The human ether-a-go-go-related gene (hERG) encodes a K⁺ channel that conducts the rapidly activating delayed rectifier K⁺ current (I_Kr) (Sanguinetti et al., 1995; Trudeau et al., 1995). Reduction in I_Kr can cause long QT syndrome (LQTS), a cardiac repolarization disorder that can lead to life-threatening arrhythmias, torsades de pointes, and sudden cardiac death (Keating and Sanguinetti, 2001). Whereas direct blockade of hERG channels by various compounds represents a common mechanism for drug-induced LQTS (Sanguinetti and Tristani-Firouzi, 2006), recent evidence indicates that drug-disrupted hERG trafficking represents another mechanism for drug-induced LQTS (Ficker et al., 2004; Cordes et al., 2005; Kuryshev et al., 2005; Rajamani et al., 2006). Probucol is a cholesterol-lowering drug that has been known for many years to cause LQTS and torsades de pointes arrhythmia in humans and experimental animals (Elharrar et al., 1979; McCaughan, 1982; Jones et al., 1984). In a previous report studying wild-type and M124T mutant hERG channels expressed in Xenopus oocyte, Hayashi et al. (2004) showed that probucol (30 μM) did not affect the hERG current amplitude during depolarization steps but decreased the tail currents as a result of a ∼10-mV shift of activation curve to the depolarized direction. It is unknown whether probucol causes similar effects on hERG channels expressed in mammalian cell lines.

Presently, much of the available data of I_Kr are obtained in hERG channels expressed in mammalian cell lines or Xenopus oocyte. Because the pore-forming subunits of K⁺ channels incorporate modulatory (β) subunits (Abbott et al., 1999), kinase-anchoring proteins (Gong et al., 1999), cytoskeletal elements, and other proteins, it is necessary to directly study native I_Kr. However, because of difficulties such as isolating I_Kr from other cardiac K⁺ currents, there has been no reported study showing native I_Kr trafficking at the protein level. In the present study, we identified I_Kr in neonatal rat ventricular myocytes, which can be recorded at a sufficiently robust level using Cs⁺ as the charge carrier in whole-cell clamp recordings. Using this native I_Kr model, we found that clinically relevant concentrations of probucol reduce I_Kr and prolong the cardiac action potential duration via reducing I_Kr membrane expression but not via channel blockade. Our study showed that primary culture of neonatal rat cardiomyocytes represents an effective model system for studying cardiac I_Kr trafficking.

Materials and Methods

Neonatal Rat Ventricular Myocyte Isolation. Experimental protocols used for animal studies were approved by the University of Manitoba Animal Care Committee. Single ventricular myocytes were isolated from 1- to 2-day-old Sprague-Dawley rats of either sex by enzymatic dissociation as described previously (Baetz et al., 2005). Cells were cultured in Dulbecco's modified Eagle's medium/Ham's F-12 medium (Invitrogen, Burlington, OH) with 10% fetal bovine serum. Cardiomyocytes grown on glass coverslips for the electrophysiology study had a “spindle” morphology with a mean capacitance of 13.8 ± 1.3 pF (n = 33).

Molecular Biology. The hERG-human embryonic kidney (HEK) cells (HEK-293 cells stably expressing hERG channels) were a gift from Dr. Craig January (University of Wisconsin-Madison, Madison, WI) (Zhou et al., 1998b). In this cell line, hERG cDNA (Trudeau et al., 1995) was subcloned into BamHI/EcoRI sites of the pcDNA3 vector (Invitrogen). hERG cDNA in pcDNA3 was obtained from Dr. Gail Robertson (University of Wisconsin-Madison) (Trudeau et al., 1995). For immunofluorescence staining of the cell surface hERG channels, a hemagglutinin (HA)-epitope tag of the sequence ⁴³⁶TEEGPPATNSEHYPYDVPDYAVTFEECGY (boldface, insertion; italicized, HA epitope) was inserted into the extracellularly located S1-S2 loop of hERG channels to generate hERG-HAex via polymerase chain reaction using overlap extension method as described previously (Zhang, 2006). The hERG-HAex was transfected to HEK-293 cells, and a stable hERG-HAex cell line (hERG-HAex-HEK) was created using G418. The insertion of HA did not change the electrophysiological and trafficking properties of hERG channels (data not shown), consistent with the results from Ficker et al. (2003).

