Swelling-Activated Potassium Channel in Porcine Pigmented Ciliary Epithelial Cells

Abstract

Purpose.: Ion channels in the ciliary epithelium play critical roles in the formation of aqueous humor in the eye. The present study identified a novel, swelling-activated K⁺ current in freshly dissociated porcine pigmented ciliary epithelial cells.

Methods.: Ciliary epithelial cells were freshly dissociated from porcine eyes. Whole-cell currents were recorded by the patch-clamp technique in pigmented and nonpigmented ciliary epithelial cell (PCE–NPCE) pairs or single PCE cells.

Results.: The 0-current potential was −49 ± 13 mV in PCE–NPCE cell pairs (n = 97) and −52 ± 12 mV in single PCE cells (n = 30). Whole-cell currents in these cells were dominated by an outwardly rectifying K⁺ current activated by potentials more positive than −90 mV, which never inactivated during prolonged depolarization. The K⁺ current was significantly augmented by hypotonic cell perfusion. External Ba²⁺ was a blocker of this K⁺ conductance (IC₅₀ of 0.38 mM), but the conductance was insensitive to external TEA⁺. Linopirdine, a specific inhibitor of KCNQ channels, effectively blocked the K⁺ current in these PCE cells.

Conclusions.: Porcine PCE cells express a swelling-activated K⁺ channel, which may be a member of the KCNQ/Kv7 channel family. This K⁺ channel is active near resting potentials and could contribute to the regulation of cell volume and water transport via the ciliary epithelia.

The ciliary body epithelium is a transporting epithelium, composed of two epithelial layers, a nonpigmented ciliary epithelial (NPCE) cell layer facing the vitreal space, and a pigmented ciliary epithelial (PCE) cell layer facing the ciliary stroma. By transporting ions and water between the ciliary body and vitreous space, the ciliary epithelium plays critical roles in the formation of aqueous humor in the eye. 

Studies on ion transport via the ciliary epithelium have mainly focused on components underlying chloride secretion, which limit aqueous humor production. ^1,2 In particular, Cl⁻ channels in NPCE cells mediate the final step of Cl⁻ release and thereby regulate the rate of aqueous humor secretion. The molecular identity of NPCE Cl⁻ channels has not been precisely elucidated, but several lines of evidence indicate that the ClC-3 member of the ClC family of channels and transporters likely plays a major role in Cl⁻ inflow. ^3,4 In addition to anion channels, epithelial potassium channels also contribute to critical cellular functions including the secretion of ions and water across the epithelia. The rate of anion secretion depends on the activity of K⁺ channels since cell hyperpolarization caused by the opening of K⁺ channels increases the driving force for anion efflux. ⁵ Several previous studies have described some K⁺ channels expressed in the ciliary epithelium, including voltage-gated delayed rectifiers, ^{6 –8} Ca²⁺-activated outward K⁺ channels ^7,9,10 and an inward rectifier activated by hyperpolarization. ^6,11 It is unknown whether these K⁺ currents in the ciliary epithelium are sensitive to cell swelling. 

In the retinal pigment epithelium (RPE), another ocular secretory epithelium, K⁺ conductance is mainly composed of two types of K⁺ channels: an inward rectifier Kir7.1 ^12,13 and a M-type K⁺ channel. ^14,15 The latter was so named because of its resemblance to the M-current, a K⁺ current distributed in neurons and other excitable cells. This M-type K⁺ current in RPE is a voltage-gated K⁺ channel with an activation threshold near the equilibrium potential for K⁺ and thus contributes to the resting potential of the cell. One of the characteristic properties of the M-type K⁺ current in the RPE is its insensitivity to tetraethylammonium (TEA), a classic K⁺ channel blocker. ^14,15  

In this study, we identified a swelling-activated, noninactivating outward K⁺ current in freshly dissociated PCE cells. The K⁺ current was essentially insensitive to TEA and resembled the M-type K⁺ current of RPE in terms of voltage-dependence, kinetics, and blocker sensitivities. 

