Effective Perturbations of the Amplitude, Gating, and Hysteresis of IK(DR) Caused by PT-2385, an HIF-2α Inhibitor

PT-2385 is currently regarded as a potent and selective inhibitor of hypoxia-inducible factor-2α (HIF-2α), with potential antineoplastic activity. However, the membrane ion channels changed by this compound are obscure, although it is reasonable to assume that the compound might act on surface membrane before entering the cell´s interior. In this study, we intended to explore whether it and related compounds make any adjustments to the plasmalemmal ionic currents of pituitary tumor (GH3) cells and human 13-06-MG glioma cells. Cell exposure to PT-2385 suppressed the peak or late amplitude of delayed-rectifier K+ current (IK(DR)) in a time- and concentration-dependent manner, with IC50 values of 8.1 or 2.2 µM, respectively, while the KD value in PT-2385-induced shortening in the slow component of IK(DR) inactivation was estimated to be 2.9 µM. The PT-2385-mediated block of IK(DR) in GH3 cells was little-affected by the further application of diazoxide, cilostazol, or sorafenib. Increasing PT-2385 concentrations shifted the steady-state inactivation curve of IK(DR) towards a more hyperpolarized potential, with no change in the gating charge of the current, and also prolonged the time-dependent recovery of the IK(DR) block. The hysteretic strength of IK(DR) elicited by upright or inverted isosceles-triangular ramp voltage was decreased during exposure to PT-2385; meanwhile, the activation energy involved in the gating of IK(DR) elicitation was noticeably raised in its presence. Alternatively, the presence of PT-2385 in human 13-06-MG glioma cells effectively decreased the amplitude of IK(DR). Considering all of the experimental results together, the effects of PT-2385 on ionic currents demonstrated herein could be non-canonical and tend to be upstream of the inhibition of HIF-2α. This action therefore probably contributes to down-streaming mechanisms through the changes that it or other structurally resemblant compounds lead to in the perturbations of the functional activities of pituitary cells or neoplastic astrocytes, in the case that in vivo observations occur.


Cell Preparations
The preparation of GH 3 or 13-06-MG is reported in the Supplementary Materials. References [51,52] are cited in the Supplementary Materials.

Electrophysiological Measurements
GH 3 or 13-06-MG cells were harvested and rapidly transferred to a customized chamber shortly before the electrical recordings. The chamber was positioned on the stage of an inverted microscope. Cells were kept for immersion in normal Tyrode's solution at 20-25 • C; the composition of this solution is described above. Patch-clamp recordings were undertaken under whole-cell mode with either an RK-400 (Biologic, Echirolles, France) or an AxoClamp 2B amplifier (Molecular Devices; Kim Forest, Tainan, Taiwan) [52,53]. Patch electrodes with tip resistance of 3-5 MΩ were made from Kimax-51 capillaries (#34500 (1.5-1.8 mm in outer diameter); Dogger, Tainan, Taiwan), using either a PP-830 vertical puller (Narishige, Tokyo, Japan) or a P-97 horizontal puller (Sutter, Novato, CA), and their tips were then fire-polished with MF-83 microforge (Narishige). The signals, which comprised voltages and current tracings, were stored online at 10 kHz in a touchscreen computer (ASUSPRO-BU401LG, ASUS, Tainan, Taiwan) equipped with Digidata 1440A interface (Molecular Devices), controlled by pCLAMP 10.7 software (Molecular Devices). The potentials were revised for the liquid-liquid junction potential that appeared when the composition of the pipette solution was different from the solution of the bath.

Data Analyses
To determine the concentration-dependent inhibitory effect of PT-2385 on I K(DR) , ionic current (i.e., I K(DR) ) was evoked by a depolarizing pulse (1-s duration) from −50 to +50 mV, and current amplitudes (i.e., peak or late I K(DR) ) taken with or without the addition of varying PT-2385 concentrations were measured at the beginning or end of test potential. The I K(DR) in the control period (i.e., PT-2385 was not present) was considered to be 1.0, and the values obtained with different concentrations of PT-2385 were then compared and analyzed. The PT-2385 concentration yielding a 50% reduction (i.e., IC 50 ) of peak or late I K(DR) was evaluated with goodness-of-fit assessments by using the 3-parameter logistic model (i.e., a modified form of sigmoidal Hill equation). That is, where [PT] is the PT-2385 concentration applied, IC 50 is the PT-2385 concentration required for 50% inhibition of peak or late I K(DR) , and n H is the Hill coefficient; moreover, maximal inhibition (i.e., 1 − a) was approximated from the equation. This equation converged reliably to produce the best-fit line and parameter estimates. The relationships either between the normalized amplitude and the conditioning potential (i.e., the quasi-steady-state inactivation curve of I K(DR) ) or between the instantaneous current and the upsloping end of ITRV (i.e., instantaneous current-voltage [I-V] relationship of the current) were constructed and then fitted with a Boltzmann function in the following form: Membranes 2021, 11, 636 4 of 21 In this equation, I max is the maximal amplitude of I K(DR) activated in response to step or ramp depolarization, V 1/2 is the voltage at which there is half-maximal inactivation or activation of the current, q is the apparent gating charge, F is the Faraday constant, R is the ideal gas constant, and T is the absolute temperature.
The free energy involved in the gating of I K(DR) (∆G 0 ) was evaluated on the assumption that there is a 2-state (i.e., in an equilibrium between closed (resting) and open states) gating model existing in the K V channel. The ∆G 0 for I K(DR) activation at 0 mV would be equal to q × F × V 1/2 [54][55][56][57]. The standard errors of ∆G 0 (i.e., σ qFV 1/2 ) were calculated according to: where σ q or σ V 1/2 represents the standard error in q or V 1/2 , respectively.