Patch-Clamp Recording Method. Whole-cell patch-clamp method was used (Hamill et al., 1981). The compositions of pipette and bath solutions for recording various currents are summarized in Table 1. Probucol was dissolved in ethanol to make 10 to 30 mM stock solutions. Procedures of whole-cell patch-clamp method and Cs⁺-carried I_Kr recording were performed as described previously (Zhang, 2006). Patch-clamp experiments were performed at room temperature (23 ± 1°C).

Western Blot Analysis. Membrane proteins from hERG-HEK cells were isolated using Mem-PER Eukaryotic Membrane Protein Extraction Reagent Kit (Pierce Biotechnology, Rockford, IL). Membrane protein (10 μg/lane) was separated on 7% SDS-polyacrylamide electrophoresis gels and transferred onto nitrocellulose membranes. The membranes were blocked using Western Breeze Blocking Reagent (Invitrogen) and incubated with primary goat anti-hERG antibody (C-20, Santa Cruz Biotechnology, Santa Cruz, CA) and secondary anti-goat Western Breeze Chromogenic Detection Kit (Invitrogen). For cleavage of cell surface proteins, cells were washed with phosphate-buffered saline (PBS) and treated with 200 μg/ml proteinase K (Sigma-Aldrich, St. Louis, MO) in a physiological buffer (10 mM HEPES, 150 mM NaCl, and 2 mM CaCl₂, pH 7.4) at 37°C. The reaction was terminated by adding ice-cold PBS containing 6 mM phenylmethylsulfonyl fluoride and 25 mM EDTA, and membrane protein was then extracted for Western blot analysis.

To extract membrane proteins from neonatal rat cardiac myocytes, cells from 100-mm plates were rinsed with ice-cold PBS and scraped off into a 1-ml solution containing 200 mM NaCl, 33 mM NaF, 10 mM EDTA, 50 mM HEPES (pH 7.4 with NaOH), plus a protease inhibitor mixture. The cells were homogenized and spun at 500g for 10 min. The membrane fractions were pelleted from the low-speed supernatants by centrifugation at 100,000 rpm for 1 h at 4°C and resuspended in 50 mM Tris-HCl, 15 mM mercaptoethanol, and 1% SDS. The membrane proteins (50 μg/sample) were boiled in sample buffer and electrophoresed on a 7% polyacrylamide SDS gel. The membrane proteins were then electrophoretically transferred onto nitrocellulose membrane using a trans-blot system (Bio-Rad, Hercules, CA). After transfer, the filters were blocked with 5% nonfat dry milk and 0.1% Tween 20 in Tris-buffered saline (TBS) for 1 h. The filters were then incubated with goat polyclonal anti-hERG (C-20) antibody at a 1:200 dilution at 4°C overnight. The filters were then washed with TBS/Tween 20 (TBST) solution and incubated with horseradish peroxidase-conjugated donkey anti-goat immunoglobulin diluted 1:60,000 in TBST for 1 h at room temperature. After washing with TBST, bound antibodies were detected with an enhanced chemiluminescence detection kit.

Isolation of Cell Surface Protein with Biotinylating Reagent. A Cell Surface Protein Isolation Kit (Pierce) was used to study the effects of probucol on the cell surface hERG expression. hERG-HEK cells were prepared in 100-mm cell culture plates at 90% confluence. The cells treated with vehicle control (0.3% ethanol) or 100 μM probucol for 48 h were washed twice with ice-cold PBS and labeled with 10 ml of membrane-impermeant biotinylating reagent Sulfo-NHS-SS-Biotin (Pierce) for 30 min at 4°C. The quenching solution (0.5 ml) was then added to quench the reaction. Cells were then lysed with 0.5 ml of lysis buffer with a protease inhibitor mixture. After centrifugation at 10,000g for 2 min at 4°C, the cell lysate was precipitated with Immobilized NeutrAvidin Gel (agarose beads) (Pierce). The bound proteins were released by incubating the resin with SDS-polyacrylamide gel electrophoresis sample buffer (62.5 mM Tris-HCl, pH 6.8, 1% SDS, and 10% glycerol) containing 50 mM dithiothreitol. The biotinylated cell surface protein was subjected to 7% SDS-polyacrylamide gel electrophoresis, analyzed using primary goat anti-hERG antibody (C-20, Santa Cruz Biotechnology) and horseradish peroxidase-conjugated donkey anti-goat secondary antibody, and detected with an enhanced chemiluminescence detection kit.