Methods

Solutions

External and internal solutions were buffered with HEPES as in previous patch-clamp studies of corneal epithelium, ¹⁶ and RPE. ¹⁵ Briefly, the standard Ringer's solution consisted of (in mM) 135 NaCl, 5.0 KCl, 10 HEPES, 10 glucose, 1.8 CaCl₂, and 1.0 MgCl₂, titrated to pH 7.4 with NaOH. The osmolarity of standard Ringer's solution was 302 ± 6 mOsmol/L. The standard pipette solution consisted of (in mM) 30 KCl, 83 potassium gluconate, 5.0 HEPES, 5.5 EGTA-KOH, 0.5 CaCl₂, and 2.0 MgCl₂, titrated to pH 7.2 with KOH. The osmolarity of standard pipette solution was 257 ± 5 mOsmol/L. An amphotericin B-stock solution (1.2 mg amphotericin B/50 μL DMSO) was added to 2.0 mL of the pipette solution to give a final concentration of 240 μg/mL. The cell isolation medium was similar to that used in a previous study on RPE and contained (in mM) 135 N-methyl-d-glucamine (NMDG)-Cl, 5.0 KCl, 10 HEPES, 3.0 EDTA-KOH, 10 glucose, 3.0 cysteine, 1.0 glutathione, 1.0 L(+)-ascorbic acid, 1.0 taurine, and 0.2 mg/mL papain (type III), and was titrated to pH 7.4 with NaOH. In experiments in which the concentrations of K⁺ or TEA⁺ were varied, NaCl was replaced by an equimolar amount of the appropriate Cl⁻ salt. A hypotonic solution was made by diluting standard Ringer's solution to 20% and adjusting [K⁺] to 5 mM to give an osmolarity of 239 ± 2 mOsmol/L. Linopirdine was dissolved in DMSO and then diluted in Ringer's solution. Throughout the experiments, the final concentration of DMSO in the perfusates was no more than 0.2%, which did not affect the whole-cell currents in this study. 

NaCl, KCl, CaCl₂, and L(+)-ascorbic acid were obtained from Wako Chemical Co. (Osaka, Japan). All other reagents were obtained from Sigma-Aldrich (St. Louis, MO). 

Cell Isolation

All experiments were conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Porcine eyes were enucleated just after death at a local abattoir and transferred to the laboratory within 30 minutes. A piece of ciliary body with epithelium was dissected and incubated in cell isolation medium for 30 to 40 minutes. The tissue was transferred to standard Ringer's solution containing 0.1% bovine serum albumin for 3 minutes and then incubated in standard Ringer's solution for 10 minutes, followed by gentle vortexing. Isolated cells were observed in a sample of the suspension by phase-contrast microscopy. The cell suspension was stored at 4°C for up to 24 hours before use. 

Cell Perfusion

Isolated ciliary epithelial cells were transferred to a Lucite perfusion chamber (RC-5/25; Warner Instrument, Hamden, CT) and settled for 10 to 20 minutes before perfusion. The perfusate flowed into the chamber under gravity at a flow rate of 0.7 mL/min by a perfusion system (BPS-4; ALA Scientific Instruments, Westbury, NY) and was continuously removed by suction. Fluid height was adjusted to give a chamber volume of approximately 0.5 mL and a complete solution exchange within 2 minutes. All experiments were conducted at room temperature (20–25°C). 

Electrophysiological Methods

Patch pipettes were pulled from borosilicate glass tubing (BF150-110-10; Sutter Instruments, San Rafael, CA) with a multistage, programable puller (P-97; Sutter Instruments). The pipette input resistance was between 1 and 3 MΩ. Under phase-contrast microscopy (Eclipse TE300; Nikon, Tokyo, Japan), a target cell was selected, and a pipette tip was pressed onto the cell membrane using a micromanipulator (MP-285; Sutter Instruments) to establish a gigaohm seal. Currents under voltage clamp were recorded by an amplifier system (EPC-8; Heka, Lambrecht/Pfalz, Germany). The built-in low pass filter was set to 3 kHz unless noted otherwise. Recordings were referenced to an Ag-AgCl electrode (EP-2; WPI, Sarasota, FL). The membrane capacitance was compensated by built-in circuits. The apparent membrane potential was corrected for the pipette tip potential (10 mV). Statistical data are presented as the mean ± SD. Data were fitted by a nonlinear, least-squares fitting program (IGOR Pro; Wavemetrics, Lake Oswego, OR). 