Statistical Analyses
Linear or nonlinear curve fitting to experimental data sets presented herein was performed using either the Microsoft Solver function embedded in Excel ® 2016 (Microsoft) or the OriginPro 2021 program (OriginLab ® ; Scientific Formosa, Kaohsiung, Taiwan) [58]. Values of experimental measurements were presented as the mean ± standard error of mean (SEM) and sample sizes (n) indicative of the cell number which the experimental results were collected from, and the error bars were hence plotted as the SEM. The Student's t test (paired or unpaired) was initially employed; however, as the statistical difference among different groups was necessarily evaluated, we performed either analysis of variance (ANOVA)-1 or ANOVA-2 (with or without repeated measure) followed by Duncan's post hoc test. Differences of mean values between groups were considered significant at p < 0.05.

Effect of PT-2385 on Delayed-Rectifier K + Current (I K(DR) ) Measured from Pituitary Tumor (GH 3 ) Cells
In this study, we initially exploited the whole-cell configuration of standard patchclamp technique to assess the perturbations of PT-2385 in the ionic currents of GH 3 cells. In order to record the currents of I K(DR) , cells were bathed in Ca 2+ -free Tyrode's solution containing 0.5 mM CdCl 2 and 1 µM tetrodotoxin (TTX) while the recording pipet was backfilled with K + -containing internal solution. TTX or CdCl 2 was used to block voltagegated Na + or Ca 2+ currents, respectively. As the whole-cell mode was established, the examined cell was maintained at the potential of −50 mV in voltage clamp mode and the depolarizing step was subsequently applied (1-s duration) to +50 mV for the elicitation of I K(DR) [27,31,32,59]. Of interest, 1 min after cell exposure to PT-2385, the amplitude of I K(DR) evoked by the 1-s-long step depolarization from -50 to +50 mV became progressively decreased ( Figure 1A). For example, at the level of +50 mV, the addition of 3 µM PT-2385 was found to diminish the peak or late amplitude of I K(DR) to 609 ± 63 or 318 ± 41 pA (n = 8, p < 0.05), respectively, from control values of 818 ± 72 or 529 ± 51 pA (n = 8). After washout of the agent, the peak or late amplitude of the current returned to 807 ± 69 or 517 ± 48 pA (n = 8, p < 0.05). In the continued presence of 3 µM PT-2385, the further addition of 10 mM tetraethylammonium chloride (TEA) almost fully abolished current amplitude. Additionally, in the presence of 3 µM PT-2385, the slow-component value in the inactivation time constant (τ inact(S) ) of I K(DR) activated by a 1-s depolarizing pulse to +50 mV was rapidly decreased to 786 ± 119 msec (n = 8, p < 0.05) from a control value of 1514 ± 217 msec (n = 8); however, a minimal change in fast component of current inactivation was seen in its presence. The results strongly suggest that PT-2385-mediated inhibition of I K(DR) in GH 3 cells is accompanied by an enhancement of the current inactivation rate ( Figure 1A). non-linear least-squares fit to the experimental data, the value of the IC50 (i.e., concentration required from a half-maximal inhibition) for the inhibitory effect of PT-2385 on peak or late IK(DR) was yielded as 8.1 or 2.2 µM, respectively. These results reflect that PT-2385 can exercise a depressant action on IK(DR) in these cells, and that the late component of IK(DR) activated by step depolarization was subject to being inhibited to a higher extent than the peak (or transient) component of the current.
Membranes 2021, 11, x FOR PEER REVIEW Figure 1. Inhibitory effect of PT-2385 (PT) on delayed-rectifier K + curren tuitary tumor (GH3) cells. Whole-cell patch-clamp experiments were un cells in Ca 2+ -free Tyrode's solution containing 1 µM tetrodotoxin (TTX the recording pipet was filled with K + -containing solution. In the uppe IK(DR) tracings were obtained for the control (i.e., PT-2385 was not prese posure to 0.3 µM PT-2385 (b), 1 µM PT-2385 (c), 3 µM PT-2385 (d), or tetraethylammonium chloride (TEA). The voltage-clamp profile is illustr In the lower part of (A), the time courses of current inactivation in the a of 3 µM PT-2385 were well approximated by a 2-exponential decay (ind The values of fast or slow components (i.e., τinact(S)) in the inactivation tim in the absence and presence of 3 µM PT-2385 were 63 or 1514 msec, and 6 (B) The concentration-response curve of PT-2385-induced block of pea GH3 cells. The continuous line represents the best fit to the modified Hi the Materials and Methods. The IC50 values for PT-2385-induced inhibiti reliably estimated to be 8.1 or 2.2 µM, respectively. Each point represe 12). (C) Relationship of the PT-2385 concentration as a function of the s tivation rate constant (1/τinact(S)) (mean ± SEM; n = 8 for each point). Of n directly proportional to the PT-2385 concentration. Based on a minimal in the text, the value of k+1 * or k−1 was estimated to be 0.124 s −1 µM −1 or 0 the KD value (k−1/k+1 * , i.e., dissociation constant) turned out to be 0.036 µM is closely similar to the IC50 value required for its inhibitory effect on lat