hERG Small Inhibitory RNA Transfection. To inhibit ether-a-go-go-related gene (ERG) mRNA, cultured neonatal rat cardiomyocytes and hERG-HEK cells were transfected with hERG small inhibitory RNA (siRNA) (Santa Cruz Biotechnology) or mouse ERG siRNA (of which one strand targets identical stretches in nucleotide sequence of rat ERG, Santa Cruz Biotechnology) using Lipofectamine 2000 (Invitrogen, Burlington, ON, Canada) according to the protocols recommended by the suppliers. Scrambled siRNA (Santa Cruz Biotechnology) was used as a negative control. Experiments were performed 48 h after transfection. For electrophysiology studies, pIRES2-EGFP (Clontech, Palo Alto, CA) was included in the transfection reagent during ERG siRNA transfection, and fluorescent-positive cells were selected for I_Kr or hERG current recordings.

Immunocytochemistry. For immunofluorescent studies, hERG HAex-HEK cells were plated on coverslips for growth under control conditions and in the presence of probucol. Cells were fixed under a nonpermeabilized condition with 4% paraformaldehyde (Sigma-Aldrich) for 10 min at room temperature. After washing three times with PBS, cells were blocked with PBS containing 5% bovine serum albumin and 2% skim milk. The cells were immunostained with a rabbit anti-HA primary antibody (Anti-HA, 1:500, Sigma-Aldrich) and a green-fluorescent Alexa Fluor 488-conjugated donkey anti-rabbit IgG secondary antibody (1:250, Invitrogen). Cells were visualized using a Nikon (Mississauga, ON, Canada) TE2000-U research microscope.

Data are expressed as the mean ± S.E.M. A one-way analysis of variance or Student's t test was used to determine the significance of differences between control and test groups. A p value of 0.05 or less was considered significant.

Results

Identification of I_Kr in Neonatal Rat Ventricular Myocytes.Figure 1Aa illustrates K⁺ currents recorded from a neonatal rat ventricular myocyte. The pipette solution contained 135 mM K⁺, and the bath solution contained 5 mM K⁺ (Table 1). From a holding potential of –60 mV, the cell was depolarized to voltages between –50 and 50 mV in 10-mV increments. On depolarizing steps, 68% (26 of 38) of cells displayed the transient outward current (I_to). Notably, 95% (36 of 38) of cells displayed the delayed outward K⁺ current that was accompanied by the slow-decay tail current on voltage returns to –50 mV. Both delayed outward currents and tail currents were completely abolished by the methanesulfonanilide compound E-4031 (1 μM), a specific blocker of I_Kr (Fig. 1Ab). The E-4031-sensitive currents were obtained by subtracting the whole-cell current after application of 1 μM E-4031 from that before application of E-4031 in the same cell. Figure 1Ac shows the E-4031-sensitive currents. Figure 1B shows the averaged current-voltage (I-V) relationships of the E-4031-sensitive currents obtained from four cells. The currents at the end of depolarizing steps displayed the inward rectification that is characteristic to I_Kr (Fig. 1B, ▪). The tail currents were plotted against the depolarizing voltages, and the data were fitted to the Boltzmann equation (Fig. 1B, •). The half-activation voltage (V_1/2) was –6.2 ± 1.7 mV, and the slope factor was 6.9 ± 1.5 mV (n = 4 cells).