Results

Whole-Cell Currents

Pairs of PCE and NPCE cells with broad contact area (Fig. 1) were selected for the recording. The PCE–NPCE cell pairs are typically similar to that observed in a previous report. ¹⁷ In the initial and major phase of this study, we found M-type K⁺ current ^14,15 expressed in the PCE–NPCE cell pairs, as shown in Figure 2. Thereafter, we tried to record from isolated single ciliary epithelial cells, and found the identical M-type K⁺ current as was observed in single PE cells (n = 30) but never in single NPCE cells (n = 11). Therefore, data obtained from PCE–NPCE cell pairs (n = 97) and single PCE cells (n = 30) were pooled in this study. The 0-current potential (V ₀) was measured in the 0-current clamp mode after establishment of a perforated patch configuration: the mean V ₀ was −49 ± 13 mV in PCE–NPCE cell pairs (n = 97) and −52 ± 12 mV in single PCE cells (n = 30). The membrane capacitance and input resistance were 47 ± 34 pF and 489 ± 324 MΩ in PCE–NPCE cell pairs (n = 93), which were somewhat larger than those values of 37 ± 29 pF and 389 ± 277 MΩ measured in single PCE cells (n = 29). Student's t-tests showed no significant difference in these values between the two groups. 

Figure 1.

 

Photomicrograph of isolated porcine ciliary epithelial cells. A pair of PCE and NPCE cells freshly isolated from porcine eyes. In this experiment, a PCE cell membrane was clamped by a patch pipette. Bar, 50 μm.

Figure 2.

 

Whole-cell current in porcine ciliary epithelium and its dependence on holding potential. (A) Whole-cell currents in a porcine PCE–NPCE pair were elicited by 1-second voltage steps ranging from +10 to −110 mV in −10-mV increments from a holding potential (HP) of −10 mV. (B) In the same cell, currents were elicited by the same range of voltage steps from an HP of −70 mV. (C) I-V relationships are depicted with a HP of −10 and −70 mV. The steady state current amplitudes were averaged between 900 and 1000 ms in the whole-cell currents in (A).

Effect of Holding Potential

Representative families of whole-cell currents in a PCE–NPCE cell pair are depicted in Figure 2. The currents in Figure 2A were generated by a series of voltage pulses from a holding potential of −10 mV. Hyperpolarizing pulses produced time-dependent relaxations in current that reversed at approximately −80 mV. The current relaxations reflect the deactivation of a conductance. The following tail currents elicited by a depolarizing voltage step to the holding potential (−10 mV) from various hyperpolarized prepulses reflect activation of a conductance (see later). Figure 2B shows currents in the same cell evoked by a series of voltage steps to the same test potentials from a holding potential of −70 mV. Voltage pulses to potentials in the range between −110 and −70 mV produced time-independent currents with no rectification. On the other hand, depolarizing voltage pulses to potentials more positive than −60 mV produced a slowly activating current that did not inactivate during the 2-second depolarization. 

Figure 2C compares the steady state current–voltage relationships generated from holding potentials of −70 and −10 mV. The two I-V curves were nearly identical, indicating that the conductance underlying this outward current was fully open when the membrane was held at −10 mV and that it did not inactivate during the 10-second interval between pulses. Electrophysiological properties of this K⁺ current in PCE cells were quite similar to those of the M-type K⁺ current in RPE. ^13,14  

Selectivity

With 5 mM K⁺ in the bath, the current relaxations produced by hyperpolarizing pulses reversed between −90 and −80 mV (Figs. 2A, 3A). The reversal potential (V _r) was analyzed by interpolation from plots of the deactivation current amplitude versus membrane voltage (not shown). For four cells, V _r averaged −84 ± 2 mV, which is very close to the theoretical E_K (−83 mV). To evaluate sensitivities for K⁺, V _r was also examined at higher [K⁺]_o. Figure 3B shows an example of the K⁺ current activated in 140 mM K⁺ Ringer's solution, which reversed between −7 and +3 mV. In three cells superfused with 140 mM K⁺ Ringer's solution (E_K = 0 mV), V _r averaged −1 ± 3 mV. Overall, V _r was not significantly different from E_K, indicating that the time-dependent outward current is largely carried by K⁺. 