Kinetic Estimate of IK(DR) Block by PT-2385
During cell exposure to PT-2385, apart from the decreased I tivation rate of the current responding to step depolarization ten this regard, we extended the study to estimate the inactivation k ated inhibition of IK(DR) activated by a long depolarizing pulse were exposed to varying PT-2385 concentrations. The concentr IK(DR)-inactivation rate caused by adding this compound was con lustrated in Figure 1C. The experimental observations demons IK(DR) resulted in a concentration-dependent increase in the rate Consequently, the effect on IK(DR) identified in GH3 cells can be re The correlation between the concentration of PT-2385 and the relative amplitude of peak or late I K(DR) was further determined and established. In these experiments, we held the examined cell in a voltage clamp at −50 mV, and the depolarizing pulse (1-s duration) to +50 mV was applied, while current amplitudes in the presence of varying PT-2385 concentrations were taken at the start or end-pulse of step depolarization. As illustrated in Figure 1B, the presence of various concentrations of PT-2385 differentially suppressed the peak and late amplitudes of I K(DR) in a concentration-dependent manner. By virtue of a non-linear least-squares fit to the experimental data, the value of the IC 50 (i.e., concentration required from a half-maximal inhibition) for the inhibitory effect of PT-2385 on peak or late I K(DR) was yielded as 8.1 or 2.2 µM, respectively. These results reflect that PT-2385 can exercise a depressant action on I K(DR) in these cells, and that the Membranes 2021, 11, 636 6 of 21 late component of I K(DR) activated by step depolarization was subject to being inhibited to a higher extent than the peak (or transient) component of the current.

Kinetic Estimate of I K(DR) Block by PT-2385
During cell exposure to PT-2385, apart from the decreased I K(DR) amplitude, the inactivation rate of the current responding to step depolarization tended to become raised. In this regard, we extended the study to estimate the inactivation kinetics of PT-2385mediated inhibition of I K(DR) activated by a long depolarizing pulse in situations where cells were exposed to varying PT-2385 concentrations. The concentration dependence of the I K(DR) -inactivation rate caused by adding this compound was constructed and is hence illustrated in Figure 1C. The experimental observations demonstrated that its effects on I K(DR) resulted in a concentration-dependent increase in the rate of current inactivation. Consequently, the effect on I K(DR) identified in GH 3 cells can be reasonably explained by a state-dependent blocking mechanism through which this compound is allowed to preferentially bind to the open or open-inactivated state of the K V channels. The reaction scheme for this viewpoint is thus simply

O·[B]
Alternatively, the dynamical system yielding three equations becomes where α or β is called the kinetic constant for the opening or closing of the K V channel, respectively; k +1 * or k −1 is that for blocking (i.e., The blocking (forward, k +1 * ) or unblocking (backward, k −1 ) rate constant was estimated based on the slow component in the inactivation time constants (τ inact(S) ) of I K(DR) during cell exposure to varying PT-2385 concentrations ( Figure 1C). The relationship can be satisfied by 1 In this equation, k +1 * or k −1 was respectively derived from the slope or from the y-axis intercept at [B] = 0 (i.e., the point at which the line crosses the y-axis) of the linear regression at which the reciprocal time constants of the current (1/τ inact(S) ) versus the PT-2385 concentration was interpolated. The resultant relationship between 1/τ inact(S) and [B] was found in the straight line through experimental data points with a correlation coefficient of 0.95 ( Figure 1C). The blocking or unblocking rate constant was thereafter yielded to be 0.124 s −1 µM −1 or 0.362 s −1 , respectively. According to the evolving rate constants, dividing k −1 by k +1 * gave a dissociation constant (K D ) of 2.9 µM during cell exposure to PT-2385, a value that was noted to be close to the effective IC 50 (2.2. µM) of this compound needed for inhibition of late I K(DR) .

Comparison among the Effects of PT-2385 (PT), PT-2385 Plus Diazoxide (Diaz), and PT-2385 Plus Cilostazol (Cil), or PT-2385 Plus Sorafenib (SOR) on the Amplitude of I K(DR)
We continued to evaluate and then compare the effect of PT-2385 and PT-2385 plus different related compounds (e.g., Diaz, Cil, SOR) on I K(DR) magnitude. As summarized in Figure 2, the amplitude of I K(DR) was effectively inhibited by the addition of 3 µM PT-2385. However, as cells were continually exposed to PT-2385 (3 µM), the subsequent addition of either diazoxide (Diaz, 30 µM), cilostazol (Cil, 10 µM), or sorafenib (SOR, 10 µM) did not result in any further modifications to the PT-2385-mediated block of peak or late I K(DR) . Diaz or Cil has been reported to activate ATP-sensitive K + (K ATP ) channels or large-conductance Ca 2+ -activated K + (BK Ca ) channels, respectively [34,60,61], while SOR was previously recognized to suppress the activity of tyrosine kinase [62]. Additionally, the presence of Diaz (30 µM), Cil (10 µM), or SOR (10 µM) alone was not noted to have significant effects on the I K(DR) amplitude observed in GH 3 cells. Under this circumstance, it is reasonable to assume that the PT-2835-mediated block of I K(DR) in GH 3 cells tends to be independent of either the inhibition of K ATP or BK Ca channels, or the stimulation of tyrosine-kinase activity.

