Study of membrane potential in T lymphocytes subpopulations using flow cytometry

Background Ion channels are involved in the control of membrane potential (ψ) in a variety of cells. The maintenance of ψ in human T lymphocytes is essential for T-cell activation and was suggested to depend mostly on the voltage-gated Kv1.3 channel. Blockage of Kv1.3 inhibits cytokine production and lymphocyte proliferation in vitro and suppresses immune response in vivo. T lymphocytes are a heterogeneous cell population and the expression of Kv1.3 varies among cell subsets. Oxonol diBA-C4-(3) was used to determine ψ by flow cytometry. The presence of distinct T cell subsets was evaluated by immunophenotyping techniques and the contribution of Kv1.3 channels for the maintenance of ψ was investigated using selective blockers. Results The distribution of ψ in T lymphocytes varied among blood donors and did not always follow a unimodal pattern. T lymphocytes were divided into CD3+/CD45RO- and CD3+/CD45RO+ subsets, whose peak channel values of ψ were -58 ± 3.6 mV and -37 ± 4.1 mV, respectively. MgTX (specific inhibitor of Kv1.3 channels) had no significant effect in the ψ of CD3+/CD45RO- subsets but depolarized CD3+/CD45RO+ cells to -27 ± 5.1 mV. Conclusion Combination of optical methods for determination of ψ by flow cytometry with immuophenotyping techniques opens new possibilities for the study of ion channels in the biology of heterogeneous cell populations such as T lymphocyte subsets.


Background
Electrical potential differences are generated across the cytoplasmic membranes of animal cells by concentration gradients of ions such as Na + , K + , Cland H + . The maintenance of membrane potential (ψ) depends on ion chan-nels, ion pumps and eletrogenic transporters. Ion channels also regulate various cell functions such as: electrical excitability of myocytes and neurons [1], cell proliferation [2][3][4] and hormone secretion [5,6]. The study of ψ variations require the use of electrophysiological methods [1,7], the patch-clamp being the gold-standard technique [7], because it allows detailed biophysical characterization of ion channels [8,9] and, combined with pharmacological tools, the study of their contribution to ψ [9,10]. However, patch-clamp analysis is restricted to one cell at a time, limiting its application for the study of large and heterogeneous cell populations. Optical methods for the determination of ψ were introduced by Cohen et al. [11] and are an alternative for the study of ψ variations in a large number of cells within a reasonably short period of time. These optical methods are based on the use of fluorescent dyes, which respond to membrane polarity stimuli causing changes in fluorescence [12]. Combination of optical methods for the measurement of ψ with flow cytometry (Fluorescence Activated Cell Sorter -FACS) techniques opens new possibilities for the study of ion channels in the biology of heterogeneous cell populations.
Human T lymphocytes are a good example of a heterogeneous cell population in which the study of ion channels and their contribution for ψ is of great interest. The activation of T lymphocytes during the immune response requires continuous Ca 2+ influx across the plasma membrane [13,14]. The voltage-gated K + channel, Kv1.3 [8,15] and the Ca 2+ -activated-K + channel, KCa3.1 modulate calcium influx by regulating the ψ and providing electrical driving force for continuous Ca 2+ entry [8,16]. While KCa3.1 blockers are able to prevent proliferation in mitogen-activated lymphocytes [16], blockage of Kv1.3 channels by specific inhibitors, such as margatoxin (MgTX) prevent proliferation in resting T cells. Blockage of Kv1.3 channels causes a depolarization of the ψ leading to a reduction in the intracellular Ca 2+ concentration [8,16]. As a consequence, cytokine production and cell proliferation are inhibited [15], which attenuates immune response in vivo [2]. Data in the literature regarding expression of Kv1.3 and control of ψ were obtained with path-clamp techniques on isolated T cells activated in vitro [17][18][19]. Peripheral T cells, however, are composed of non-activated (naive) T cells, pre-activated T blasts and memory T cells. Data obtained by optical methods estimate that the ψ of peripheral T cells vary between -70 and -45 mV [20][21][22], suggesting that different subsets of T cells present in peripheral blood have distinct ψ.
The membrane potential-sensitive fluorescent dye oxonol (diBA-C4-(3) was chosen due to advantages over other dyes: i) it is non-cytotoxic, ii) not shown to block ion channels and iii) it is not extruded by the glycoprotein efflux pump [23,24]. In the present work we combine oxonol with FACS-immunophenotyping techniques in order to characterize the ψ in specific sub-populations of human T lymphocytes [25]. We use specific inhibitors of potassium channels to evaluate the role of voltage-gated K + channels in controlling the ψ in naive and in memory T cells.

