Modulation of K_2P2.1 and K_2P10.1 K⁺ channel sensitivity to carvedilol by alternative mRNA translation initiation

Abstract

Background and Purpose

The β-receptor antagonist carvedilol blocks a range of ion channels. K_2P2.1 (TREK1) and K_2P10.1 (TREK2) channels are expressed in the heart and regulated by alternative translation initiation (ATI) of their mRNA, producing functionally distinct channel variants. The first objective was to investigate acute effects of carvedilol on human K_2P2.1 and K_2P10.1 channels. Second, we sought to study ATI-dependent modulation of K_2P K⁺ current sensitivity to carvedilol.

Experimental Approach

Using standard electrophysiological techniques, we recorded currents from wild-type and mutant K_2P2.1 and K_2P10.1 channels in Xenopus oocytes and HEK 293 cells.

Key Results

Carvedilol concentration-dependently inhibited K_2P2.1 channels (IC_50,oocytes = 20.3 μM; IC_50,HEK = 1.6 μM) and this inhibition was frequency-independent. When K_2P2.1 isoforms generated by ATI were studied separately in oocytes, the IC₅₀ value for carvedilol inhibition of full-length channels (16.5 μM) was almost 5-fold less than that for the truncated channel variant (IC₅₀ = 79.0 μM). Similarly, the related K_2P10.1 channels were blocked by carvedilol (IC_50,oocytes = 24.0 μM; IC_50,HEK = 7.6 μM) and subject to ATI-dependent modulation of drug sensitivity.

Conclusions and Implications

Carvedilol targets K_2P2.1 and K_2P10.1 K⁺ channels. This previously unrecognized mechanism supports a general role of cardiac K_2P channels as antiarrhythmic drug targets. Furthermore, the work reveals that the sensitivity of the cardiac ion channels K_2P2.1 and K_2P10.1 to block was modulated by alternative mRNA translation initiation.

Introduction

Two-pore-domain potassium (K_2P) channels stabilize the resting membrane potential and facilitate action potential repolarization (channel nomenclature follows Alexander et al., 2013). Dynamic and polymodal regulation of K_2P currents determines cellular excitability (Goldstein et al., 2001; Gierten et al., 2008;2012; Thomas et al., 2008; Sandoz et al., 2011; Staudacher et al., 2011a; Kisselbach et al., 2012; Rahm et al., 2012,2014; Seyler et al., 2012). In the heart, K_2P3.1 inhibition or genetic inactivation of K_2P3.1 currents results in action potential prolongation, suggesting that these channels might represent targets for antiarrhythmic therapy (Putzke et al., 2007; Gierten et al., 2010; Ravens, 2010; Decher et al., 2011; Donner et al., 2011; Limberg et al., 2011; Staudacher et al., 2011b; Petric et al., 2012; Schmidt et al., 2012). Prolongation of action potentials and cardiac effective refractory periods is a characteristic of class III antiarrhythmic drugs, reducing membrane excitability and decreasing arrhythmia susceptibility. In addition to K_2P3.1, expression of K_2P2.1 (TREK1, TWIK-related K⁺ channel) channels has been shown in atria and ventricles at mRNA, protein and functional levels (Fink et al., 1996; Reyes et al., 1998; Aimond et al., 2000; Medhurst et al., 2001; Terrenoire et al., 2001; Liu and Saint, 2004; Tan et al., 2004; Li et al., 2006; Zhang et al., 2008; Decher et al., 2011; Zhao et al., 2011; Goonetilleke and Quayle, 2012; Schmidt et al., 2014). The closely related K_2P10.1 (TREK2) channels have been detected in the heart as well (Medhurst et al., 2001; Liu and Saint, 2004; Li et al., 2006; Staudacher et al., 2011a; Gierten et al., 2012). However, data delineating the cardiac functions of K_2P2.1 and K_2P10.1 channels are limited. Synthesis of different K_2P proteins with different N terminal domains from a single mRNA by alternative mRNA translation initiation (ATI) was recently revealed as a novel mechanism to increase the number of functional TREK protein subunits, producing two (K_2P2.1) or three (K_2P10.1) endogenous variants (Simkin et al., 2008; Thomas et al., 2008). ATI regulates K_2P current magnitude, ion selectivity and intersubunit interactions (Simkin et al., 2008; Thomas et al., 2008; Veale et al., 2010; Eckert et al., 2011). Effects of ATI on K_2P channel sensitivity to antiarrhythmic drugs have not been investigated to date.

