All cell lines used in this study were derived from the peripheral blood of patients diagnosed with T-ALL: Jurkat cell line was established from 14-year-old boy in first relapse in 1976; MOLT-3 cell line from 19-year-old man in relapse in 1971; CCRF-CEM cell line from 3-year-old Caucasian girl in relapse (terminal) in 1964. All cell lines were purchased from ATCC, and cryopreserved in liquid nitrogen. To prevent passage-related variations in data, the cells with passage number less than 20 were used in experiments. Cells were cultured in a Roswell Park Memorial Institute 1640 (RPMI 1640) Advanced medium supplemented with 5% heat inactivated fetal bovine serum, 100 U/ml penicillin, 100 mg/ml streptomycin, 10 mM N-2-hydroxyethylpiperazine-N′-2 ethanesulfonic acid (HEPES), 2 mM glutamine (GlutaMAX^TM Supplement) (all from Invitrogen) at 37°C in humidified atmosphere (5% de CO₂ and 95% air). Cells were maintained in the logarithmic growth phase by daily medium refreshment.

Mature CD4⁺ lymphocytes were obtained from peripheral blood of healthy 22–32 years old volunteers. Peripheral blood 10 ml- probes were taken by trained personal under aseptic condition with sterile material for each donor. The protocol was approved by Bioethics and Biosecurity Committee of the Biomedical Research Centre and Faculty of Medicine, in accord with Federal (Artículo 100, Ley General de Salud), State, and local laws. Informed consent was obtained in written form from all participants. Blood samples were diluted 1:1 with PBS. The mononuclear cells (CMN) were separated by Ficoll gradient (Ficoll-Paque PLUS, GE Healthcare). CMN were subjected to negative selection (to avoid activation) with CD4⁺ T cell isolation kit, following Miltenyi Biotec specifications. The CD4⁺ T cells were cultured in complete RPMI 1640 medium, in 24-wells plates (no more than 1.5 × 10⁶ cells per ml), for 24 h. For polyclonal activation, resting lymphocytes were pretreated in 96-well plates with antiCD3 monoclonal antibodies (5 μg/ml) (BD, 555336) for 2 h at 37°C; excess medium was removed and 1 × 10⁵ cells were incubated with (2 μg/ml) anti CD28 monoclonal antibodies (BD, 555725) for 4 days. Ligation of CD3/CD28 provides a powerful antigen-independent stimulus by cross-linking T cell receptor (TCR) resulting in transit from quiescent (resting) to proliferation state. This event, termed as blast formation is accompanied by a huge increase in biomass and size (Teague et al., 1993; Grumont et al., 2004; Li and Kurlander, 2010).

Patch-clamp recordings were performed in a whole-cell configuration, immediately after gaining the access to whole cell and after at least 15 min of incubation, to evaluate initial potassium current density and voltage dependence and to measure the voltage dependence at a stable state, respectively. Patch pipettes were pulled from Kwik-Fil 1B150F-4 capillaries (World Precision Instruments, Sarasota, FL, United States) in six steps on a Brown/Flamming model P-97 puller (Sutter Instruments, Novato, CA, United States) and fire polished, using LPZ 101 microforge (List Medical, Germany). Bath solution for recording of voltage dependent potassium (Kv) channels contained (in mM): 150 NaCl, 5 KCl, 1 MgCl₂, 2.5 CaCl₂, and 10 HEPES NaOH (pH 7.4). For records of Ca²⁺-activated potassium (KCa) channels bath contained (in mM): 160 KCl, 1 MgCl₂, 2.5 CaCl₂, and 10 HEPES-KOH (pH 7.4) or, for K⁺- free condition all KCl were replaced for with equimolar NaCl. Patch electrodes for Kv records were filled with a solution containing (in mM): 134 KCl, 2 MgCl₂, 1 CaCl₂, 10 EGTA, and 10 HEPES–KOH (pH 7.4). For KCa records the pipette solution contained (in mM): 130 K-Aspartate, 4.75 CaCl₂, 2 MgCl₂, 5 EGTA, and 10 HEPES-KOH (pH 7.4, pCa 6). Liquid junction potentials (LJP) between pipette and bath solutions were calculated using the application within Clampex Version 10.3.2.1. program (Molecular Devices, LLC). Assuming that in the whole cell mode this LJP will be zeroed, instead of cancelation of the electrical potential difference between open patch-clamp electrode and the bath at the beginning of each experiment, this voltage difference was set to a respective calculated LJP value. All chemicals were of analytical quality and purchased from Sigma-Aldrich. The resistance of patch electrodes was 3–5 MΩ. Experiments were performed at 28–30°C. Current measurements were performed using an Axopatch 200A Integrating Patch-clamp amplifier (Axon Instruments, Foster City, CA, United States). Records were low-pass filtered at 2 kHz, digitized using a DigiData 1200 Interface (Axon Instruments), transferred to a personal computer, and analyzed using the pClamp 6.0 software package (Axon Instruments).