Figure 3.

 

Reversal potential of the time-dependent K⁺ current and its relationship to extracellular K⁺ concentration. (A) Superfused with standard Ringer's solution ([K⁺]_o = 5 mM), currents elicited by membrane voltage steps ranging from −100 to −70 mV from a holding potential (HP) of −10 mV. The polarity of the deactivating currents reversed between −90 and −80 mV. (B) Superfused with a high K⁺ solution ([K⁺]_o = 140 mM), K⁺ currents elicited by membrane voltage steps ranging from −37 to +33 mV from an HP of −107 mV. In 140 mM [K⁺]_o solution, the activating currents reversed between −7 and +3 mV.

Conductance–Voltage Relationship

The data in Figure 2B indicated that the M-type K⁺ current activated at −60 mV but at more negative potentials, the activation of a conductance could be underestimated because of its small size and slow kinetics. Therefore, we analyzed the tail current elicited by a depolarizing voltage step to a set potential from various hyperpolarized prepulses to determine the voltage activation range of K⁺ conductance. Figure 4A depicts a family of tail currents produced by voltage steps to −10 mV after 1-second prepulses to the potentials indicated. When the prepulse potentials were in the range of −110 to −90 mV, the tail current began from the same initial level and then rose exponentially to a new steady state. In contrast, prepulses to potentials between −80 and −50 mV produced smaller tail currents that began at a higher initial level but rose to the same steady state. Finally, when the voltage was stepped to −10 mV from −40 and +10 mV, the initial current level was roughly the same as the steady state current. The results of these experiments indicate that the M-type K⁺ current was completely deactivated at −90 mV and more negative potentials; partially activated between −80 and −30 mV; and completely activated at −20 mV and more positive potentials. 

Figure 4.

 

Voltage dependence of the K⁺ conductance. (A) Tail current activation at −10 mV after prepulses to the potentials indicated. Tail current amplitude is inversely proportional to the fraction of K⁺ channels activated at prepulse potential. (B) Conductance–voltage relationship. K⁺ conductance at each voltage is normalized to maximum conductance. Data represent mean ± SEM for 12 cells. The continuous curve is the least-squares fit of the data to a Boltzmann function: G(V)/G _max = 1/{1 + exp[(V _1/2 − V_m )/k_n ]}, where V _1/2 and k_n are −63.3 and 10.8 mV, respectively.

The conductance at each prepulse potential, G(V), can be expressed as  

 

where G_max is maximum conductance, and ΔG(−10) is the change in conductance produced by stepping the voltage to −10 mV, calculated by dividing the tail current amplitude by the driving force on K⁺ at −10 mV (73 mV). Similar results were obtained in 11 other cells, giving a mean value for G_max of 2.3 ± 0.9 nS (n = 12). Figure 4B summarizes the results of these experiments and plots normalized conductance as a function of prepulse potential. The continuous curve is the least-squares fit of the data to the Boltzmann function:  

 

with values for V_1/2 and k_n of −63.3 and 10.8 mV, respectively. 

Blocker Sensitivity

To further identify channel type, we examined the sensitivity of the M-type K ⁺ current in the PCE cells to several K⁺ channel blockers. M-type K⁺ current in the RPE ^14,15 is known to be sensitive to Ba²⁺ and Cs⁺ but not to TEA⁺. Figure 5A depicts families of whole-cell currents recorded in the absence (top) and the presence (bottom) of 50 mM external TEA⁺. TEA⁺ had little effect on the amplitude or kinetics of the time-dependent K⁺ current. The conductance–voltage relationship obtained from tail currents is depicted in Figure 5B, which shows that even in 50 mM TEA⁺ the maximum conductance was 15% smaller than it was in the absence of TEA⁺. The maximum K⁺ conductance changed to 101% ± 3% for four cells in the presence of 20 mM TEA⁺, and 91% ± 6% for other four cells in the presence of 50 mM TEA⁺. Ba²⁺, an effective blocker of the RPE M-type K⁺ current, also inhibited the K⁺ current in PCE. Figure 6A shows that 2 mM Ba²⁺ considerably inhibited the K⁺ current (by 82%) compared with the control. For three cells, the best fit of the data to a first-order equation indicates an apparent IC₅₀ of 0.38 mM (Fig. 6B). Electrophysiological properties and blocker sensitivities indicate that the swelling-activated K⁺ current in PCE belongs to a family of KCNQ channels. ¹⁸ Thus, we investigated the effect of linopirdine, a specific inhibitor of KCNQ channels. Figure 6C represents families of whole-cell currents recorded in the absence (top) and the presence (bottom) of linopirdine. The K⁺ current component (tail currents) was almost completely blocked by 10 μM linopirdine. In four cells, linopirdine (10–25 μM) depolarized the resting potentials to −41 ± 8 mV from the control value of −59 ± 4 mV. Figure 6D summarizes the results of experiments in which the effects of various concentrations of linopirdine were tested on the K⁺ current, which indicated an apparent IC₅₀ of 0.54 μM. 

