaOH). This solution was used as the pipette solution for inside-out patches. During the recording, the cells were continuously perfused at a rate of 1 ml/min with control or drug-containing solutions. Fluoxetine was obtained from Tocris Cookson (Bristol, UK). Norfluoxetine was obtained from Research Biochemicals Inc. (Natick, MA).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Figure 1A shows representative recordings of the Kv1.3 current expressed in CHO cells. Under control conditions, the Kv1.3 current rapidly rose to a peak and displayed slow inactivation during a 200-ms depolarizing pulse as reported previously (Hahn et al., 1996). In the presence of fluoxetine (1 and 10 µM), the peak amplitude of Kv1.3 currents was slightly reduced and there was an acceleration in the apparent rate of current decay. The current was initially activated in a manner similar to that of the control but subsequently declined markedly. This decline occurred much faster and to a greater extent than the slow inactivation observed in the control (156.4 ± 9.9 ms for control; 89.3 ± 6.5 ms for 10 µM fluoxetine, n = 7). Therefore, the reduction in Kv1.3 currents at the end of the pulse was used as an index of inhibition. The concentration dependence of the inhibition of Kv1.3 induced by fluoxetine, in a range of concentrations between 1 and 30 µM, is presented in Fig. 1B. Nonlinear least-squares fit of the data yielded an apparent IC₅₀ value of 5.9 µM with a Hill coefficient of 1.3. 

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Concentration-dependent inhibition of Kv1.3 by fluoxetine. A, whole-cell Kv1.3 currents were elicited by 200-ms step depolarizations to +40 mV from a holding potential of 80 mV at 30-s intervals. Control current and current after the addition of 1 and 10 µM fluoxetine are shown. B, concentration-response curve for Kv1.3 inhibition by fluoxetine. Reduction in current (relative to control) at the end of depolarizing steps from 80 to +40 mV was used as an index of inhibition. Nonlinear least-squares fit of the data yielded an IC50 value of 5.9 µM and a Hill coefficient of 1.3 (n = 7).

Current-voltage relations for Kv1.3 currents at the end of the voltage step, as in Fig. 2, indicated that fluoxetine reduced Kv1.3 currents for the entire voltage range over which this current was activated (Fig. 2C). To quantify the effects of voltage on the drug-channel interaction, the relative current (I_fluoxetine/I_control) was plotted as a function of membrane potential together with the activation curve (Fig. 2D). The current was activated at 40 mV and the conductance of the channel was fully saturated at 0 mV. In the presence of 10 µM fluoxetine, the inhibition increased steeply between 20 and 0 mV, which corresponded with the voltage range for channel opening. Over potentials where conductance is saturated (0 to +40 mV), inhibition continued to increase with a shallow voltage dependence. For instance, with 10 µM fluoxetine, the degree of inhibition increased in a voltage-dependent manner from 54.6 ± 5.9% at 0 mV to 64.8 ± 3.5% at +40 mV (p < .05, ANOVA, n = 6). This result indicates that fluoxetine binds primarily to the open channel. By fitting these data according to the Woodhull model, the calculated fractional electrical distance () was 0.29. 

View larger version (30K): [in this window] [in a new window]   Fig. 2.   Effects of fluoxetine on Kv1.3 current-voltage relationships. Whole-cell currents elicited by 200-ms test pulses to potentials between 50 and + 40 mV in steps of 10 mV under control conditions (A) and after the addition of 10 µM fluoxetine (B). C, resultant current-voltage relationships taken at the end of the test pulses in the absence () and presence () of fluoxetine. D, voltage dependence of current inhibition by fluoxetine. Normalized inhibition shown as relative current (Ifluoxetine/Icontrol) from data in C. The dotted line represents the activation curve of control Kv1.3. The solid line represents the result of the fit of equation (see Materials and Methods).