For pharmacological analysis, the following drugs were introduced by bath perfusion to a final concentration as indicated: specific Kv 1.3 blocker MgTx (1 nM), specific KCa2.2 channel blocker apamin (1 nM), specific KCa3.1 blocker clotrimazole (200 nM), and efficient blocker of Kv1.3 and KCa 3.1 channels ChTx (50 nM). All drugs were from Tocris Bioscience.

Data are presented as means ± SE, with a number of individually tested cells, n, as indicated for each condition. For the comparison of the density of Kv and KCa currents statistical significance was calculated using Prism software version 6.0 (GraphPad, La Jolla, CA, United States) from adequately powered sample sizes for one-tailed tests using one-way ANOVA, unpaired Tuckey t-test with Bonferroni correction for multiple comparisons. Minimal criterion for statistical significance was defined as P ≤ 0.05. Actual P-values are provided in the figure legends.

Activation curves for the Kv1.3 current were fitted to a single Boltzmann function (Eq. 1):

where Z is a number of elementary charges to be translocated across the whole voltage drop upon the channel activation (gating charge), V_1/2 is a potential, at which half of the channels turned open (midpoint potential), V is membrane voltage, F is Faraday constant, R is universal gas constant, and T is absolute temperature. In case of CEM cells, the data points were better described by a sum of two different Boltzmann functions (Eq. 2):

where 𝜃 and (1-𝜃) are the weights of respective fractions. Inactivation curves can be well described by a single Boltzmann function with a small, but significant offset (off) (Eq. 3):

When two-pulse protocol was used to evaluate the steady-state inactivation (Figure 2), I₊₅₀/I₊₅₀(max) at the left side of the equation was substituted for I_2nd/I_1st, the amplitude ratio of Kv1.3 currents, evoked by second and first voltage steps to +20 mV.

Kv1.3 is the only member of the family of voltage-dependent Kv channels known to be expressed in healthy human T cells and Jurkat cell line (Cahalan and Chandy, 2009). Due to its characteristic pattern of time-dependent activation upon depolarization, use-dependent inactivation, and high affinity specific block by MgTx, Kv1.3 current can be with ease functionally identified in electrophysiological records. Here, we verified the identity of the Kv current in CEM and MOLT-3 leukemic T cell lines and performed an accurate comparison of its density and parameters of voltage activation and inactivation with Kv1.3 current in resting and activated helper T CD4⁺ and Jurkat cells.