Inhibitory Effect of PT-2385 on Mean Current versus Voltage (I-V) Relationship of IK(DR)
In another separate set of experiments, we voltage-clamped at -50 mV and then applied a series of command voltage pulses (1-s duration) from -60 to +50 mV in 10 mV increments to the examined GH3 cells. Under this voltage-clamp protocol, a family of IK(DR) was robustly elicited, and the currents were manifested by an outwardly directed rectifying property with a small relaxation in the time course of current inactivation [31,41,45,59,62]. Of note, after 1 min of cell exposure to 1 µM PT-2385, the IK(DR) magnitude The whole experiments were performed in cells bathed in Ca 2+ -free Tyrode's solution and the recording pipet was filled with K + -enriched internal solution. The I K(DR) amplitude activated by a 1-s depolarizing step from −50 to +50 mV was taken at the end-pulse of command potential. In this set of experiments on PT-2385 plus tested compounds (e.g., Diaz, Cil, and SOR), each tested compound was subsequently added to the bath in the continued presence of PT-2385 (3 µM). * Significantly different from control (p < 0.05).

Inhibitory Effect of PT-2385 on Mean Current versus Voltage (I-V) Relationship of I K(DR)
In another separate set of experiments, we voltage-clamped at -50 mV and then applied a series of command voltage pulses (1-s duration) from -60 to +50 mV in 10 mV increments to the examined GH 3 cells. Under this voltage-clamp protocol, a family of I K(DR) Membranes 2021, 11, 636 8 of 21 was robustly elicited, and the currents were manifested by an outwardly directed rectifying property with a small relaxation in the time course of current inactivation [31,41,45,59,62]. Of note, after 1 min of cell exposure to 1 µM PT-2385, the I K(DR) magnitude became reduced, particularly at the potentials ranging between 0 and +50 mV ( Figure 3A). Figure 3B respectively illustrates the mean I-V relationships of I K(DR) measured at the start (initial peak) and end-pulse (late or sustained) of each potential obtained in the control period (upper panel) or during exposure to 1 µM PT-2385 (lower panel). For example, at the level of +50 mV, PT-2385 at a concentration of 1 µM led to decreased peak amplitude from 803 ± 62 to 667 ± 46 pA (n = 8, p < 0.05), while, at the same level of voltage pulse, this compound at the same concentration reduced the late I K(DR) from 635 ± 61 to 381 ± 41 pA (n = 8, p < 0.05). After the agent was removed, the peak or late amplitude of I K(DR) was back to 786 ± 58 or 611 ± 57 pA (n = 7, p < 0.05). It is conceivable, therefore, that the potency of the PT-2385-mediated block of late or sustained I K(DR) activated by different step depolarizations (i.e., the potentials ranging between 0 and −50 mV) was higher than that for its blocking on the instantaneous peak component of the current.

The Steady-State Inactivation Curve of I K(DR) during Exposure to PT-2385
When cells were exposed to varying concentrations of PT-2385, the I K(DR) activated via membrane depolarization displayed an evident peak, followed by an exponential relaxation to a steady-state level. Hence, we also investigated the quasi-steady-state inactivation curve of I K(DR) achieved in the presence of PT-2385 by exploiting a two-step voltage protocol. In this set of current-recording experiments, we placed cells in Ca 2+ -free Tyrode's solution, and then filled up the pipet with K + -enriched solution during the measurements. Once a wholecell configuration was established, we delivered a two-pulse protocol to the examined cells in situations where varying PT-2385 concentrations were present. According to the least-squares minimization procedure, the best-fitting parameters (i.e., V 1/2 and q) for the I K(DR) -inactivation curve by a test of goodness of fit were convergently obtained in the presence of 1 or 3 µM PT-2385. Consequently, we constructed the normalized amplitude of I K(DR) (i.e., I/I max ) against the conditioning command potentials ( Figure 4A,B), and the continuous sigmoidal curve by adjusting the line to minimize the vertical deviations between the data points, and the line was approximately fitted with a single Boltzmann isotherm, as detailed in Materials and Methods. The values of V 1/2 during exposure to 1 µM and 3 µM PT-2385 were calculated to be −34.7 ± 1.8 and −41.3 ± 1.8 mV (n = 8), respectively, while the q values (i.e., apparent gating charge) were 2.0 ± 0.3 and 2.1 ± 0.3 e (n = 8), respectively. These experimental results showed that during exposure to varying PT-2385 concentrations (i.e., 1 and 3 µM), the V 1/2 value of the I K(DR) -inactivation curve in GH 3 cells was modified, despite its ineffectiveness in changing the apparent gating charge of the current (i.e., the steepness of the inactivation curve).  In these experiments, cell membrane potentials were held at −50 mV, and the depolarizing command pulse to a series of voltage steps (1-s duration) ranging between −60 and +50 mV in 10-mV increments (as indicated in the uppermost of (A)) was thereafter applied to evoke the currents. (A) Representative current traces in the control period (upper image) and during exposure to 1 µM PT- mann isotherm, as detailed in Materials and Methods. The values of V1/2 during exposure to 1 µM and 3 µM PT-2385 were calculated to be −34.7 ± 1.8 and −41.3 ± 1.8 mV (n = 8), respectively, while the q values (i.e., apparent gating charge) were 2.0 ± 0.3 and 2.1 ± 0.3 e (n = 8), respectively. These experimental results showed that during exposure to varying PT-2385 concentrations (i.e., 1 and 3 µM), the V1/2 value of the IK(DR)-inactivation curve in GH3 cells was modified, despite its ineffectiveness in changing the apparent gating charge of the current (i.e., the steepness of the inactivation curve). We also investigated the recovery from I K(DR) block by PT-2385 by the use of another two-step voltage-clamp profile. The voltage-clamp protocol applied consists of an initial (i.e., the first conditioning) depolarizing step from −50 to +50 mV, sufficiently long to ensure the block to reach a steady-state level; subsequently, following different inter-pulse intervals, a second command pulse was applied at the same potential as the first conditioning one ( Figure 5A). As the ratios (second current/first current, or relative amplitude) of the peak amplitude of I K(DR) activated by the first and the second conditioning pulse were derived for a measure of recovery from I K(DR) block in the presence of PT-2385, the value of the recovery time constant was constructed and hence plotted as a function of the inter-pulse interval ( Figure 5B). The time course for the recovery of I K(DR) block in the presence of 1 or 3 µM PT-2385 was noted to be described by a single-exponential function. By minimizing the squares of the deviations, the time constant for current recovery in the presence of 1 µM PT-2385 was estimated to be 0.59 ± 0.04 s (n = 7), significantly smaller than the value of 0.83 ± 0.06 s (n = 7, p < 0.05) which was obtained during exposure to 3 µM PT-2385. The experimental observations thus tempted us to imply that the slowing of I K(DR) -current recovery caused by adding PT-2385 would result from its preferential block in the open or open-inactivated state of the K V channel. ence of 1 or 3 µM PT-2385 was noted to be described by a single-exponential function. By minimizing the squares of the deviations, the time constant for current recovery in the presence of 1 µM PT-2385 was estimated to be 0.59 ± 0.04 s (n = 7), significantly smaller than the value of 0.83 ± 0.06 s (n = 7, p < 0.05) which was obtained during exposure to 3 µM PT-2385. The experimental observations thus tempted us to imply that the slowing of IK(DR)-current recovery caused by adding PT-2385 would result from its preferential block in the open or open-inactivated state of the KV channel.