Carvedilol is a non-selective β-adrenoreceptor antagonist with a multichannel blocking pharmacological profile. Carvedilol targets a variety of ion currents including K⁺ channels (Karle et al., 2001; Staudacher et al., 2011b). The clinical use of carvedilol is associated with reduced morbidity and mortality in heart failure patients (Poole-Wilson et al., 2003; Remme et al., 2007). In addition, suppression of atrial fibrillation and ventricular tachyarrhythmia as well as pronounced prolongation of atrial and ventricular effective refractory periods suggest specific antiarrhythmic effects of the drug (Senior et al., 1992; Brunvand et al., 1996; Cice et al., 2000; Takusagawa et al., 2000; Katritsis et al., 2003; Merritt et al., 2003; Ramaswamy, 2003; Acikel et al., 2008; Kanoupakis et al., 2008).

We hypothesized that inhibition of K_2P2.1 and K_2P10.1 background currents contributed to the antiarrhythmic effect of carvedilol. Furthermore, this study was designed to investigate modulation of the drug sensitivity of cardiac K_2P channels by ATI.

Methods

Molecular biology

Human cDNAs encoding wild-type and mutant K_2P2.1 (TREK1b) (GenBank accession number, EF165335) were kindly provided by Dr. Steve Goldstein (Brandeis University, Waltham, MA, USA) (Thomas et al., 2008). Human K_2P2.1 was amplified from a heart cDNA library (Clontech, Palo Alto, CA, USA), inserted into the pCR2.1-TOPO vector (Invitrogen, Carlsbad, CA, USA) and subcloned into pMAX, a dual-purpose expression vector containing a CMV promoter for mammalian expression and a T7 promoter for cRNA synthesis. Human K_2P10.1 (TREK2c) (GenBank accession number EU978939) was amplified from a brain cDNA library as described (Gierten et al., 2008; Staudacher et al., 2011a). Complementary DNA was inserted into the pCR2.1-TOPO vector and subcloned into pRAT, a dual-purpose expression vector containing a CMV promoter for mammalian expression and a T7 promoter for cRNA synthesis (Bockenhauer et al., 2001). Generation of mutant K_2P2.1 and K_2P10.1 clones was previously reported (Thomas et al., 2008; Staudacher et al., 2011a). Briefly, K_2P2.1 and K_2P10.1 mutations described in the text were made with the QuikChange Site-Directed Mutagenesis kit (Stratagene, La Jolla, CA, USA). N terminal channel truncation was achieved using PCR. All cDNA constructs were confirmed by DNA sequencing. Complementary RNAs were transcribed after vector linearization using T7 RNA polymerase and the mMessage mMachine kit (Ambion, Austin, TX, USA). Transcripts were quantified by spectrophotometry and cRNA integrity was assessed by agarose gel electrophoresis.

Oocyte preparation and injection

All animal care and experimental procedures complied with the Directive 2010/63/EU of the European Parliament and the United States National Institutes of Health Guide for the Care and Use of Laboratory Animals (NIH Publication no. 85-23, revised 1996). Approval was granted by the local Animal Welfare Committee (reference number A-38/11). All studies involving animals are reported in accordance with the ARRIVE guidelines for reporting experiments involving animals (Kilkenny et al., 2010; McGrath et al., 2010). The preparation of oocytes has been reported earlier in detail (Gierten et al., 2008; Staudacher et al., 2011b; Seyler et al., 2012). Briefly, oocytes were isolated from Xenopus laevis ovarian lobes after surgical removal under tricaine anaesthesia (1 g·L⁻¹; pH 7.5). Oocyte collection was alternated between left and right ovaries, and a maximum of three surgeries were performed on one individual frog. After the final collection of oocytes, anaesthetized frogs were killed by decerebration and pithing. Complementary RNA (0.1–25 ng; 46 nl per oocyte) was injected into stage V–VI defolliculated Xenopus oocytes.