E-4031-sensitive I_Kr is small, and its recording represents a tedious task. Moreover, any alterations of K⁺ currents during recordings before and after E-4031 will make the subtraction inaccurate. To address this difficulty, we recorded the pure I_Kr in neonatal rat ventricular myocytes using isotonic Cs⁺ solutions (135 mM Cs_i⁺/135 mM Cs_o⁺) (Table 1). We recently showed that hERG and I_Kr channels display unique Cs⁺ permeability (Zhang, 2006). Because Cs⁺ blocks most cardiac K⁺ channels, Cs⁺-carried I_Kr represents a simple and reliable way to directly record I_Kr (Zhang, 2006). Figure 1C shows a family of Cs⁺ currents obtained from a single cardiomyocyte. From a holding potential of –80 mV, depolarizations in 10-mV increments to voltages between –70 and +70 mV for 1 s were applied to evoke currents. Depolarizing steps to voltages more positive than 0 mV induced outward currents that inactivated in a voltage-dependent manner. The following tail currents at –80 mV displayed an initial rising phase, which is usually described as a “hook,” reflecting the rapid recovery of inactivated channels to the open state before deactivation, and is unique to I_Kr. Figure 1D shows the I-V relationships of peak currents (▴) and currents at the end of 1-s depolarizing steps (▪). Figure 1E shows the tail current activation curve. The V_1/2 and slope factor (k) were –41.7 ± 3.4 and 5.8 ± 0.3 mV, respectively (n = 8 cells). The relatively negative V_1/2 was caused by the absence of Ca²⁺ in the bath solution (Zhang, 2006). The average Cs⁺ tail current density measured at –80 mV following full channel activation was 17.9 ± 3.3 pA/pF (n = 8 cells).

To further show that the Cs⁺ current recorded from neonatal rat cardiomyocytes indeed represents the Cs⁺-carried I_Kr, the voltage-dependent inactivation and recovery from inactivation, which are unique to I_Kr, were analyzed. Figure 2A shows the voltage dependence of inactivation of Cs⁺-carried I_Kr. The membrane was initially depolarized to +60 mV for 500 ms to inactivate the channels. A 20-ms repolarizing step to –100 mV was applied to recover inactivated channels to the open state. Before deactivation, the membrane was depolarized to various voltages to induce channel inactivation (current decay), which was fitted to a single exponential function to obtain the inactivation time constant (τ_inact). Figure 2B shows the recovery from inactivation of the Cs⁺-carried I_Kr. The channels were activated and inactivated by a 500-ms depolarizing pulse to +60 mV. Repolarizing voltages between –10 and –120 mV were applied to record tail currents. The time constants of recovery from inactivation (τ_rec) were obtained by fitting the rising phase of the tail currents to a single exponential function at different voltages (Sanguinetti et al., 1995). Values of τ_inact (▾, n = 6) and τ_rec (▴, n = 6) are summarized and plotted against the membrane voltages (Fig. 2C). It has been known that extracellular Cs⁺ slows hERG and I_Kr inactivation (Zhang, 2006). Consistent with the Cs⁺ current in neonatal rat ventricular myocytes being Cs⁺-carried I_Kr, the Cs⁺ current inactivation was slowed by Cs⁺ (Fig. 2, D–F). As well, the Cs⁺ current was entirely sensitive to hERG/I_Kr blocker E-4031, with IC₅₀ and Hill coefficient of 2.0 ± 0.4 and 1.2 μM, respectively (Fig. 2, G–I, n = 4).

Western blot analysis was used to identify I_Kr proteins in neonatal ventricular myocytes. As shown in Fig. 3A, the hERG C-terminal antibody (HERG C-20; Santa Cruz Biotechnology) consistently identified four bands at sizes of approximately 150, 130, 95, and 85 kDa (n = 5 different myocyte isolations). The two higher molecular mass bands (150 and 130 kDa) are similar in size with mature glycosylated and immature core-glycosylated rat ERG1a (Jones et al., 2004). The two lower molecular mass bands are consistent in size with mature glycosylated and core-glycosylated rat ERG1b (Jones et al., 2004). To confirm that the 150- and 95-kDa bands represent the plasma membrane forms of ERG, we cleaved cell surface proteins by treating neonatal cardiomyocytes with proteinase K. As shown in Fig. 3A, proteinase K treatment significantly reduced the 150- and 95-kDa bands, and this was accompanied by the appearance of an additional 60- to 70-kDa band. This treatment did not affect the 130- or 85-kDa bands of the ERG proteins (n = 5). Figure 3B shows Western blot of hERG proteins extracted from hERG-HEK cells (hERG1a). As reported previously (Zhou et al., 1998b; Kuryshev et al., 2005), hERG displayed two bands with molecular masses of 155 and 135 kDa (Fig. 3B), representing the mature fully glycosylated membrane form (155 kDa) and the immature core-glycosylated endoplasmic reticulum (ER) form (135 kDa) (Zhou et al., 1998a,b; Ficker et al., 2004; Kuryshev et al., 2005). The 155-kDa hERG protein is localized on the plasma membrane because it was cleaved by proteinase K treatment (Fig. 3B, n = 3).