Effect of PT-2385 on the Hysteretic Behavior of I K(DR) Triggered by Isosceles-Triangular Ramp Voltage (ITRV) with Varying Durations
Previous studies have demonstrated the capability of I K(DR) strength to influence either varying patterns of bursting firing or action potential configurations in many types of electrically excitable cells [31,37,38,[40][41][42][63][64][65]. Therefore, we continued to determine whether and how the presence of PT-2385 could adjust the I K(DR) strength activated in response to long-lasting ITRV with varying durations. In this set of experiments, in the absence or presence of PT-2385, we held the examined cell at −50 mV, and an upsloping (forward) limb from −80 to +60 mV followed by a downsloping (backward) limb back to −80 mV (i.e., upright ITRV) with varying durations (between 0.4 and 3.2 s) were thereafter applied ( Figure 6A,B). Under this experimental condition, the voltage-dependent hysteresis of I K(DR) in response to upright ITRV with varying durations was observed. Of additional interest, during cell exposure to 1 or 3 µM PT-2385, the strength of current responding to both rising and falling limbs of upright ITRV was progressively decreased with increasing the PT-2385 concentrations ( Figure 6A). For example, as the duration of isosceles-triangular ramp pulse applied was set at 3.2 s (i.e., the ramp speed of ±87.5 mV/s), the value of ∆area (i.e., the difference in area encircled by the curve in the forward and backward direction) for voltage-dependent hysteresis in the control period (i.e., PT-2385 was not present) was 3.18 ± 0.41 mV·nA (n = 8), while at the duration of 3.2 s (i.e., ramp speed at ±87.5 mV/s), the value of ∆area (indicated in shaded area) in the presence of 3 µM PT-2385 was significantly reduced to 1.18 ± 0.25 mV·nA (n = 8, p < 0.05) ( Figure 6C). Findings from these data led us to propose that there was the occurrence of voltage-dependent hysteresis for I K(DR) activation in response to upright ITRV in GH 3 cells [66], and that the hysteretic strength of the current was noticeably reduced with increasing PT-2385 concentration.
value of ∆area (i.e., the difference in area encircled by the curve in the forward and backward direction) for voltage-dependent hysteresis in the control period (i.e., PT-2385 was not present) was 3.18 ± 0.41 mV·nA (n = 8), while at the duration of 3.2 s (i.e., ramp speed at ±87.5 mV/s), the value of ∆area (indicated in shaded area) in the presence of 3 µM PT-2385 was significantly reduced to 1.18 ± 0.25 mV·nA (n = 8, p < 0.05) ( Figure 6C). Findings from these data led us to propose that there was the occurrence of voltage-dependent hysteresis for IK(DR) activation in response to upright ITRV in GH3 cells [66], and that the hysteretic strength of the current was noticeably reduced with increasing PT-2385 concentration.

Effect of PT-2385 on the Activation Energy Required for IK(DR) Elicitation by ITRV
In another set of experiments, we ascertained whether PT-2385 could perturb any modifications in the free energy during IK(DR) elicitation caused by ITRV. As illustrated in Figure 7A, in the absence or presence of 3 µM PT-2385, the instantaneous (or transient) activation curve of the current elicited by the upsloping (forward) end of ITRV with a ramp speed of 87.5 mV/s was obtained. Thereafter, the Boltzmann isotherm (detailed in Materials and Methods) was performed with a test of goodness to fit to experimental data points in order to estimate the values of q and V1/2 for such instantaneous activation curves of ITRV-induced IK(DR). Based on the values of q and V1/2 during the upsloping end of upright ITRV, the free energy entailed in the gating of IK(DR) activation at 0 mV (∆G0) in the absence or presence of varying PT-2385 concentrations (i.e., ∆G0 = q × F × V1/2) was derived

