Depressive Effectiveness of Vigabatrin (y-Vinyl-GABA), an Antiepileptic Drug, in Intermediate-conductance Calcium-Activated Potassium Channels in Human Glioma Cells

Background: Vigabatrin (VGB) is an approved non-traditional antiepileptic drug that has been revealed to have potential for treating brain tumors; however, its effect on ionic channels in glioma cells remains largely unclear. Methods: With the aid of patch-clamp technology, we investigated the effects of VGB on various ionic currents in the glioblastoma multiforme cell line 13-06-MG. Results: In cell-attached conguration, VGB concentration-dependently reduced the activity of intermediate-conductance Ca 2+ -activated K + (IK Ca ) channels, while DCEBIO (5,6-dichloro-1-ethyl-1,3-dihydro-2H-benzimidazol-2-one) counteracted the VGB-induced inhibition of IK Ca channels. However, the activity of neither large-conductance Ca 2+ -activated (BK Ca ) nor inwardly rectifying K + (K IR ) channels were affected by the presence of VGB in human 13-06-MG cells. However, in the continued presence of VGB, the addition of GAL-021 or BaCl 2 effectively suppressed BK Ca and K IR channels. Conclusions: The inhibitory effect of VGB on IK Ca channels demonstrated in the current study could be an important underlying mechanism of VGB-induced antineoplastic (e.g., anti-glioma) actions.

VGB has been reported to decrease oligodendrocyte precursor cell proliferation as well as to increase the number of mature oligodendrocytes [14]. Interestingly, it has been also disclosed to have promising therapeutic e cacy for treating brain metastases in vivo [15]. However, the ionic mechanism through which VGB exerts anti-neoplastic actions is not yet determined. In this study, we sought to investigate its ionic mechanism which could be linked to anti-neoplastic actions in the glioblastoma multiforme cell line (i.e., human 13-06-MG glioma cells).
For recording large-conductance Ca 2+ -activated (BK Ca ) channels, we kept cells in a high K + -bathing solution, and its composition was 145 mM KCl, 0.53 mM MgCl 2 , and 5 mM HEPES adjusted with KOH to 7.2, and the pipette solution contained 145 mM KCl, 2 mM MgCl 2 , and 5 mM HEPES titrated with KOH to 7.2. In this study, we obtained the reagent water from a Milli-Q water puri cation system (Merck, Ltd., Taipei, Taiwan). The culture medium and pipette solution were ltered on the day of use with an Acrodisc Ò syringe lter with a Supor Ò membrane (Bio-Check; New Taipei City, Taiwan).

Cell preparations
The glioblastoma multiforme cell line (13-06-MG) used in this study was kindly provided by Professor Dr. Carol A. Kruse (Department of Neurosurgery, Ronald Reagan UCLA Medical Center, LA, U.S.A). The 13-06-MG cells were cultured at a density of 10 6 /ml in high glucose (4 g/l) Dulbecco's modi ed Eagle media (Invitrogen, Carlsbad, CA, USA) supplemented with 10% heat-inactivated fetal bovine serum, 100 U/ml penicillin and 10 μg/ml streptomycin. Cells were maintained at 37˚C in a 5% CO 2 incubator as monolayer cultures and thereafter sub-cultured weekly. Fresh media was added every 2-3 days in order to ensure a healthy cell population. To verify the presence of glial cells, we identi ed them by displaying glial brillary acidic protein, which is a cytoskeletal protein.
To evaluate concentration-dependent inhibition of VGB on the probability of IK Ca channels that would be open, we kept 13-06-MG cells to be bathed in normal Tyrode's solution containing 1.8 mM CaCl 2 , and each cell examined was voltage-clamped at -80 mV relative to the bath. The probability of channel opening was measured in the control or during cell exposure to different concentrations (0.3-100 μM) of VGB; and, these values were then compared with those taken after the addition of TRAM-34 (3 μM).
TRAM-34 is a known selective blocker of IK Ca channels. The concentration required to suppress 50% of channel activity was determined by means of a Hill function: where IC 50 or n H is the concentration required for a 50% inhibition or the Hill coe cient, respectively; [C] the VGB concentration; and E max the maximal reduction in channel opening probability (i.e., TRAM-34sensitive channel activity) caused by VGB.

Statistical analyses
Linear or nonlinear curve-tting (e.g., sigmoidal or exponential curve) to the data sets collected was performed by using either Microsoft Excel Ò (Redmond, WA) or OriginPro 2016 (Microcal). The experimental data are presented as the mean ± standard error of the mean (SEM) with sample sizes (n) indicating the number of 13-06-MG cells from which the results was acquired. The Student's t-test (paired or unpaired) or one-way analysis of variance (ANOVA) followed by a post-hoc Fisher's least-signi cant difference test, was performed to analyze multiple groups. The data were examined using a nonparametric Kruskal-Wallis test, subject to possible violation in the normality underlying ANOVA. Differences were considered statistically signi cant when the P-value was below 0.05.