The voltage dependence of activation and steady-state inactivation in the presence of fluoxetine was evaluated to investigate whether the inhibition was due to a shift of activation and inactivation curves (Fig. 3). The activation curve of Kv1.3 was not affected by 10 µM fluoxetine. V₅₀ and k were 23.3 mV and 6.9 for control and 24.4 mV and 6.1 in the presence of 10 µM fluoxetine, respectively (Fig. 3A). A similar lack of effect of fluoxetine on the voltage dependence of steady-state inactivation was observed. In the absence of drug, V₅₀ and k were 44.6 mV and 5.2, respectively. In the presence of 10 µM fluoxetine, V₅₀ and k were not significantly changed and were 47.2 mV and 4.9, respectively (Fig. 3B).

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Lack of effect of fluoxetine on activation and steady-state inactivation properties. A, activation curves in the absence () and presence of () fluoxetine (n = 6). The activation curve was obtained from deactivating tail current amplitude at 50 mV after 100-ms depolarizing steps to potentials between 50 to +20 mV in steps of 5 mV from a holding potentials of 80 mV. Data show the conductance (G) at each step voltage standardized to the conductance at +20 mV (Gmax). B, steady-state inactivation curves in the absence () and presence () of fluoxetine (n = 4). Data show the current activated by a test step to +40 mV after a 30-s conditioning prepulse at different voltages. The peak current amplitude in the test step (I) was normalized to the peak amplitude measured after a prepulse at 80 mV (Imax) and is plotted against prepulse voltage. Curves are least-squares fit of a Boltzmann equation (see Materials and Methods).

Figure 4 shows the effects of fluoxetine on Kv1.3 current deactivation kinetics. The reversal potential for Kv1.3 currents was approximately 80 mV and was not changed in the presence of 10 µM fluoxetine (Fig. 4A). To compare the time course of decay of the tail current, outward tail currents were recorded at a potential of 50 mV after a 200-ms step depolarization to +40 mV in control conditions and in the presence of 10 µM fluoxetine and have been superimposed (Fig. 4B). In the absence of fluoxetine, Kv1.3 currents deactivated with a time constant of 15.6 ± 1.4 ms (n = 4). In the presence of fluoxetine, the tail current amplitude was reduced and the subsequent decline of the current was slower (24.1 ± 2.5 ms, n = 4, p < .05, Student's t test) than in control conditions, which resulted in a crossover phenomenon. These results also support an open channel interaction between fluoxetine and the Kv1.3 channel.

View larger version (22K): [in this window] [in a new window]   Fig. 4.   Effects of fluoxetine on Kv1.3 deactivating tail currents. A, cells were depolarized to +40 mV for 100 ms from a holding 80 mV and then repolarized to various voltages between 95 and 60 mV in steps of 5 mV under control conditions (left) and in the presence of 10 µM fluoxetine (right). B, whole-cell Kv1.3 currents were elicited by 200-ms step depolarizations to +40 mV from a holding potential of 80 mV. The cells were returned to 50 mV to induce outward tail currents. The tail currents were well fitted to a single exponential decay. In the presence of 10 µM fluoxetine, the initial amplitude was reduced and the subsequent slower decline resulted in the crossover phenomenon with the control tracing. The dotted line represents the zero level of current. Inset, mean time constants of the tail current decay under control conditions and in the presence of fluoxetine. *p < .05 (Student's t test, n = 4).