Validation of FACS estimates of ψ
The calculation of ψ was based on the Nernst equation: ψ = RT/F*ln(Ox i /Ox e ), where R is the universal gas constant, T is the absolute temperature, F is the Faraday's constant and Ox i and Ox e are the internal and external concentrations of oxonol, respectively. The calibration curve was determined using different concentrations of extracellular oxonol. Since the external and internal concentrations of the dye are equal when the ψ is the same (Ox i = Ox e *exp ψF/RT), one can assume a new calibration curve based on Ox i . The ratio Ox i /Ox e was calculated based on the acquisition of a fixed sample (ψ equal to 0 mV) using the same Ox e for both. Afterwards the ratio value of Ox i /Ox e was used to calculate the ψ based on the Nernst equation [26].
We characterized the variation of ψ in Kv1.3-transfected CHO (CHO-Kv1.3) cells exposed to different concentrations of extracellular K + ([K + ] e = 5-145 mM). Figure 1A and 1B shows the values of ψ measured by patch-clamp and the oxonol fluorescence by FACS, respectively. The values of ψ measured by either FACS or by patch-clamp were compared ( fig 1C). The curves show an overlay in the range of -40 to +10 mV, indicating the reliability of ψ measurements obtained with FACS.
Our aim was also to test if this technique could discriminate between distinct ψ generate by different concentration of pharmacological blockers. CHO-Kv1.3 cells were treated with two toxins: MgTx (black bars) or iberiotoxin (IbTx; grey bars) and the ψ was measured by FACS ( fig.  1D). The IbTx was chosen as a control blocker for the Kv1.3 channel, since it is the most potent and high-affinity blocker for the high-conductance calcium-activated potassium channel (BK Ca ) and it has none or low affinity for the Kv1 channels [27]. Addition of MgTx depolarized CHO-Kv1.3 cells on a dose-dependent manner shifting ψ from -50.3 ± 2.5 to -3.6 ± 2.5 when the concentration of 10 nM was used. IbTx, which is a selective blocker of the BK Ca channel [28], had no effect on the ψ of CHO-Kv1.3 cells.
In order to evaluate the ability of this method to distinguish cell populations with different ψ, we used two established cell lines, CHO and CHO-Kv1.3 [29]. These results indicate that it is possible to characterize the ψ of different cell populations using FACS and to evaluate the contribution of ion channels for maintenance of ψ by using specific ion channel blockers.

Distribution of ψ on peripheral blood lymphocytes
Human mononuclear cells from peripheral blood (PBMC) were immunostained with CD3 and CD45RO mAb and loaded with oxonol in order to evaluate the ψ in peripheral blood lymphocytes (PBL). PBL were gated according to their physical characteristics and the patterns of CD3 and CD45RO were analyzed in Figure 3A. Two subsets of T lymphocytes (CD3 + cells) can be identified in relation to the expression of CD45RO. Thus, CD3 + / Validation of ψ quantification by FACS  Figure  3B shows the CD45RO + subset according to the oxonol fluorescence distribution. The distribution of ψ in CD3 + cells did not always follow a unimodal pattern, as the example illustrated in figure 3B. In two out of six blood donors, we obtained a bimodal distribution of ψ. Separation of CD3 + cells according to the expression of CD45RO allowed characterization of ψ in the two cell subsets. Thus, CD3 + /CD45ROcells were hyperpolarized in relation to CD3 + /CD45RO + cells ( fig. 3C), the peak channel values of ψ being -58 ± 3.6 mV and -37 ± 4.1 mV, respectively (P = 0.0087). Figure 4 shows the histograms of ψ distribution obtained with CD3 + cells from three different donors (A, B and C) in control (gray line) and upon treatment with either 10 nM MgTX (black line) or high [K + ] e (145 mM; dotted line). The three control CD3 + samples represent the distinct distribution patterns of ψ. Thus, donors A and C showed a unimodal pattern on the control trace, whereas donor B showed a bimodal pattern. Addition of high [K + ] e depolarized all the samples to approximately 0 mV (peak channel evaluation) and generated a unimodal pattern in all histograms studied. In contrast, addition of MgTX produced diverse degrees of response, causing partial depolarization, which broadened the ψ distribution ( Fig. 4A) or generated a bimodal pattern (Fig. 4C). Percentage of cells which depolarized upon treatment with MgTx and overlapped with the histogram acquire after treatment with high [K + ] e were determined. Addition of MgTx did not depolarize all population when compared with the high [K + ] e treatment; it rather depolarized a small part (black marker) which accounts for 27, 39 and 12% of CD3 + cells from the different blood donors ( fig. 4A,B and 4C, respectively).