The siRNA of hERG were used to interfere with ERG mRNA and hERG/I_Kr expression. The siRNA of hERG (Santa Cruz Biotechnology; mRNA accession no. NM_000238) has a sense strand sequence CCAUCAAGGACAAGUAUGU that targets the nucleotides of S5 side of the P-loop of hERG (between 1817 and 1835). This sequence also targets the nucleotides between 1823 and 1841 of rat ERG with one nucleotide difference. Scrambled siRNA (Santa Cruz Biotechnology) was used as control. Transfection with hERG siRNA reduced four bands of rat ERG proteins with a similar extent, with 130-kDa band being reduced by 75 ± 4% in neonatal rat cardiomyocytes (Fig. 3C, n = 4). Transfection with hERG siRNA reduced both hERG 135- and 155-kDa bands, with 135-kDa band being reduced by 83 ± 7% in hERG-HEK cells (Fig. 3D, n = 3). For electrophysiology studies, hERG siRNA or scrambled siRNA was transfected along with green fluorescent protein (GFP) (pIRES2-EGFP; Clontech). The I_Kr Cs⁺ current in neonatal cardiomyocytes and the hERG K⁺ current in hERG-HEK cells were recorded in GFP-positive cells. Compared with I_Kr or hERG currents recorded in cells transfected with control siRNA, I_Kr was reduced by 81 ± 9% (n = 10 for control siRNA; n = 16 for hERG siRNA; p < 0.01), and hERG current was reduced by 87 ± 2% (n = 8 for control siRNA; n = 9 for hERG siRNA; p < 0.01) in cells transfected with hERG siRNA (Fig. 3, E and F). Furthermore, in neonatal rat cardiomyocytes, transfection of a mixture of three mouse ERG siRNA, which are highly homologous to rat ERG (strand A 100% targeting region 987-1005 of rat nucleotide sequence, accession no. NM_053949.1; strand B targeting 1469–1487 with one nucleotide difference; strand C targeting 2635–2653 with one nucleotide difference), reduced I_Kr by 83 ± 6% (n = 8 for control siRNA; n = 9 for mouse ERG siRNA; p < 0.01).

Probucol Reduces ERG Expression. Our results indicate that Cs⁺-carried I_Kr in neonatal rat cardiomyocytes can be recorded at a level that is large enough and sufficiently robust to evaluate I_Kr alterations. Next, we investigated the mechanisms of probucol-induced LQTS by studying the effects of probucol on I_Kr Cs⁺ current in neonatal rat ventricular myocytes. Probucol is a cholesterol-lowering drug that causes LQTS in humans (Elharrar et al., 1979; McCaughan, 1982; Jones et al., 1984; Hayashi et al., 2004). We found that acute application of 100 μM probucol had no effect on I_Kr (Fig. 4A; n = 5). Figure 4A shows Cs⁺-carried I_Kr recorded from a cardiomyocyte before and after acute application of 100 μM probucol. The tail current-activation voltage relationships were shown at the bottom of Fig. 4A. The V_1/2 and slope factor were –38.5 ± 1.6 and 6.7 ± 0.5 mV, respectively, in the absence of probucol (n = 5 cells). They were –39.5 ± 1.7 and 6.6 ± 0.6 mV in the presence of probucol (n = 5 cells, p > 0.05). Thus, probucol had no acute effect on I_Kr in neonatal cardiomyocytes. However, chronic treatment of cardiomyocytes with probucol substantially reduced I_Kr. Figure 4B shows Cs⁺-carried I_Kr from cardiomyocytes cultured in the absence (0.3% ethanol vehicle) or presence of 100 μM probucol for 48 h. The tail current-activation voltage relationships were summarized from five cells in the absence of probucol and from seven cells in the presence of 30 μM probucol. The V_1/2 and slope factor were –42.2 ± 3.3 and 6.1 ± 0.7 mV, respectively, in control. They were –39.1 ± 2.9 and 6.6 ± 0.5 mV in probucol-treated cells (p > 0.05). Thus, whereas chronic application of probucol reduced I_Kr amplitude, it did not affect the voltage dependence of activation of I_Kr. Probucol-induced I_Kr reduction was concentration-dependent. The bottom of Fig. 4B shows the summarized I_Kr tail current amplitudes from myocytes treated with various concentrations of probucol for 48 h (n = 5–12 for each concentration). The IC₅₀ and Hill coefficient were 20.6 ± 1.6 μM and 0.9 ± 0.1, respectively.