To minimize the contribution of KCa current, free cytosolic (patch pipette) Ca²⁺ was set to virtually zero (few nM). To evaluate the Kv voltage- and time-dependent activation, from a holding potential of -70 mV, shortly passing to -100 mV to verify the quality of the seal, a sequence of depolarized pulses from -50 to +25 mV was applied. The duration of test pulses was long enough to reach the maximal activation, but as short as possible to avoid a substantial inactivation. At the end of depolarizing pulse the voltage was switched to -50 mV, and, subsequently returned to a resting potential of -70 mV (Figure 1, left), where it was kept for 15 s to preclude a significant accumulation of inactivation from pulse to pulse. Relative magnitude of tail current at -50 mV was taken as a measure of the channel activation at the end of the test pulse (Figure 3, left). This voltage protocol corresponded to the beginning of experiment, immediately after gaining into whole cell configuration. As also observed by other authors (Cahalan et al., 1985; Pahapill and Schlichter, 1990), Kv1.3 voltage dependence was shifted to more negative potentials upon few minutes of incubation in the whole cell mode. Thus, the measurement was repeated after at least 15 min of incubation, but holding potential was lowered to -80 mV, depolarizing pulses were started from -60 mV and tail currents were evaluated at -60 mV for this case. Alternatively to tail current evaluation, we have also performed a more traditional analysis, measuring the voltage dependence of the peak current conductance, G/G_max. Both approaches generally give consistent results; yet, a small but significant difference between the two datasets may be detected in case of CEM and Jurkat cells (Supplementary Figure S2). It should be noted that the rectification of the whole cell current may be caused not only by a voltage-dependent increase of channels’ open probability, but also by any voltage-dependent change of the single channel conductance. Therefore, the accuracy of the analysis based on the macroscopic conductance measurements depends on the correctness of the assumption of the constant single channel conductance (Ohmic behavior). Intrinsic inward rectification of the unitary Kv1.3 current appears to be “linearized” by Goldmann-type rectification due to an outward directed K⁺ gradient upon physiological conditions, resulting in a quasi-Ohmic behavior (Cahalan et al., 1985; Pahapill and Schlichter, 1990). However, some super-linear trends could be detected in unitary I/V curves, reported by Pahapill and Schlichter (1992). Thus, we preferred the tail current analysis, because the relative activity as a function of voltage may be directly read from the original recordings and the result is not dependent on the assumption of Ohmic behavior of the unitary current. At the end of each experiment, MgTx (1 nM) was added to the bath and the activation voltage protocol was run again. For all cell lines, as well as resting and activated healthy CD4+ lymphocytes, time- and voltage-dependent Kv currents were completely abolished by MgTx (Figure 1), which implies that the Kv current was mediated exclusively by Kv1.3 channels. It should be noted here, that MgTx at 1 nM is also a potent blocker of the Kv1.2 channels (Bartok et al., 2014). However, Kv1.2 current has biophysical properties clearly different the Kv1.3 one. In particular, it inactivates very slowly, with a characteristic time about 15 s (Sprunger et al., 1996), which is inconsistent with a relatively fast inactivation of the Kv current, reported in the present study (Figure 1).

The voltage-dependence of Kv1.3 current inactivation was obtained by application of rectangular 5 s depolarizing voltage pulses, starting from -60 mV for initial whole cell recording (Figure 1, right) or -80 mV after long (>15 min) incubation in the whole cell mode, in 5 mV increments. There were 45 s pauses between consecutive test pulses, to ensure the channels’ recuperation from the inactivation state. To estimate the fraction of a non-inactivated current, short pulse to +50 mV was applied after each test pulse, and respective amplitude was taken relative to a maximum and plotted against the test voltage value (Figure 3, right). This relatively short protocol primarily reflects the inactivation, occurring directly from the open channel state. It should be noted, that Kv1.3 current is believed to inactivate also from a closed state; this process starts at potentials, where the probability of open state is still very low and may proceed with characteristic times as slow as 100 s (Cahalan et al., 1985). Therefore, to evaluate the Kv1.3 channel inactivation at a steady-state, we incubated whole cell samples at -90 mV for >1 min and then apply a pair of depolarizing pulses to +20 mV, separated over 3 min in time by a test holding voltage, ranging between -90 and -20 mV (in 10 mV steps). The degree of equilibrium inactivation as a function of holding voltage was calculated by dividing the amplitude of the peak current evoked by the second pulse through that induced by the first pulse (see a typical experiment in Figure 2). For self-explicable reasons such a long protocol may not be used for evaluation of the inactivation process at the beginning of the whole cell recording. Yet, after reaching a stable state of whole cell recording at incubation times >15 min, the results, obtained by the two protocols, short (5 s) and long (3 min) may be compared. As to be expected, a steady-state inactivation curve was shifted to more negative potentials as compared to the curve, obtained by the usage of a shorter voltage protocol. Yet, this shift was relatively small, ranging from 1 to 4 mV, depending on the cell line (Figure 3, right, Table 1). A small shift may be explained by the fact that the inactivation may occur not directly from the closed state, but via a transient opening. The direct inactivation from open state is a relatively fast process with a characteristic time in the range of 10^-1 s (Figure 1). Multiplying this time by the open/closed equilibrium constant plus 1, one yields the estimate for a characteristic time for the inactivation from a closed state via open one. Thus, when the probability of the Kv1.3 channel to open will approach ∼10% (at voltages above -50 mV, Figure 3), the inactivation from a closed state via equilibration with the open state will occur only 10-times slower than a direct inactivation from the open state, i.e., it will be essentially completed within 5 s of the test pulse of our short protocol. After correction for a relatively small error in the determination of the midpoint potential, the short protocol as one in Figure 1, besides giving direct information on the kinetics of inactivation via the open state, could serve also as a quick estimate of the equilibrium inactivation at any moment.