Results
VGB and the activity of IK Ca channels in 13-06-MG cells Experiments to evaluate the effect of VGB on IK Ca channel activity were performed. In this set of experiments, 13-06-MG cells were bathed in normal Tyrode's solution containing 1.8 mM CaCl 2 and single-channel current recordings were made. The probability of IK Ca channel opening was measured at -80 mV relative to the bath. In the presence of VGB, the IK Ca channels were signi cantly less likely to be open, compared with the control ( Figure 1A). Similar effects were observed after TRAM-34 was added to the control group ( Figure 1B). IK Ca channels that were closed in VGB-treated cells were reopened after the cells were treated with DCEBIO, an activator of IK Ca channels. This data is summarized in Figure 1C, which shows the effects of control, extracellular Ca 2+ (0 mM), extracellular Ca 2+ (3.6 mM), VGB, TRAM-34 (3 μM), and VGB (10 μM) plus DCEBIO (10 μM) on IK Ca channel activity. Each bar indicates the mean ± SEM (n=9-11). As cells were exposed to Tyrode's solution containing 3.6 mM CaCl 2 , the presence of VGB (10 mM) effectively decreased IK Ca channel activity, while it had minimal effect on it in cells bathed in Ca 2+ -free Tyrode solution. Therefore, the results enable us to indicate that the IK Ca channels measured from these cells was sensitive either to the level of extracellular Ca 2+ or to block by TRAM-34, and that VGG-mediated inhibition of IK Ca channel was attenuated by further application of DCEBIO.
VGB effect on single-channel conductance of IK Ca channels How VGB treatment affected IK Ca channels at different membrane potentials was further evaluated.
Plots of current amplitude as a function of holding potential were then constructed. Single-channel amplitudes at the potentials ranging between -80 and +20 mV were measured. Original current traces of single channel activities at the different levels of membrane potential relative to the bath obtained in the absence (left) and presence (right) of VGB (10 mM) were shown ( Figure 2A). The single-channel conductance of IK Ca channels calculated from a linear I-V relationship in the control was further calculated to yield 32.4±4 pS (n=9) over the voltage ranging between -80 and +20 mV ( Figure 2B). Of notice, the conductance measured at negative potentials was greater than that at positive voltages.
However, the single-channel slope conductance (32.1±4 pS; n=9, P>0.05) of IK Ca channels was not signi cantly changed after VGB (10 mM) treatment, despite the observed reduction in the probability of channel openings.

Concentration-dependent inhibitory effect of VGB on the activity of IK Ca channels
The relationship of the percentage suppression of IK Ca channel activity versus VGB concentration was further analyzed. In this set of experiments, each cell was maintained at -80 mV relative to the bath, and the channel open probability in the absence and presence of different VGB concentrations was measured.
As depicted in Figure 3, the addition of VGB (0.3-100 μM) suppressed the activity of IK Ca channels in a concentration-dependent manner. The IC 50 value required for its inhibitory effect on channel activity in 13-06-MG cells was calculated to be 4.21 μM, and it at a concentration of 100 μM nearly abolished the probability of channel openings. These ndings led us to indicate that VGB is able to exert a depressive action on the activity of IK Ca channels expressed in 13-06-MG cells.
Effect of VGB and VGB plus GAL-021 on the probability of BK Ca channel opening We further examined whether the presence of VGB could affect the activity of BK Ca channels in 13-06-MG cells. In these experiments, cells were immersed in a high-K + solution that contained 1.8 mM CaCl 2 , and the examined cells were held at +80 mV. As the cells were exposed to 10 μM VGB, the probability of BK Ca channels opening was not altered (Figure 4). However, following the addition of GAL-021 (10 μM) channel activity was signi cantly decreased. GAL-021 has been previously reported to be a blocker of BK Ca channels [18]. Unlike IK Ca channels, which were suppressed by VGB, the BK Ca channels were resistant to being blocked by this agent.

Effect of VGB and VGB plus BaCl 2 on K IR channel activity
In another set of single-channel current recordings, we tested whether other K + channels (i.e., K IR channels) could be affected by the presence of VGB. Cells were bathed in Ca 2+ -free Tyrode's solution and the holding potential was set at -80 mV relative to the bath. However, the presence of 10 μM VGB was unable to produce any modi cations in K IR channel activity in these cells ( Figure 5). However, the subsequent addition of 1 mM BaCl 2 in the continued presence of 10 μM VGB, effectively suppressed the probability of channel opening. BaCl 2 is regarded as an inhibitor of K IR channels [19].