Cell culture

HEK 293 cells were cultured in DMEM (Invitrogen) supplemented with 10% FBS, 2 mM L-glutamine, 100 000 U·L⁻¹ penicillin and 100 mg·L⁻¹ streptomycin in an atmosphere of 95% humidified air and 5% CO₂ at 37°C. Cells were passaged regularly. Transient transfections (5 μg cDNA/35 mm dish) were performed using FuGENE HD transfection reagent (Promega, Madison, WI, USA) according to the manufacturer's instructions.

Electrophysiology

Whole-cell currents were recorded from Xenopus oocytes using two-electrode voltage-clamp electrophysiology with an Oocyte Clamp amplifier (Warner Instruments, Hamden, CT, USA) and pCLAMP9 (Axon Instruments, Foster City, CA, USA) software 2–3 days after injection, as described (Thomas et al., 2008; Gierten et al., 2012). Data were sampled at 2 kHz and filtered at 1 kHz. Electrodes were filled with 3 M KCl. The extracellular solution contained 96 mM NaCl, 4 mM KCl, 1.1 mM CaCl₂, 1 mM MgCl₂, 5 mM HEPES (pH 7.4, adjusted with NaOH).

Whole-cell patch-clamp recordings from HEK 293 cells were carried out using a RK-400 amplifier (Bio-Logic SAS, Claix, France) as reported (Gierten et al., 2008; Schmidt et al., 2012; Seyler et al., 2012). Electrodes were filled with the following solution: 100 mM K-aspartate, 20 mM KCl, 2 mM MgCl₂, 1 mM CaCl₂, 10 mM EGTA, 10 mM HEPES, 40 mM glucose, 5 mM K-ATP (pH adjusted to 7.2 with KOH). The external solution for these experiments was composed of 140 mM NaCl, 5 mM KCl, 1 mM MgCl₂, 1.8 mM CaCl₂, 10 mM HEPES, 10 mM glucose (pH adjusted to 7.4 with NaOH). Identification of transfected cells was performed by requiring outward current amplitudes of at least 200 pA and leak currents of ≤10% of the initial maximum current. Recordings were carried out under constant perfusion at room temperature, and no leak subtraction was done during the experiments.

Drug administration

Carvedilol, metoprolol and propranolol (all from Sigma-Aldrich, St. Louis, MO, USA) were prepared as 100 mM stock solutions in DMSO (carvedilol, propranolol) or H₂O (metoprolol) and stored at −20°C. On the day of experiments, aliquots of the stock solutions were diluted to the desired concentrations with the bath solution. Wild-type and mutant K_2P2.1 or K_2P10.1 currents recorded from Xenopus oocytes were not significantly altered upon application of 0.1% DMSO (v*v⁻¹; maximum bath concentration) for 20 min (n = 4 to 10 cells were studied for each channel, P > 0.05).

Biochemistry

Pooled tissue-specific (heart, brain, aorta, lung, skeletal muscle, stomach, small intestine, kidney, uterus and prostate) total protein samples (‘Protein Medleys’) were obtained commercially from Clontech (Mountain View, CA, USA). The samples (30 μg per well) were subjected to SDS-PAGE on precast 10% gels (‘Ready Gels’, Bio-Rad, Hercules, CA, USA), followed by wet-transfer onto nitrocellulose paper and immunoblot analysis using anti-K_2P2.1 antibodies raised against recombinant C terminal rat K_2P2.1 protein containing 68 amino acids (1:400; Thomas et al., 2008). Successful detection of human K_2P2.1 protein using this antibody has been demonstrated previously (Thomas et al., 2008). Secondary fluorescence-labelled goat anti-rabbit antibodies were used at 1:5000 dilutions, and secondary labelling of fluorescent conjugates was documented with an Odyssey scanner (Li-Cor Biosciences, Lincoln, NE, USA).