At this point, it is important to know whether fluoxetine blockade of Kv1.3 occurred from the intracellular or extracellular side of the membrane. In an attempt to determine on which side of the membrane Kv1.3 channels could be blocked by fluoxetine, the time courses for onset of inhibition by 10 µM fluoxetine under whole-cell and inside-out recordings were compared (Fig. 5). Cells were repetitively pulsed from 80 to +40 mV every 30 s. Both phases were well described by a single exponential decay function with on time constants (_on). For whole-cell recordings, _on was 43.9 ± 5.0 s (n = 5). The time course for inhibition with inside-out recordings was significantly faster, and _on was 17.4 ± 1.8 s (n = 4). In addition, in the presence of fluoxetine at the same concentrations, the current was greatly reduced in inside-out recordings compared with that in whole-cell recordings. One explanation for the short exposure times required to achieve steady-state inhibition and the increased sensitivity by fluoxetine in inside-out recordings compared with whole-cell recordings is that the site of action of this drug is on the intracellular side of Kv1.3 channels. To examine the possibility that fluoxetine had an intracellular site of action, inside-out recordings were performed to confirm this interpretation. The averaged normalized current values obtained in inside-out patches in the presence of various concentrations of fluoxetine are shown in Fig. 6. The drug concentration needed to induce 50% current inhibition in inside-out patches was 1.7 µM (Fig. 6B). Compared with whole-cell recordings, Kv1.3 currents in inside-out patches were more sensitive to fluoxetine. These results strongly suggest that the inhibition of Kv1.3 by fluoxetine is due to binding to an intracellular site. In addition, in inside-out patches, the main effect of fluoxetine was to accelerate the rate of Kv1.3 current decay, whereas the peak current was only slightly reduced, as was shown in whole-cell recordings (Fig. 6A). Under control conditions, Kv1.3 current decay was well fitted to a single exponential with a time constant of 112.6 ± 11.6 ms. After the addition of 1, 3, and 10 µM fluoxetine, Kv1.3 current decay was significantly accelerated and was 74.9 ± 6.2, 58.4 ± 5.6, and 39.4 ± 3.3 ms, respectively (Fig. 6C).

View larger version (21K): [in this window] [in a new window]   Fig. 5.   Comparison of time courses for onset of Kv1.3 inhibition by fluoxetine in the whole-cell () and inside-out () configurations. Currents were induced by 200-ms step depolarizations to +40 mV from a holding potential of 80 mV at 30-s intervals. The amplitudes of currents were plotted as a function of time. All currents were measured at the end of the pulse. The time constant of onset of inhibition (on) was obtained from monoexponential fits to data. The bar indicates the time of application of fluoxetine.

View larger version (18K): [in this window] [in a new window]   Fig. 6.   Effects of fluoxetine on Kv1.3 recorded from inside-out patches. A, Kv1.3 currents recorded from an inside-out patch were elicited by 200-ms step depolarizations to +40 mV from a holding potential of 80 mV at 30-s intervals in the absence and presence of 1 and 10 µM fluoxetine. B, concentration-response curve for Kv1.3 inhibition by fluoxetine. Reduction of current (relative to control) at the end of depolarizing steps from 80 to +40 mV was used as an index of inhibition. Nonlinear least-squares fit of the data yielded an IC50 value of 1.7 µM and a Hill coefficient of 0.9 (n = 7). C, effects of fluoxetine on Kv1.3 inactivation kinetics. The time constants were estimated from single exponential fits to the decay phase. The time constant obtained at +40 mV has been plotted against fluoxetine concentration (n = 7).

After administration, fluoxetine is metabolized with norfluoxetine as the major metabolic product; thus, it is of interest to examine whether norfluoxetine also inhibits Kv1.3 channels. Figure 7 shows the effects of norfluoxetine on Kv1.3 in whole-cell recordings. From the analysis of the concentration-response curve, norfluoxetine was about 4-fold more potent at inhibiting Kv1.3 than fluoxetine, displaying an IC₅₀ value of 1.4 µM. Also, like fluoxetine, norfluoxetine accelerated the rate of Kv1.3 current decay.

View larger version (15K): [in this window] [in a new window]   Fig. 7.   Effects of norfluoxetine, the major metabolite of fluoxetine, on Kv1.3 currents. A, whole-cell Kv1.3 currents were elicited by 200-ms step depolarizations to +40 mV from a holding potential of 80 mV at 30-s intervals. Control current and current after the addition of 1 and 10 µM norfluoxetine are indicated. B, concentration-response curve for Kv1.3 inhibition by norfluoxetine. Reduction of current (relative to control) at the end of depolarizing steps from 80 to +40 mV was used as an index of inhibition. Nonlinear least-squares fit of the data yielded an IC50 value of 1.4 µM and a Hill coefficient of 1.1 (n = 5).