Discussion
VGB is an anti-epileptic agent that is reported to be an inhibitor of gamma-aminobutyric acid breakdown. It has been approved for use as an adjunctive treatment for resistant epilepsy, and as a monotherapy for infantile spasms or West syndrome [2,3]. In this study, we found that VGB dose-dependently reduced the probability of IK Ca channel openings, and that this reduction in channel activity is voltage-dependent and associated with an increase in mean channel time closed. The reduction of the channel open probability accounts primarily for its suppression in IK Ca channel activity, despite the inability to modify singlechannel conductance in IK Ca channels. However, the activity of neither BK ca nor K ir channels was perturbed by the presence of VGB. Therefore, in addition to the inhibition of GABA breakdown, this study revealed that VGB suppressed the activity of IK Ca channels. This effect could be partly responsible for its suppression of neoplastic cells [20]. Therefore, caution needs to be appropriately exercised when the effect of this compound is explained solely by its action on GABA-ergic dysregulation [14]. However, whether there is functional coupling between GABA-receptor(s) signaling and IK Ca channel activity remains to be further studied.
The single-channel conductance of IK Ca channels in human glioma cells (13-06-MG) was calculated to be 32 pS, a value similar to the prototypical IK Ca channels present in other cell types [7,13,21], but apparently less than that of BK Ca channels [22,23]. VGB-mediated inhibition of IK Ca channel activity depends on membrane voltage and it is thought to occur via a direct interaction with the KCa3.1 channel protein in glioma cells.
In this study, the IC 50 value required for VGB-induced inhibition of IK Ca channels was 4.21 μM. There is a wide range of serum/plasma concentrations (0.8-36 mg/L) associated with successful epilepsy treatment [24]. The concentration in cerebrospinal uid was noted to be approximately 30-40% of plasma concentration, supporting that the IC 50 value of VGB in the current study could be of clinical relevance.
The presence of VGB inhibits IK Ca channels in humans at these relatively low concentrations, and in contrast to other GABA compounds, it is lipophilic and able to cross the blood-brain barrier [25]. Therefore, ndings from the present observations could be important in determining VGB's in vivo anti-neoplastic mechanism.
Different types of kinetic behaviors perturbed by VGB might facilitate its inhibition of IK Ca channel activity. VGB has no discernible effect on IK Ca single-channel conductance; therefore, the VGB molecule unlikely acts within the channel's central pore. However, the mean closed time of the channel was conceivably lengthened in its presence. Based on minimal kinetic analyses, we were able to characterize VGB-mediated inhibition of IK Ca channels by a greater a nity for the IK Ca channel in the closed (or resting) state. The activity of IK Ca channels has been previously reported to regulate the proliferation of prostate cancer cells by controlling Ca 2+ entry into these cells [8]. However, signi cant changes in neither BK Ca nor Kir channel activity were observed in these cells. The effectiveness of VGB in inhibiting IK Ca channels demonstrated here in glioma cells does not arise secondary to the reduction of intracellular Ca 2+ [26]. In the present study, VGB inhibited IK Ca channel activity within a few minutes in the 13-06-MG cells. As the onset of inhibition was rapid, its action on channel activity was most unlikely to ascribe from the binding to nuclear DNAs. The mechanism through which the VGB molecule binds to and then interact with IK Ca channels tends to be direct and not genomic, despite the detailed mechanism of VGB action remains to be further resolved.
An earlier study in which immunolabelling of KCa3.1 channels was performed, disclosed that IK Ca channels tended to be differentially expressed in excitatory and inhibitory neurons of the central nervous system [21]. Different isoforms of KCa3.1 might also be present in various types of body tissue, including gliomas; however, whether VGB is capable of modifying different types of IK Ca channels remains unknown. Further studies investigating the extent to which VGB-induced effects on glioma cells may be attributed to direct inhibitory perturbations on IK Ca channels, are thus imperatively warranted.
Of notice, the expression and function of glial Kir channels have been previously studied in retinal Müller glial cells, Schwann cells, astrocytes, and oligodendrocytes. Expression of Kir4.1 was identi ed in brain and retinal glial cells, while those of Kir2.1 and Kir2.3 were reported to be present in Schwann cells [27,28]. Whether VGA can perturb the activity of different types of Kir channels in glial cells still remains to be further resolved.
Interestingly, one in vitro study suggested that VGB should not be used for prophylaxis or the short-term treatment of epilepsy in glioblastoma [20]. However, another study suggested that blocking GABA ux into the TCA cycle, either through genetic depletion of GAD1 or pharmacological treatment with VGB, signi cantly suppressed aggressive metastatic outgrowth in the brain. Furthermore, it suggests that VGB might bring an additional bene t of stabilizing tumor-induced seizures [15].
Our previous study on temozolomide, which demonstrated its inhibitory effect on IK Ca accompanied by membrane depolarization, could describe an important underlying mechanism of temozolomide-induced antineoplastic actions [29]. Supportively, it has been reported ionizing radiation could stimulate BK Ca channel activity, resulting in Ca 2+ /calmodulin-dependent kinaces II, leading to glioblastoma cell migration [30]. As KCa3.1 has been reported to confer radioresistance to breast cancer cells [31], strategies targeting KCa3.1 in anti-cancer treatment may have good potential in modulating anti-tumor immune activity [32].
The inhibitory effect of VGB on IK Ca channels demonstrated herein sheds light on and supports the potential of VGB on antineoplastic actions. The possible link between vigabatrin/IK Ca channel activity and neoplastic cell behavior, including migration, spread, survival and proliferation is worth further investigation.

Conclusion
Our study demonstrated that the inhibitory effect of VGB on IK Ca channels could be an important underlying mechanism of VGB-induced antineoplastic actions.

Declarations
Ethics approval and consent to participate: Not applicable. This study did not involve human participants and animals.
Consent for publication: Not applicable.
Availability of data and materials: The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.