Data analysis

Origin (OriginLab, Northampton, MA, USA) software was used for data analysis. Data are expressed as mean ± SEM. Concentration-response relationships for drug-induced block were fit with a Hill equation of the following form: I_drug/I_control = 1/[1 + (D/IC₅₀)ⁿ], where I indicates current, D is the drug concentration, n is the Hill coefficient and IC₅₀ is the concentration necessary for 50% block. We applied paired and unpaired Student's t-tests (two-tailed tests) to compare statistical significance of the results: P < 0.05 was considered statistically significant. The multiple comparisons shown in Figure 2F and in Table 3 were performed using one-way anova. If the hypothesis of equal means could be rejected at the 0.05-level, pair-wise comparisons of groups were made and the probability values were adjusted for multiple comparisons using the Bonferroni correction.

Figure 2.

Biophysical characteristics of hK_2P2.1 blockade by carvedilol in oocytes. (A) Open rectification of K_2P2.1 currents evoked by voltage ramps from −140 to +60 mV. Typical recordings in the absence of the drug and after superfusion with 100 μM carvedilol (20 min) are superimposed. (B–D) Blockade during extended voltage pulses. Currents activated by 7.5 s depolarizing voltage steps to +20 mV are shown under control conditions and after administration of 100 μM carvedilol (20 min). (C, D) The extent of current inhibition during the first 400 ms of the depolarizing voltage step is displayed as % (C, linear time scale; D, logarithmic time scale). (E) Carvedilol block of K_2P2.1 is frequency-independent. Relative current amplitudes recorded at +20 mV membrane potential during carvedilol application (0.5 and 0.1 Hz stimulation rates) are plotted versus time (n = 7 oocytes were studied at each rate). For the purpose of clear presentation, not all measurements are displayed. (F) Effects of metoprolol and propranolol on K_2P2.1 channels. Mean relative current amplitudes after administration of 100 μM carvedilol (Carvedil; n = 15), 100 μM metoprolol (Metopr; n = 13) and 100 μM propranolol (Propan; n = 14) are displayed. K_2P current inhibition by carvedilol was significantly greater than those induced by metoprolol and propranolol (*P < 0.05 vs. metoprolol; ***P < 0.001 vs. carvedilol). Dashed lines indicate zero current levels.

Table 3.

Statistical comparisons of relative current inhibition by indicated carvedilol concentrations between full-length (hK_2P10.1 M60I M72I) and truncated (hK_2P10.1 Δ1–59 M72I, hK_2P10.1 Δ1–71) hK_2P10.1 channel variants

	0.1 μM	1 μM	10 μM	100 μM	300 μM
P-value (n, hK2P10.1 M60I M72I; n, hK2P10.1 Δ1–59 M72I)	0.013* (7; 6)	0.032ns (6; 7)	0.001** (6; 6)	0.001** (7; 8)	0.086ns (7; 7)
P-value (n, hK2P10.1 M60I M72I; n, hK2P10.1 Δ1–71)	0.87ns (7; 6)	0.20ns (6; 7)	0.001** (6; 8)	<0.0001*** (7; 7)	<0.0001*** (7; 7)
P-value (n, hK2P10.1 Δ1–59 M72I; n, hK2P10.1 Δ1–71)	0.18ns (6; 6)	0.26ns (7; 7)	0.60ns (6; 8)	0.002** (8; 7)	0.006* (7; 7)

^*P < 0.05 (adjusted to 0.05/3 = 0.017);

^**P < 0.01 (adjusted to 0.01/3 = 0.003);

^***P < 0.001 (adjusted to 0.001/3 = 0.0003); ns, not statistically significant; ANOVA with Bonferroni adjustment.