Finally, to check whether the fluoxetine-induced inhibition demonstrated in cell transfected with Kv1.3 channels could also be observed in native T lymphocytes, we investigated the effects of fluoxetine on human T lymphocytes (Fig. 8). Fluoxetine (10 µM) also decreased the outward K⁺ current in human T lymphocytes by 21.4 ± 4.9% (n = 3). Therefore, the same effects of fluoxetine on Kv1.3 were observed both in Kv1.3 transfected CHO cells and in human T lymphocytes.

View larger version (11K): [in this window] [in a new window]   Fig. 8.   Effects of fluoxetine on voltage-activated K+ currents in human T lymphocytes. Whole-cell currents were elicited by 200-ms step depolarizations to +40 mV from a holding potential of 80 mV at 30-s intervals. The traces of the currents obtained in the absence and presence of 10 µM fluoxetine are shown (n = 3).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In the present study, we demonstrated that fluoxetine is a potent blocker of the Kv1.3 channel stably expressed in CHO cells and native human T lymphocytes. Furthermore, norfluoxetine, a major metabolite of fluoxetine, also inhibited Kv1.3. Fluoxetine produced a concentration- and voltage-dependent inhibition of Kv1.3 channels. The main effect of fluoxetine was to accelerate Kv1.3 current decay during step depolarization, thus reducing current amplitude at the end of the test pulse. This effect was observed at concentrations as low as 1 µM, which is within the range of concentrations observed clinically.

Fluoxetine-induced acceleration of the rate of decline of Kv1.3 currents can be due to several mechanisms, including block of open channels. As shown in Figs. 1 and 6, there was little inhibition of Kv1.3 at the onset of depolarization. This finding indicated that fluoxetine does not preferentially bind to the resting state of the channel. However, on depolarization, fluoxetine caused an acceleration of the time course of decay of Kv1.3 currents in a concentration-dependent manner. A possible explanation is that fluoxetine preferentially binds to the open state of the channel or accelerates current inactivation. However, it is unlikely that this acceleration of current inactivation was due to alteration of gating kinetics because activation and steady-state inactivation properties of Kv1.3 are not affected by fluoxetine. In addition, the interaction of fluoxetine with the Kv1.3 channel was voltage-dependent. The inhibition increased steeply in the voltage range of channel activation, thus providing further strong evidence that the channels must open before fluoxetine can bind and block permeation. Consistent with the predominantly cationic form of this drug at physiological pH (pK_a ~10), fluoxetine displays a greater block at more depolarized potentials. A fractional electrical distance calculated from the Woodhull model  = 0.29 was obtained in this experiment with 10 µM fluoxetine. This  value was similar to the value  = 0.24 obtained in previous reports for fluoxetine in Kv1.1 channels (Tytgat et al., 1997). The  value indicated that the positively charged fluoxetine senses about 29% of the applied transmembrane electrical field as referenced from the intracellular side. This finding reflects blockade of the open channel from the inside, consistent with the effect of fluoxetine in inside-out patches (see below). Open channel block can also affect the time course of the tail current. In our experiment, fluoxetine decreased the peak tail current amplitude and slowed the time course of tail currents, producing a crossover phenomenon of the tail currents. This phenomenon has been also reported with the block of other open state K⁺ channel blockers (Snyders et al., 1992; Delpón et al., 1996; Tytgat et al., 1997). However, the slower deactivation kinetics, in the presence of fluoxetine, might also be interpreted as drug-bound inactivated channels taking longer to deactivate due to a structural requirement for returning to their closed resting state via a transition through the open state. Because many Shaker-type K⁺ channels exhibit C-type inactivation during a prolonged depolarizing pulse (Hoshi et al., 1991), we cannot exclude an interaction of fluoxetine with the inactivated state of Kv1.3. On the other hand, the reduction of the peak current occurring at higher concentrations could be attributed to an interaction of the drug with the closed resting state of the channel (tonic block). Although the instantaneous block on depolarization can occur if an open channel block developed before the time of the peak current, the present data are not sufficient to validate such a hypothesis and the possibility of inhibition of the channel in the resting state cannot be completely ruled out.