Effect of PT-2385 on the Activation Energy Required for I K(DR) Elicitation by ITRV
In another set of experiments, we ascertained whether PT-2385 could perturb any modifications in the free energy during I K(DR) elicitation caused by ITRV. As illustrated in Figure 7A, in the absence or presence of 3 µM PT-2385, the instantaneous (or transient) activation curve of the current elicited by the upsloping (forward) end of ITRV with a ramp speed of 87.5 mV/s was obtained. Thereafter, the Boltzmann isotherm (detailed in Materials and Methods) was performed with a test of goodness to fit to experimental data points in order to estimate the values of q and V 1/2 for such instantaneous activation curves of ITRV-induced I K(DR) . Based on the values of q and V 1/2 during the upsloping end of upright ITRV, the free energy entailed in the gating of I K(DR) activation at 0 mV (∆G 0 ) in the absence or presence of varying PT-2385 concentrations (i.e., ∆G 0 = q × F × V 1/2 ) was derived and is thereafter illustrated in Figure 7B. Of note, as the PT-2385 concentration was raised, the activation energy (∆G 0 value) required for I K(DR) elicitation was progressively augmented during upright ITRV with a duration of 3.6 s. For example, as GH 3 cells were exposed to 3 µM PT-2385, the ∆G 0 value was significantly elevated to 2.98 ± 0.14 kJ/mol from a control value of 1.98 ± 0.07 kJ/mol (n = 7, p < 0.05).

Effect of PT-2385 on the Hysteresis Activated During Inverted ITRV
We continued to determine whether there is hysteresis of IK(DR) elicited by inverted ITRV at varying durations. As demonstrated in Figure 8A, in keeping with the abovestated results from upright ITRV, the hysteretic strength of the current by inverted ITRV in the absence and presence of PT-2385 was clearly observed in combination with a gradual decline in peak IK(DR) with decreasing ramp speed. As shown in Figure 8B, the relationship of the instantaneous current versus membrane potential was constructed during inverted ITRV with ramp speed of ±87.5 mV/s. Our experimental observations demonstrate the effectiveness of PT-2385 in diminishing the inverted ITRV-induced strength of voltage-dependent hysteresis; meanwhile, the hysteretic magnitude activated by inverted ITRV tends to be stronger than that by upright ITRV with the same ramp speed.

Effect of PT-2385 on the Hysteresis Activated during Inverted ITRV
We continued to determine whether there is hysteresis of I K(DR) elicited by inverted ITRV at varying durations. As demonstrated in Figure 8A, in keeping with the above-stated results from upright ITRV, the hysteretic strength of the current by inverted ITRV in the absence and presence of PT-2385 was clearly observed in combination with a gradual decline in peak I K(DR) with decreasing ramp speed. As shown in Figure 8B, the relationship of the instantaneous current versus membrane potential was constructed during inverted ITRV with ramp speed of ±87.5 mV/s. Our experimental observations demonstrate the effectiveness of PT-2385 in diminishing the inverted ITRV-induced strength of voltagedependent hysteresis; meanwhile, the hysteretic magnitude activated by inverted ITRV tends to be stronger than that by upright ITRV with the same ramp speed.

Ability of PT-2385 to Inhibit IK(DR) Amplitude in Human 13-06-MG Glioma Cells
PT-2385 has been recently reported to inhibit the proliferation of glioma cells [18,19]. In a final stage of experiments, we thus intended to explore whether the presence of PT-2385 was able to modulate ionic currents in other types of neoplastic cells (e.g., malignant glioma cells). As illustrated in Figure 9, in whole-cell current recordings, the magnitude of ramp pulse-induced IK(DR) in 13-06-MG glioma cells was robustly detected, as described previously [67]. After 1 min of cell exposure to PT-2385, the amplitude of IK(DR) was progressively decreased. For example, at the level of +80 mV, the addition of 3 µM PT-2385 significantly decreased the current amplitude from 1221 ± 228 to 732 ± 187 pA (n = 7, p < 0.05). After washout of the compound, the amplitude returned to 1217 ± 207 pA (n = 7, p < 0.05). Therefore, similar to the GH3 cells stated above, the IK(DR) existing in 13-06-MG cells was noted to be sensitive to inhibition by PT-2385. Apart from the inhibition of HIF-2α, PT-2385-induced changes in malignant behaviors of glioma [18,19] could, to some extent, be closely associated with its inhibitory action of IK(DR). PT-2385 has been recently reported to inhibit the proliferation of glioma cells [18,19]. In a final stage of experiments, we thus intended to explore whether the presence of PT-2385 was able to modulate ionic currents in other types of neoplastic cells (e.g., malignant glioma cells). As illustrated in Figure 9, in whole-cell current recordings, the magnitude of ramp pulse-induced I K(DR) in 13-06-MG glioma cells was robustly detected, as described previously [67]. After 1 min of cell exposure to PT-2385, the amplitude of I K(DR) was progressively decreased. For example, at the level of +80 mV, the addition of 3 µM PT-2385 significantly decreased the current amplitude from 1221 ± 228 to 732 ± 187 pA (n = 7, p < 0.05). After washout of the compound, the amplitude returned to 1217 ± 207 pA (n = 7, p < 0.05). Therefore, similar to the GH 3 cells stated above, the I K(DR) existing in 13-06-MG cells was noted to be sensitive to inhibition by PT-2385. Apart from the inhibition of HIF-2α, PT-2385-induced changes in malignant behaviors of glioma [18,19] could, to some extent, be closely associated with its inhibitory action of I K(DR) .