Despite the fact that the Kv current for all leukemic cell lines and healthy T cells was mediated by Kv1.3 channels, its voltage dependence shows some differences (Figure 3). Activation curves in case of healthy T lymphocytes, and leukemic Jurkat and MOLT-3 cells could be well described by a single Boltzmann function (Eq. 1 in section “Materials and Methods”). However, in CEM cells single Boltzmann curve provided a poor description of the data. A much better description can be provided by a sum of two different Boltzmann functions (Eq. 2). Single Boltzmann curve describes experimental points equally good with a probability P < 0.0003, as yielded by F-test. One of the components in the CEM voltage dependence was characterized by very similar parameters values to those obtained for Jurkat, MOLT-3 and resting T lymphocytes; only a midpoint potential for activated T cells has shown a significantly different value in early whole cell recordings (Table 1). Activation curves shifted to more negative potentials upon incubation in whole cell mode. At a stable state reached at later times they display little if any difference for healthy lymphocytes and cell lines, but CEM, which preserved a biphasic behavior at all times. The second fraction (about 1/3 of all active Kv1.3 channels) in CEM was characterized by ∼17 mV more positive midpoint potential value and two-times weaker voltage dependence (Table 1).

Parameter values for inactivation curves are collected in the Table 1 and show striking similarity for all leukemic T cell lines used in this study and healthy T cells at a final stable state. However, for early whole cell recordings with resting T cells the midpoint potential value for inactivation appeared to be more negative by ∼10 mV as compared to activated T cells and cell lines. This, however, may be an artifact, dealt with a much smaller size of resting T cells, membrane capacitance of ∼1.7 pF as compared to ∼5 pF for activated T cells and leukemic lymphoblasts. Thus, with resting T cells the washout process would be more rapid so that even early measurements in the whole cell mode might reflect already an altered state. Consequently, with resting T cells substantially smaller relative shifts between initial and final curves were observed both for activation and inactivation, as compared to those with larger cells (Figure 3).

KCa current in human T lymphocytes and leukemic T cell line Jurkat is activated by cytosolic Ca²⁺ in a highly cooperative manner, with a full activation at free Ca²⁺ = 1 μM. However, in healthy T cells, either resting or activated, KCa current is sensitive to ChTx, whereas in Jurkat cells respective current is mainly composed by ChTx insensitive KCa channels with a high sensitivity to apamin (Grissmer et al., 1992; Grissmer et al., 1993). Currently it is known that the channels, encoding KCa currents in healthy T and Jurkat cells, are intermediate conductance IKCa (KCa3.1, KCNN4) and small conductance SKCa (KCa2.2, KCNN2) ones (Cahalan and Chandy, 2009). The only report on non-Jurkat leukemic cell line (CEM) suggests that a substantial part of the KCa current in this model is ChTx-sensitive (Zegarra-Moran et al., 1999). Thus, the composition of the KCa current in T leukemic cells is an open issue.