Because fluoxetine is highly lipid-soluble and is permeable to biological membranes, externally applied fluoxetine inhibited Kv1.3 currents by binding to an extracellular site or by diffusing through the cell membrane and gaining access to the binding site from the internal surface. However, our results strongly suggest that inhibition of Kv1.3 by fluoxetine is due to binding of fluoxetine to an intracellular site. From the analysis of the concentration dependence of inhibition, it was evident that the Kv1.3 channel was more sensitive to fluoxetine applied from the intracellular side of the membrane. The discrepancy in the IC₅₀ values obtained for inside-out and whole-cell recordings suggests that the Kv1.3 channel binding site for fluoxetine is on the intracellular side of the channels. The IC₅₀ value of fluoxetine for inhibiting Kv1.3 expressed in CHO cells is much less than the value reported for the blockade of Kv1.1 expressed in Xenopus oocytes (IC₅₀ = 300-700 µM). Moreover, the slow and incomplete reversal reported in previous reports (Tytgat et al., 1997; Hahn et al., 1999) probably is the result of fluoxetine accumulation within the cells. If this is indeed the case, the rate-limiting step for inhibition would be the rate of diffusion into the plasma membrane. A slower rate of diffusion of fluoxetine through the plasma membrane may be the reason for the lower potency and the long exposure time required to inhibit Kv1.3 in whole-cell recordings compared with inside-out recordings. Taken together, all these results suggest that fluoxetine accesses its binding site of Kv1.3 from the intracellular side of the channel.

Kv1.3 has been demonstrated to be the predominant voltage-activated K⁺ channel in human T lymphocytes, and its electrophysiological properties have been well characterized (Matteson and Deutsch, 1984; DeCoursey et al., 1985). The role of Kv1.3 in human T lymphocytes remains unclear, but the possible involvement of this K⁺ channel in T lymphocyte activation has been reported by several investigators. There is some evidence that drugs that block Kv1.3 channels also inhibit the activation and proliferation of T lymphocytes (Chandy et al., 1984; Lin et al., 1993). In humans, therapeutic doses of fluoxetine used in the management of depression result in plasma levels of approximately 1 µM. Thus, it is possible that concentrations required to inhibit Kv1.3 in our experiments can be achieved in various clinical conditions. In addition, fluoxetine is extensively metabolized to norfluoxetine. Norfluoxetine also is a serotonin reuptake inhibitor and its pharmacological action is similar to that of the parent drug (Lucas, 1992). The half-lives of both drugs are relatively long (4-15 days). In our experiment, both fluoxetine and norfluoxetine inhibited Kv1.3 with an IC₅₀ value of ~1 µM, which is comparable to therapeutic plasma concentrations. Although the clinical relevance of the Kv1.3 channel inhibition by fluoxetine and norfluoxetine is unclear at the present time, there is considerable evidence demonstrating the effects of fluoxetine on immune cell function. More recently, fluoxetine was found to decrease the mitogen-induced lymphocyte proliferation and natural killer cell cytolytic activity in rats (Pellegrino and Bayer, 1998). Presently, it is unclear whether the effects of fluoxetine on the immune system may be directly related to inhibition of voltage-activated K⁺ channels in human T lymphocytes. Because the potential of Kv1.3 as a target for an immunosuppressive drug has been recognized, the fluoxetine-induced inhibition of Kv1.3 channels could be of clinical relevance with IC₅₀ values similar to therapeutic plasma concentrations.

We conclude that fluoxetine and its metabolite, norfluoxetine, inhibit Kv1.3 channels, a major voltage-activated K⁺ channel in human T lymphocytes, at clinically relevant concentrations and that such effects may have pharmacological significance.