Figure 4, C and D, shows the effects of probucol on hERG K⁺ currents. Acute bath application of 100 μM probucol affected neither the hERG current amplitude nor the voltage dependence of the activation curve of hERG channels (Fig. 4C). The V_1/2 and slope factor were –5.1 ± 2.3 and 7.2 ± 0.8 mV, respectively, in control. They were –7.2 ± 2.7 and 7.3 ± 0.7 mV in the presence of probucol (n = 5 cells, p > 0.05). To test whether probucol blocks hERG channels from the internal side of the membrane, we included 100 μM probucol in the pipette solution. We found that the hERG current with 100 μM probucol in the pipette solution was not different from the currents without probucol in the pipette solution (n = 4, data not shown). Thus, probucol does not directly block hERG channels. However, inclusion of probucol in the cell culture medium reduced the hERG current. Figure 4D shows hERG currents from hERG-HEK cells cultured in the control medium (containing 0.3% ethanol vehicle) and in the presence of 30 μM probucol for 48 h. The hERG activation curves under control and probucol-treated conditions are also shown in Fig. 4D. Probucol did not change the voltage dependence of the hERG channel activation (V_1/2 =–7.5 ± 1.9 mV, k = 6.6 ± 0.2 mV, n = 14 for control; V_1/2 =–3.4 ± 1.7 mV, k = 7.8 ± 0.4 mV, n = 12 for 30 μM probucol, p > 0.05) but significantly reduced the hERG current. To study concentration dependence of probucol effects on hERG current amplitude, tail currents from each probucol-treated cell were normalized to the mean control value (n = 14 cells). The summarized relative tail currents were plotted against probucol concentrations and fitted to Hill equation (n = 9–22 cells at each concentration). The IC₅₀ and Hill coefficient were 10.6 μM and 1.0, respectively (Fig. 4D, bottom).

The chronic nature of probucol-induced hERG and I_Kr reduction suggests that probucol may decrease ERG membrane expression. To visualize the surface hERG expression, we performed immunofluorescence staining of cell surface hERG-HAex channels using an anti-HA antibody that recognizes the HA-epitope located in the extracellular S1-S2 linker of hERG channels. Because we did not permeabilize hERG-HAex-HEK cells, only the extracellularly exposed hERG was stained. As shown in Fig. 5A, the cell surface staining of hERG is visible throughout the control cell. Treatment of hERG-HAex-HEK cells with 100 μM probucol for 48 h significantly reduced the cell surface hERG staining. Figure 5B shows Western blots of hERG proteins extracted from wild-type hERG-HEK cells cultured in the absence (0.3% ethanol) and presence of 100 μM probucol for 48 h. Probucol essentially eliminated the mature plasma membrane form of hERG channels. Figure 5B, right shows the densities of the hERG 155- and 135-kDa bands after probucol treatment relative to their respective control values. Probucol treatment reduced the density of the 155-kDa band by 88 ± 7% (n = 10, p < 0.01) and increased the density of the 135-kDa band by 13 ± 10% (n = 10, p > 0.05). To confirm the effects of probucol on the surface membrane form of hERG channels, we isolated surface membrane protein using the biotinylation method (Cell Surface Protein Isolation Kit, no. 89881, Pierce). The isolated membrane protein from hERG-HEK cells under control conditions (0.3% ethanol) or treated with 100 μM probucol were analyzed using Western blot with anti-hERG antibody (C-20, Santa Cruz Biotechnology) (Fig. 5C). Probucol treatment reduced surface membrane hERG by 84 ± 9% (n = 4). To examine the effects of probucol on I_Kr expression, Western blots of ERG proteins extracted from neonatal cardiomyocytes cultured in the absence (0.3% ethanol) or presence of 100 μM probucol for 48 h were compared in Fig. 5D. Probucol treatment significantly reduced the 150- and 95-kDa bands and did not affect the 130- and 85-kDa bands (n = 4). Thus, probucol reduced cell surface form of ERG proteins.