To evaluate KCa current, we have elevated free Ca²⁺ in pipette (intracellular solution) to 1 μM. Yet, Kv1.3 current, unless blocked by some external agent, will be also present under these conditions. To separate Kv1.3 from KCa currents, we applied ramp-wave protocols, starting from very negative potentials (–150 mV), where Kv1.3 channels are closed. To increase the inward current, which at potentials <-40 mV is dominated by KCa, we have increased the external K⁺ concentration. Finally, it was important to keep at minimum the contribution of a non-specific leak current. That it was the case in our experiments is evidenced by following data. First, substitution of K⁺ for Na⁺ in external solution virtually abolished the inward current (Figure 4, K-free bath). On contrary, leak conductance does not differentiate between Na⁺ and K⁺, so that inward leak current at -150 mV should be the same in Na⁺ and K⁺ bath. This result also implies that the inward current at -150 mV with high K⁺ bath is mediated by K⁺-selective channels. This K⁺ current is Ca²⁺-activated, because it is abolished at low (8 nM) intracellular Ca²⁺. Second line of evidence for the negligible contribution of leak is the high sensitivity of inward current within -50 to -150 mV range to specific blockers of KCa2.2 or KCa3.1 channels, either to 1 nM apamin (Jurkat) or to 200 nM clotrimazole (other cell lines and healthy T cells), respectively (Figure 4). In addition, ChTx, which blocks Kv1.3 and KCa3.1 channels with identical affinity, K_d ∼ 3.5 nM (Grissmer et al., 1993), at 50 nM concentration produced almost a complete block of the whole cell current, except the case of Jurkat, where voltage-independent component resided. At the same time, the effect of 1 nM apamin on the whole cell current in all cases, but Jurkat, was negligible. The data, presented in Figure 4 strongly evidence that in cases of human T lymphocytes, CEM and MOLT-3 cell lines KCa current is mediated exclusively by ChTx- and clotrimazole- sensitive KCa3.1 channels. On the contrary, in Jurkat cells we could not detect KCa3.1 channels in the whole cell mode; within a margin of error, the whole cell KCa current in this model was mediated by apamin-sensitive KCa2.2 channels. Although there were some indications in early literature on the presence of low-conductance KCa channels in rat thymus and human T and B cells, there was no evidence on their apamin sensitivity, and, also, in available cases (human T cells) all KCa conductances were sensitive to ChTx (Leonard et al., 1992; Jäger et al., 2000).

Specific current densities were evaluated either as early as possible after breaking into whole cell mode (case of Kv1.3), or, in case of KCa few minutes were taken to allowed the perfusion of Ca²⁺ buffer from the pipette to interior of the cell. Joint recording of KCa and Kv currents in whole cell mode reveals large differences in the current density between Jurkat, resting and activated healthy T cells and other T cell lines (Supplementary Figure S3). In these records at large negative potentials the whole cell current is dominated by KCa, whereas around -40 mV the Kv activation starts (increase of the inward current); when it reaches its maximum (the minimum of the current), whole cell conductance is defined by the sum of Kv and KCa currents. KCa current density was the highest in activated T cells, followed by MOLT-3 and CEM, it was lower in resting T cells, and the lowest in Jurkat cells. Healthy T cells and all cell lines but Jurkat displayed a relatively large Kv current.

Quantitative analysis shows that KCa current density increased more than 10-times upon T cells activation (Figure 5B). When expressed on the number of channels copies/cell basis, the difference between resting and activated T cells was even larger, 10 and 330 channels/cell, respectively, due to increased cell size upon activation, from 1.7 to 5.1 pF as an average. Density of KCa3.1 current in leukemic T cell lines MOLT-3 and CEM was substantially higher than in resting T cells from healthy individuals and about 62% as an average of that in activated T cells (Figure 5B). Irrespective to molecular identity, KCa2.2 in Jurkat vs. KCa3.1 in CEM and MOLT-3, KCa current density in Jurkat was 10 times lower.

When Kv current was assayed separately, at physiological ionic gradients and zero cytosolic Ca²⁺ (as in Figure 1), no significant difference in Kv1.3 current density was found between healthy T cells (resting or activated), MOLT-3 and CEM leukemic cell lines; in all cases specific current at 0 mV was between 170 and 180 pA/pF. In Jurkat Kv1.3 current density was 10-times lower (Figure 5A). Low functional expression of Kv in Jurkat was also reflected by a more depolarized membrane potential, measured in the current-clamp mode immediately after gaining low resistance access to the cell interior, -20 to -30 mV as compared to -40 to -50 mV in MOLT-3 and CEM cell lines (result not shown).