Figure 5.

 

Effect of TEA on the K⁺ current. (A) Families of whole-cell currents recorded in the absence (top) and the presence (bottom) of 50 mM TEA. Currents were elicited by membrane voltage steps ranging from +10 to −130 mV from a holding potential of −10 mV. (B) Effect of TEA on steady state conductance-voltage relationship. The K⁺ conductance in the absence (●) and the presence (○) of 50 mM TEA was calculated from tail currents in (A).

Figure 6.

 

Inhibitory effects of Ba²⁺ and linopirdine on the K⁺ channels. (A) Whole-cell currents elicited by voltage steps ranging from +10 to −130 mV were obtained in the absence and the presence of 2 mM Ba²⁺. (B) Dose–response curve. Relative current amplitude at each concentration was measured and summarized in 3 cells. Smooth curves represent fitting the data to the first-order equation: I _Inhibitor/I _control = 1 − {[Inhibitor]/([Inhibitor] + IC₅₀)}, with an IC₅₀ of 0.38 mM. All error bars represent SEM. (C) Whole-cell currents elicited by voltage steps ranging from +10 to −110 mV were obtained in the absence and the presence of 10 μM linopirdine. (D) Dose–response curve. Each data point depicts measurements in three cells. Smooth curves represent fitting the data to the same equation above, where the IC₅₀ was 0.54 μM. All error bars represent SEM.

Activation by Hypotonicity

Modulation of the M-type K⁺ current by cell swelling was evaluated. Cell perfusion with a 20% diluted Ringer's solution (239 ± 2 mOsmol/L) induced a significant increase in the K⁺ currents in the PCE–NPCE cell pairs, as shown in Figure 7. This current augmentation by hypotonicity was reversible when the perfusate returned to the standard osmolarity (Fig. 7B). In 14 PCE–NPCE cell pairs, exposure to 20% hypotonic solution increased the maximum conductance to 250% ± 86% of the control value. Similar augmentations of M-type K⁺ current were observed also in single PCE cells (n = 3, ratio; 184% ± 82%). The reversal potential of the hypotonicity-induced current component averaged −78 ± 8 mV (n = 7), a value more positive than the theoretical E_K (−83 mV), which indicates that chloride conductance was also activated somewhat and thereby the reversal potential shifted to the direction of theoretical E_Cl (−33 mV). The current component augmented by exposure to 20% hypotonic solution as was blocked by linopirdine (20 μM, n =3), but was essentially insensitive to external TEA⁺ (40 mM, n = 3; Fig. 7C). 

Figure 7.

 

Effect of hypotonicity on the K⁺ current. (A) Families of whole-cell currents recorded when a cell was superfused with standard Ringer's solution (302 ± 6 mOsmol/L, left) and 20% diluted hypotonic solution (239 ± 2 mOsmol/L, right). Currents were elicited by membrane voltage steps ranging from +10 to −120 mV from a holding potential of −10 mV. (B) Time course of the M-type K⁺ current. The maximum amplitude of the tail currents was monitored in the same cell as shown in (A). (C) Effects of high TEA and linopirdine in hypotonic condition. In another cell, families of whole-cell current were recorded in standard Ringer's solution, 20% diluted hypotonic solution, hypotonic (20% dilution) 40 mM TEA solution, and 20 μM linopirdine in hypotonic solution. Currents were elicited by membrane voltage steps ranging from +10 to −110 mV from a holding potential of −10 mV.