Discussion
In the present study, cell exposure to PT-2385 was able to decrease the amplitude of IK(DR) and raise the inactivation time course of the current. A concentration-dependent inhibition of peak or late IK(DR), with respectively effective IC50 values of 8.1 or 2.2 µM, was observed as GH3 cells were exposed to different PT-2385 concentrations. The value of the dissociation constant (KD) derived from quantitative estimate of the inactivation time course of IK(DR) was yielded to be 2.8 µM. Moreover, cell exposure to this compound not only decreased the magnitude of voltage-dependent hysteresis for IK(DR) in reply to upright or inverted isosceles-triangular ramp voltage (ITRV), but also increased free energy involved in the gating IK(DR) elicitation. It could also decrease the amplitude of IK(DR) in human 13-06-MG glioma cells. Collectively, the PT-2385-mediated effectiveness in the modifications of ionic currents described herein tends to be independent of an interaction with HIF-2α and it could thus be a confounding factor affecting neuronal or endocrine function.
Perhaps more noticeable with respect to the issue of the magnitude in the PT-2385mediated block of IK(DR) observed in GH3 cells is that the rising phase of the current activated in response to rapid membrane depolarization remained little-altered in its presence; however, the addition of this compound had the propensity to accelerate IK(DR) inactivation in a time-, state-, and concentration-dependent fashion. This could thus be interpreted to mean that PT-2385 is able to reach the blocking site in situations where the KV channel favorably resides in the open or open-inactivated state. Consequently, this property is reasonably explained by the first-order binding scheme (i.e., closed ↔ open ↔ open • blocked), as demonstrated above [30,59]. Pertinent to such reaction scheme is that open-blocked KV channels are not closed unless the PT-2385 dissociates from the binding site. Meanwhile, the time-dependent block produced by PT-2385 reflects that it preferentially binds to and blocks an open state (or conformation) of the channel with a KD value of 2.9 µM observed in GH3 cells, a value which is close to the IC50 value (i.e., 2.2 µM) entailed for its inhibitory effect on late IK(DR), not on peak IK(DR).
The magnitude of IK(DR) reduced by PT-2385, concomitantly with a considerable rise in current inactivation in response to long sustained depolarization, may be the cause of