Effects of MgTX or high [K + ] e on the ψ of PBL
In order to characterize the distinct degrees of depolarization upon exposure to MgTX in activated T lymphocytes, we investigated the presence of CD45ROand CD45RO + cells in the different samples and evaluated the sensitivity of each cell subset to MgTX. The proportion of CD45ROand CD45RO + varied among donors and the individuals with higher amounts of CD45RO + cells among T lymphocytes (N = 4) showed significant depolarization upon treatment with MgTX (from -45 to -15 mV, P = 0.0025), whereas those with low proportion of CD45RO + cells (N = 2) were not affected by MgTX exposure (data not shown). In view of these results, we analyzed the distribution of ψ and the effects of MgTX in CD45ROand CD45RO + subsets. Figure 5 shows the results obtained in three out of six donors (panels A, B and C), which illustrate the different patterns and degrees of response. The CD45ROand CD45RO + subsets from each donor are shown in left and right panels, respectively. The CD45RO + subsets presented a unimodal distribution of ψ in all donors studied. In contrast, CD45ROsubsets had a more variable distribution of ψ, with a bimodal pattern being seen in two out of six donors (panel B-left shows an example). Addition of high [K + ] e (dotted line) depolarized all cell subsets. In contrast, addition of MgTX had no significant effect in CD45ROsubsets (P = 0.15, N = 6) but depolarized CD45RO + cells, shifting the peak channel value of ψ from -37.2 ± 4.1 mV to -26.7 ± 5.1 mV (P = 0.0025, N = 6). By comparing CD45ROand CD45RO + subsets that depolarized at the same extent as high [K + ] e treatment (marker), we had always 2-fold increase of percentage of cells in the later subset for each donor. Percentage of cells was 5 and 13%, 29 and 64% and 17 and 37% (CD45ROand CD45RO + , respectively; fig. 5A,B and 5C).

Discussion
The main goal of the present study was to combine the methodology described by Krasznai et al. [26] with other FACS techniques and the use of specific ion channel blockers in order to study the ψ of T lymphocytes. CHO and CHO-Kv1.3 cells are well-established cell lines, widely used in electrophysiology [29][30][31][32]. The mean values of ψ determined by FACS in these cell lines are in agreement with the data from electrophysiological studies [29]. The dispersion of ψ values was higher for CHO cells (-70 to +30 mV) than for CHO-Kv1.3 cells (-70 to -30 mV). CHO cells have their ψ controlled partly by chloride channels and partly by cation channels [29]. The transfection of Kv1.3 to CHO cells sets the resting ψ to values close to -50 mV, similarly to what is observed in human peripheral T lymphocytes [8,33]. The narrower dispersion of ψ values in CHO-Kv1.3 cells as compared to CHO cells corroborates the idea that Kv1.3 is the main responsible for the control of ψ in these cells. This is confirmed by the fact that MgTX, but not IbTX, depolarizes CHO-Kv1.3 cells and enlarges the dispersion of ψ to values similar to those of CHO cells (-60 to +30 mV, fig. 1C).
PBMC are a heterogeneous population, composed of T and B lymphocytes, NK cells and monocytes. Expression of Kv1.3 has been reported in T and B lymphocytes and in monocytes/macrophages [34]. In the present study, we evaluate the ability of Kv1.3 channel to control the ψ in different T lymphocyte subsets, allying an optical method for determination of ψ by FACS with immunophenotyping techniques. Human T lymphocytes were identified by the expression of CD3 and subdivided into CD3 + / CD45ROand CD3 + /CD45RO + cells. Thus, CD3 + / CD45ROcells include naive and recently activated T lymphocytes, whereas CD3 + /CD45RO + correspond to memory T lymphocytes [25]. The proportion of CD3 + / CD45ROand CD3 + /CD45RO + varied among different donors (see figure 4) reflecting the dynamic regulation of the immune system.
When analyzed together, CD3 + cells showed different patterns of ψ distribution and variable sensitivity to MgTX, suggesting that the T lymphocyte subsets have different ψ and are differently regulated by Kv1.3. Accordingly, it has been recently shown that naive and memory T cells have differences in the expression of Kv1.3 and KCa3.1 channels. Thus, naive cells express about 200-400 Kv1.3 channels along with 8-10 KCa3.1 channels per cell, whereas memory T cells may have up to 1800 Kv1.3 channels/cell [35]. The markers for discriminating naive and memory T cells used in this study are different from the ones published by Wulff et al. [35], nevertheless there is an overlap between the subsets studied. In view of this channel distribution, it would be expected that CD3 + /CD45RO + cells, were hyperpolarized in relation to CD3 + /CD45ROcells, unlike the results shown in figure 3. However, CD3 + / CD45ROcells include naive and recently activated T blasts [25], and the latter express 500-600 KCa3.1 channels, which have been shown to shift the ψ to -80 mV [36]. Thus, the broad distribution of ψ (sometimes with a bimodal pattern) within CD3 + /CD45ROcells may be due to the presence of activated T blasts in the peripheral blood of some donors. Accordingly, CD3 + /CD45ROcells were not significantly depolarized by MgTX, suggesting a minor role of Kv1.3 channels among these cells.
Memory T cells comprehend two sub-populations, which have been classified as central memory (TCM) and effector memory (TEM) cells, based on their homing potentials and effector functions [37]. These two memory cell subsets differ in relation to the expression of Kv1.3 and KCa3.1 channels. TCM cells have 250-300 Kv1.3 channels/cell and up-regulate KCa3.1 from 20 to 500-600 channels/cell following activation, whereas TEM cells upregulate Kv1.3 channels to 1500-1800 channels/cell and down-regulate KCa3.1 to 50-100 channels/cell after repeated activation [35]. In the present study, we did not distinguish these two memory T cell subsets and we are aware that further experiments are necessary to study these T cells subsets and examine the effect of Kv1.3 and KCa3.1 channels blockers. Nonetheless, our results corroborate the notion that Kv1.3 channels are the main responsible for the control of ψ among memory T cells, since MgTX caused significant depolarization. It is noteworthy, though, that the depolarization was partial in some cases (see figure 5A), suggesting the presence of a less sensitive subset (possibly composed of TCM cells). The correlation between a specific channel and its ability to maintain the resting potential of a particular population of cells requires the use of more specific blocker since many of the available pharmacological channel blockers target more than one channel. Nevertheless, we can suggest that the Kv1.3 channel is involved in the regulation of ψ from the CD3 + /CD45RO + subset of T lymphocytes ( fig.   5).