To determine the specificity of the probucol-induced hERG/I_Kr current reduction, the effects of probucol on Na⁺ current (I_Na), transient outward K⁺ current (I_to), and inward rectifier K⁺ current (I_K1) in neonatal rat cardiomyocytes were examined. Cardiomyocytes were treated with 100 μM probucol for 48 h, and various currents were recorded using specific voltage protocols shown above the corresponding current traces in Fig. 6. The solutions used are summarized in Table 1. Probucol significantly reduced I_Kr without affecting I_Na, I_to, or I_K1.

Probucol Treatment Prolongs Ventricular Action Potential. The effects of probucol on action potentials recorded in neonatal rat ventricular myocytes were examined. Under control conditions (0.3% ethanol) after 24-h culture, 41% of the cardiomyocytes were quiescent after achieving whole-cell configuration (n = 32 cells from six isolations; 13 of 32 cells were quiescent). When 30 μM probucol was added to the cell culture medium, 58% of the ventricular myocytes were quiescent after 24-h culture (p < 0.01, n = 19 cells from six isolations; 11 of 19 cells were quiescent). Action potentials from the quiescent cells were used for analysis. Consistent with previous reports (Kang et al., 1995; Gaughan et al., 1998), the action potential in control neonatal rat cardiomyocytes displayed a much more pronounced plateau phase than those in adult rat cardiomyocytes (Fig. 7A). Probucol treatment significantly prolonged action potential duration (Fig. 7B). The action potential durations at 90% repolarization (APD₉₀) under control and probucol-treated conditions are summarized in Fig. 7C. To confirm the role of I_Kr in action potential duration, the effect of E-4031, a specific I_Kr blocker, was examined. The action potentials from neonatal cardiomyocytes in the presence of 100 nM E-4031 were compared with those in control. E-4031 (100 nM) significantly prolonged action potential duration (Fig. 7, D–F; n = 12).

Discussion

We identified I_Kr in cultured neonatal rat ventricular myocytes. ERG mRNA and protein have been found in the adult rat heart (Wymore et al., 1997; Jones et al., 2004). The existence of I_Kr in neonatal rat cardiomyocytes has not been reported. In fact, the systematic electrophysiology data on I_Kr in adult rats are lacking possibly because of difficulties in isolating I_Kr from other coexisting K⁺ currents such as I_to. We have found that Cs⁺ uniquely permeates hERG, and we have used Cs⁺ permeation to successfully isolate I_Kr in rabbit ventricular myocytes (Zhang, 2006). The clone of rat ERG is 96% identical to hERG at the amino acid level (Wymore et al., 1997). In the present study, we have recorded robust Cs⁺-carried I_Kr in neonatal rat ventricular myocytes. Because Cs⁺ slows hERG/I_Kr inactivation (Zhang, 2006), there are differences in current kinetics between K⁺- and Cs⁺-carried I_Kr (Zhang, 2006). As well, the IC₅₀ for E-4031 to block Cs⁺-carried I_Kr is higher than that to block K⁺-carried I_Kr (Figs. 1 and 2) (Zhang, 2006). Despite these differences, the Cs⁺-carried I_Kr recording represents a simple and reliable way to study I_Kr density and greatly facilitates studies of drug-induced alterations of I_Kr expression. Presently, models for studying native I_Kr trafficking at the protein level are not available. hERG turnover time is approximately 11 h in hERG-HEK cells (Ficker et al., 2003), and I_Kr turnover time in adult cardiomyocytes is likely to be much slower. Because neonatal cardiomyocytes have a more vigorous metabolism than adult cardiomyocytes, identification of I_Kr in neonatal cardiomyocyte provides us with a very useful way to analyze the native I_Kr trafficking. Significantly, using this model, we discovered that probucol reduces functional I_Kr expression. Previously, pentamidine was reported to disrupt hERG trafficking and cause LQTS (Cordes et al., 2005; Kuryshev et al., 2005). We found that like probucol, pentamidine treatment (10 μM, 48 h) significantly reduced the 150- and 95-kDa bands and did not affect the 130- and 85-kDa bands of ERG of neonatal rat cardiomyocytes (n = 4, data not shown).