Discussion

In the present study, we identified a noninactivating, outward K⁺ current in porcine PCE cells that is likely to be different from the previously reported K⁺ channels in the ciliary epithelium. To date, two types of outward K⁺ currents have been identified in the ciliary epithelium. Voltage-activated, delayed rectifier K⁺ currents were reported in studies of cultured rabbit PCE cells, ⁶ cultured NPCE cells, ⁸ and fresh bovine PCE cells. ⁷ All the currents were inhibited by TEA⁺. Another type of current is the Ca-activated K⁺ current reported in rabbit PCE cells, which was also sensitive to TEA⁺. ⁹ Carbachol-induced transient outward K⁺ currents described in fresh bovine PCE cells was also Ca²⁺-sensitive, ¹⁰ but their sensitivity to channel blockers is unknown. 

K⁺ current in the present study closely resembled the M-type K⁺ current expressed in RPE ^14,15 in terms of activation kinetics, voltage dependence, and blocker sensitivities. The common properties of these currents are that they are noninactivating and activate with a threshold of −60 mV or more negative voltages, and, thus, they contribute to the resting potential. In addition, they are inhibited by Ba²⁺ but are relatively insensitive to TEA⁺. The M-type K⁺ current of RPE was so named because of its resemblance to the M-current, a K⁺ current identified in certain neurons and other excitable cells. The underlying channel of the M-current is encoded by KCNQ genes. ^18,19 The KCNQ/Kv7 channel family is composed of a group of five voltage-gated K⁺ channel subunits (Kv7.1–Kv7.5) with six transmembrane domain–spanning proteins encoded by the KCNQ genes (KCNQ1–5). The channels are formed by homotetramers of the same KCNQ α-subunits or co-assembly of different KCNQ proteins. Linopirdine, a specific blocker of KCNQ channels, inhibited the K⁺ channel in this study suggesting that the K⁺ current in PCE cells belongs to KCNQ channel members. Sensitivities of KCNQ/KV7 channels to the classic K⁺ channel blocker TEA vary by member: KCNQ1 and -2 are relatively sensitive to TEA, with an IC₅₀ ranging from 0.3 to 5.0 mM, whereas KCNQ3, -4, and -5 are substantially insensitive to TEA, with an IC₅₀ of 30 mM or more. ^18,20 The K⁺ channel in this study was little affected, even by treatment with 50 mM TEA; hence, it is likely to belong to the latter members. 

One of the physiological roles of this K⁺ channel in PCE cells is likely to consist of a major K⁺ conductance of the apical and/or basolateral membranes, which is supported by the fact that the channel is open near the resting potential and that linopirdine, a specific blocker of this K⁺ channel depolarized the 0-current potential of the PCE cells (−59 ± 4 to −41 ± 8 mV). Another role of this K⁺ channel in PCE cells may be to mediate cell volume and water transport. As with the K⁺ channel in this study, some KCNQ channels have been reported to be activated by cell swelling. For instance, the KCNQ1 channel, which is expressed in secretory epithelia, like the distal colon or airway, is regulated by small changes in cell volume and participates in the secretion of salt and water. ²¹ In addition, recent evidence suggests that KCNQ3 and -5 also could mediate basal anion secretion from an airway epithelial cell line ⁵ and regulate Na⁺ transport across lung epithelial cells. ²²  

In conclusion, we identified a swelling-activated K⁺ channel in freshly dissociated porcine PCE cells. Electrophysiological and pharmacological properties of this channel indicate that it is a member of the KCNQ/Kv7 channel family. This K⁺ channel is active near resting potentials and may contribute to the regulation of cell volume and water transport via the ciliary epithelia. 

Footnotes

 Supported by Grants-in-Aid 16591742(MT) and 21791674(NS) from the Japanese Ministry of Education, Culture, Sports, Science and Technology.

Footnotes

 Disclosure: M. Takahira, None; M. Sakurai, None; N. Sakurada, None; K. Sugiyama, None