Effects of MgTX or high [K + ] e on subsets of CD3 + cells
The fact that Kv1.3 channel is functionally restricted regarding tissue distribution together with the improvement of experimental autoimmune encephalomyelitis [38] and delayed type hypersensitivity in animal models without causing obvious side effects has made Kv1.3 an interesting therapeutic target [2,39,40]. A rapid screening of new ion channel blockers and the determination of the exact subset of cells affected by these blockers would be of great interest in the development of new immunossupressive therapies.

Conclusion
In summary, our results indicate that FACS determination of ψ can be used for identification of ψ heterogeneity among cell populations. Combination of this method with other FACS techniques could also be used for determination of ψ in different cell cycle phases, developmental stages or activation patterns and for rapid screening of new ion channel blockers. This represents a new strategy for studying the role of ion channels in cell growth and differentiation of normal and tumoral cells. PBMC were isolated from the peripheral blood of healthy donors by centrifugation on a Ficoll gradient. PBMC were washed and incubated with RPMI 1640 medium (Sigma) supplemented with 10% (v/v) heat-inactivated bovine fetal serum (Gibco), 60 mg/L penicillin (Sigma), 100 mg/ L streptomycin (Sigma) for 30 minutes (37°C, 5% CO 2 humidified atmosphere). The study was evaluated by the National Cancer Institute (INCa-Brazil/RJ) Ethical Committee and the informed consent of all participating subjects was obtained.

Cells
Samples were fixed with ice-cold 2% formaldehyde and kept at 40°C for 60 minutes. Cells were washed with PBS and kept at room temperature before the measurements.

Flow Cytometry (FACS)
We used the method described by Krasznai et al. [26] for determination of ψ. The phosphate buffer solution (PBS) was replaced by the PSS, which is the standard solution in all experiments of electrophysiology in our laboratory. The method was validated using the PSS in CHO cells and in human lymphocytes.
PBMC were labeled with a PercP-conjugated anti-human CD3 or PE-conjugated CD45RO mouse antibody (Pharmigen, San Diego, CA, USA), Fc receptor being blocked with normal mouse serum (1:50) in PBS buffer. After a 20 minute-incubation with the antibody at 4°C, cells were washed with PSS kept at room temperature and re-suspended at a concentration of 10 6 ml-1.
Measurements were carried out at room temperature, using a Becton Dickison FACScan flow cytometer and data were analyzed using Cell Quest or FlowJo program. Forward-scatter (FSC) and side-scatter (SSC) lights were used for gating of data acquisition. Non-viable cells were identified with propidium iodide (Sigma), and were excluded from analysis. All samples were excited with the 488 nm line and oxonol, PE and PercP fluorescence emission were captured at 530/30 nm, 585/42 nm and 670 nm long pass, respectively. The calculated values of ψ within a cell subset are presented in histogram distributions and the peak channels were used for comparative analysis. All the experiments were performed at least four times and the data are presented as mean ± standard error.

Statistical Analysis
Unpaired t test was performed for comparison of the values of ψ between cell subsets and paired t test was used for comparisons of ψ after different treatments within a given cell subset. The software GraphPad Prism version 4 was applied for the analysis. The ψ was obtained from the conversion of the peak channel value generate by the Cell Quest software for each volunteer.