Discussion
In the present study, cell exposure to PT-2385 was able to decrease the amplitude of I K(DR) and raise the inactivation time course of the current. A concentration-dependent inhibition of peak or late I K(DR) , with respectively effective IC 50 values of 8.1 or 2.2 µM, was observed as GH 3 cells were exposed to different PT-2385 concentrations. The value of the dissociation constant (K D ) derived from quantitative estimate of the inactivation time course of I K(DR) was yielded to be 2.8 µM. Moreover, cell exposure to this compound not only decreased the magnitude of voltage-dependent hysteresis for I K(DR) in reply to upright or inverted isosceles-triangular ramp voltage (ITRV), but also increased free energy involved in the gating I K(DR) elicitation. It could also decrease the amplitude of I K(DR) in human 13-06-MG glioma cells. Collectively, the PT-2385-mediated effectiveness in the modifications of ionic currents described herein tends to be independent of an interaction with HIF-2α and it could thus be a confounding factor affecting neuronal or endocrine function.
Perhaps more noticeable with respect to the issue of the magnitude in the PT-2385mediated block of I K(DR) observed in GH 3 cells is that the rising phase of the current activated in response to rapid membrane depolarization remained little-altered in its presence; however, the addition of this compound had the propensity to accelerate I K(DR) inactivation in a time-, state-, and concentration-dependent fashion. This could thus be interpreted to mean that PT-2385 is able to reach the blocking site in situations where the K V channel favorably resides in the open or open-inactivated state. Consequently, this property is reasonably explained by the first-order binding scheme (i.e., closed ↔ open ↔ open·blocked ), as demonstrated above [30,59]. Pertinent to such reaction scheme is that open-blocked K V channels are not closed unless the PT-2385 dissociates from the binding site. Meanwhile, the time-dependent block produced by PT-2385 reflects that it preferentially binds to and blocks an open state (or conformation) of the channel with a K D value of 2.9 µM observed in GH 3 cells, a value which is close to the IC 50 value (i.e., 2.2 µM) entailed for its inhibitory effect on late I K(DR) , not on peak I K(DR) .
The magnitude of I K(DR) reduced by PT-2385, concomitantly with a considerable rise in current inactivation in response to long sustained depolarization, may be the cause of its perturbing effects on membrane excitability in varying types of electrically excitable cells (e.g., neurons and neuroendocrine or endocrine cells). Because the recovery from current inactivation in the presence of PT-2385 observed in this study also became slowed, a duration greater than 3 s was required for current magnitude to recover completely. The steady-state inactivation curve of I K(DR) was more hyperpolarized as the PT-2385 concentration was increased. As a consequence, a PT-2385-perturbed block of I K(DR) will even become pronounced when a train of action potentials occurs, since the I K(DR) magnitude, especially at the occurrence of transient resurgent K + currents, is decreased as a function of firing frequency [32,[40][41][42].
Recent pharmacokinetic studies have shown that after being administered orally at twice-per-day dose of 800 mg, the plasma level of PT-2385 in patients with clear cell renal cell carcinoma was reported to reach around 3000-5000 ng/mL (or 7.8-13.0 µM) [20]. Therefore, it is possible that the used concentration of PT-2385 in this study tends to be achievable in the plasma of treated patients. Any changes in the amplitude, gating, and hysteresis of I K(DR) caused by PT-2385 depend not simply on the PT-2385 concentration, but also on the preexisting resting potential and varying bursting patterns of excitable cells. From this perspective, the observed effects by PT-2385 in this study are likely to arise to the extent of clinically achievable concentrations.
There seems to be a divergence in the IC 50 value between the PT-2385-induced block of late I K(DR) (around 2.2 µM) and its effect on HIF-2α inhibition (around 27 nM) [22]. The plausible interpretation for this difference is unknown at present; however, it could be due to the varying experimental maneuvers used in the study. In terms of investigations on HIF-2α inhibition, surface-membrane components in host cells could have mostly been lysed and removed; consequently, PT-2385 could readily reach HIF-2α inside the cells [13,14,22]. However, in the present observations, intact cells were used to investigate different types of membrane ionic currents in GH 3 cells or 13-06-MG glioma cells.
Under hypoxic conditions along with possible changes in HIF expression, the activity of ATP-dependent K + (K ATP ) channels in GH 3 cells or 13-06-MG glioma cells could be elevated during exposure to PT-2385 [47][48][49][50]. The presence of PT-2385 was previously demonstrated to counteract the adverse effect of sorafenib (SOR), known to suppress the activity of tyrosine kinases, in hepatocellular carcinoma [16]. However, in our wholecell current recording, the intracellular solution contained 3 mM ATP, a value known to adequately suppress K ATP -channel activity [60,68]. However, the PT-2385-mediated inhibition of I K(DR) observed in GH 3 cells was little -affected by the subsequent application of either 30 µM diazoxide (Diaz) or 10 µM cilostazol (Cil). Diaz or Cil were respectively reported to be stimulators of K ATP channels or large-conductance Ca 2+ -activated K + (BK Ca ) channels [34,60,61]. Moreover, in the continued presence of PT-2385, the subsequent addition of sorafenib (SOR), an inhibitor of tyrosine kinase [62], failed to modify PT-2385induced block of I K(DR) in GH 3 cells. As such, a PT-induced reduction in whole-cell I K(DR) strength depicted in recent study is unlikely to be linked either to inhibition of K ATP -or BK Ca -channel activity, or to stimulation of tyrosine-kinase activity, although these channels or enzymes are functionally active in pituitary cells [60][61][62].
It needs to be emphasized that as it is administered, PT-2385 needs to enter cytosol or even the nucleus before being able interact with HIF-2α or the HIF-α/HIF-β complex [12][13][14]21]. In other words, despite being small in size, the molecules must pass through cell-surface membrane before entering the cell interior and then reaching the target. In this regard, it is reasonable to assume that the perturbations to plasmalemmal ionic currents caused by PT-2385 would precede its subsequent actions through the inhibition of HIF-2α. The extent of PT-2385-mediated modifications to ionic currents (e.g., I K(DR) ) is closely linked to the inhibition of HIF-2α, followed by changes in proliferation of different types of neoplastic cells [12][13][14]16,17,26].
The perturbations by PT-2385 of membrane ionic currents shown here tend to be acute in onset. Hence, the actions could be unlinked to its interaction with the activity of HIF-2α, as stated recently [9,12,14,25]. Both the steady-state inactivation curve of I K(DR) and recovery from current inactivation were demonstrated in the presence of PT-2385. The actions of this compound on the magnitude or kinetic gating of I K(DR) are thought to occur either through its preferential binding to the open state of the K V channel or through an interaction with the channel's open conformation.
The phenomenon of the voltage-dependent hysteresis of ionic currents has been proposed to play roles in affecting the electrical behavior of electrically excitable cells. Consistent with previous observations on other types of ionic currents [45,52,57,66], the I K(DR) present in GH 3 cells underwent hysteresis in which there is likely to be a mode shift, because the voltage sensitivity of the gating charge movement is thought to depend on the previous state of the channel [57,66]. Moreover, we further evaluated the possible perturbation by PT-2385 of the instantaneous and non-equilibrium property inherent in I K(DR) activated by upright or inverted ITRV. The results found PT-2385 s effectiveness in decreasing the ∆area of hysteresis for I K(DR) elicited by ITRV. Moreover, the presence of PT-2385 could increase the ∆G 0 value for I K(DR) activation in response to an instantaneous upsloping ramp pulse. In this regard, any modifications of I K(DR) caused by PT-2385 depend not only on the PT-2385 concentration used but also various factors that include the pre-existing resting potential or varying firing patterns of action potentials.
It is important to mention that a possible link between HIF and regulation of K Vchannel activity has been demonstrated [23,[47][48][49][50]. Another study [69] showed the ability of HIF-1 to regulate K V 1.2 channels in rat PC12 cells. It is possible that HIF-2α regulates K V -channel activity directly. Whether PT-2385 exercises an inhibitory effect on I K(DR) in GH 3 or 13-05-MG cells where HIF-2α is knocked down remains to be further investigated.
All in all, in this study, we hitherto revealed the unidentified action of PT-2385 through which it could modify the amplitude, gating, and hysteretic behavior of I K(DR) . The modifications to the magnitude, gating, and voltage-dependent hysteresis of I K(DR) caused by the presence of PT-2385 are not simply a bystander effect, and may potentially converge to act on the functional activities of neurons and endocrine or neuroendocrine cells if similar in vivo observations occur. To what extent such direct actions on I K(DR) contribute to the rapid impairment in hypoxic ventilatory drive [62] remains to be further resolved.