Probucol is a cholesterol-lowering drug that has been found to cause long QT syndrome and torsades de pointes arrhythmia in patients and sudden cardiac death in experimental animals (Elharrar et al., 1979; McCaughan, 1982; Browne et al., 1984; Dujovne et al., 1984; Matsuhashi et al., 1989; Tamura et al., 1994; Reinoehl et al., 1996; Hayashi et al., 2004). Our data indicate that chronic probucol exposure reduces the hERG current with an IC₅₀ of 10.6 μM and reduces native I_Kr with an IC₅₀ of 20.6 μM. We chose 48-h treatment of cells with probucol because the turnover of the hERG channel occurs at a rather slow rate (approximately 11 h) (Ficker et al., 2003). The recommended daily dosage of probucol for human adults is 1 g. In one reported probucol-induced LQTS and torsades de pointes arrhythmia case, the serum probucol concentration measured during the cardiac arrhythmia was 26 μg/ml (Hayashi et al., 2004), which is equivalent to 50.3 μM. In patients who received probucol 1 g daily for periods of 1 to 12 months, the mean plasma levels ranged from 18.2 to 39.2 μg/ml (35.2–75.9 μM) (Heeg and Tachizawa, 1980). Although the drug concentrations that reach cardiac myocytes are unknown, it seems that clinically relevant concentrations of probucol are able to cause I_Kr reduction and LQTS.

Approximately 200 LQTS-associated hERG mutations have been identified in humans, and missense (single amino acid substitution) mutations represent the dominant predicted protein abnormality (Anderson et al., 2006). Although some patients with hERG mutations may have a prolonged QT interval and clinically be asymptomatic, they are probably more vulnerable to drugs that interact with hERG channels. Previously, Hayashi et al. (2004) reported that hERG M124T mutation caused a mild-to-moderate channel dysfunction but manifested marked QT prolongation or torsade de pointes after taking probucol. They showed that probucol acts to alter hERG function as a result of a ∼10-mV depolarization shift of activation curve (Hayashi et al., 2004). The authors also reported that acutely applied probucol shifted the reversal potential of hERG channels (Hayashi et al., 2004). Our data obtained in hERG-HEK cells and neonatal rat ventricular myocytes showed that probucol has no acute effect on either hERG or I_Kr currents. The reason for the discrepancy between our data and those of Hayashi et al. is unknown. The study reported by Hayashi et al. was performed in Xenopus oocytes. However, usually higher concentrations of drugs are needed to block channels expressed in Xenopus oocytes compared with those needed in mammalian cells. There has been no other evidence showing that probucol blocks hERG channels. On the other hand, our finding that probucol disrupts I_Kr/hERG cell membrane expression provides a plausible explanation for probucol-induced LQTS and torsade de pointes.

hERG proteins are synthesized in the ER and transported to the cell surface via the Golgi apparatus. Misfolded or misassembled proteins are retained in the ER by its quality control mechanism. It is thought that mutations can cause misfolding of hERG proteins, resulting in trafficking defect (Anderson et al., 2006). Because probucol treatment did not reduce the intracellular forms of ERG (135-kDa band of hERG, 130- and 85-kDa bands of ERG in cardiomyocytes) (Fig. 5), the probucol-induced ERG membrane expression does not seem to be a result of inhibition of the channel synthesis. Instead, it could develop from either defective trafficking or accelerated membrane ERG degradation. It has been shown that E-4031, glycerol, or culture in low temperature can rescue some forms of trafficking deficient mutant hERG channels (Zhou et al., 1999). However, we found that none of these manipulations rescued probucol-disrupted hERG surface expression (data not shown). Probucol is a very lipophilic drug that can inhibit the cholesterol synthesis in the cell and may modify the lipid content of the membrane. Whether the lipid content of the cell contributes to hERG/I_Kr functional expression needs further investigation.

In summary, the present study provides evidence that probucol does not block hERG or I_Kr channels but disrupts ERG protein trafficking. In drug development, early screening of lead compounds for potential acute hERG channel blockade is becoming a common practice. The finding that probucol reduces hERG and I_Kr currents by reducing the number of functional ERG channels suggests that further strategies for evaluating the LQTS risk of drugs should be considered.