Results

Inhibition of human cardiac K_2P2.1 (TREK1) channels by carvedilol

Carvedilol blockade of hK_2P2.1 channels was investigated in Xenopus laevis oocytes. From a holding potential of −80 mV, pulses were applied for 500 ms to voltages between −140 and +60 mV in 20 mV increments (0.5 Hz). Representative families of current traces from one cell are shown for control conditions (Figure 1A) and after application of 100 μM carvedilol for 20 min (Figure 1B). Current amplitudes were analysed at the end of the 0 mV test pulse. After a control period (20 min) with no significant amplitude changes (I_0 min, 3.17 ± 0.4 μA; I_20 min, 3.08 ± 0.4 μA; n = 14; P = 0.18), currents decreased concentration-dependently after administration of carvedilol (Figure 1A–D) and the derived IC₅₀ value is shown in Table 1, with a Hill coefficient, n_H, of 1.6 ± 0.1 (n = 12 to 15 cells were studied at each concentration). Blockade of K_2P2.1 channels in oocytes was incomplete with maximum current reduction of 80.3 ± 2.2% by 300 μM (20 min) carvedilol (Figure 1C), similar to dronedarone inhibition of K_2P2.1 reported earlier (Schmidt et al., 2012). The onset of block was rapid (Figure 1D) and 100 μM carvedilol inhibited K_2P2.1 currents by 76.3 ± 1.9% (n = 6; P = 0.001). Inhibitory effects were partially reversible, as current levels reached 50.0 ± 2.7% of controls, 20 min after removal of the drug. We cannot exclude a minor contribution of the bath solution exchange (usually requiring ∼0.5–1 min) to the kinetics of block and unblock shown in Figure 1D. Mean voltages required to achieve 50% of maximum K_2P current amplitudes recorded at +60 mV (V_50%,K2P) yielded 23.1 ± 0.5 mV under control conditions and 22.8 ± 0.4 mV after application of 100 μM carvedilol (n = 15; P = 0.36; Figure 1E,F). We observed an apparent increase of relative (but not absolute) inward currents compared with outward currents in the presence of carvedilol (Figure 1E,F). As low levels of endogenous currents (at hyperpolarizing pulses) are not blocked by carvedilol, they are expected to become more apparent in relation to significantly blocked K_2P2.1 currents at more depolarized potentials. Similar findings were obtained with K_2P10.1 channels.

Figure 1.

Inhibition of human K_2P2.1 (TREK1) channels by carvedilol in Xenopus laevis oocytes. (A, B) Representative macroscopic currents recorded under control conditions and after application of 100 μM carvedilol. Zero current levels are indicated by dashed lines. (C) Concentration-response relationships for the effect of carvedilol on K_2P2.1 currents measured at the end of the 0 mV voltage step (n = 12 to 15 cells). (D) Time course of hK_2P2.1 current blockade by 100 μM carvedilol (n = 6). Panels E and F display I–V relationships (i.e. mean current amplitudes as function of the test pulse potential) recorded under isochronal conditions (E, original current amplitudes; F, values normalized to maximum currents; n = 15).

Table 1.

Sensitivity to carvedilol of wild-type K_2P channels and of subunits generated by alternative mRNA translation initiation

Subunit	IC50,oocytes (mean ± SEM)	Relative drug sensitivity (full-length channel/truncated subunit)	IC50,HEK (mean ± SEM)
hK2P2.1
hK2P2.1 WT	20.3 ± 1.5 μM	–	1.6 ± 0.9 μM
hK2P2.1 M57I	16.5 ± 3.7 μM	–	NI
hK2P2.1 Δ1–56	79.0 ± 33.8 μM	4.8	NI
hK2P10.1
hK2P10.1 WT	24.0 ± 11.1 μM	–	7.6 ± 0.2 μM
hK2P10.1 M60I M72I	17.5 ± 4.3 μM	–	NI
hK2P10.1 Δ1–59 M72I	45.5 ± 34.2 μM	2.6	NI
hK2P10.1 Δ1–71	87.0 ± 66.3 μM	5.0	NI

NI, not investigated; WT, wild type.

K_2P2.1 currents show outward (or open) rectification. Linear ramp voltage protocols were applied between −140 and +60 mV (500 ms) before and after application of 100 μM carvedilol (20 min) to assess drug effects on current rectification (Figure 2A). The currents displayed outward rectification before and after carvedilol administration. The degree of blockade at +20 mV ramp potential was 80.5 ± 1.3% (n = 4; P = 0.001). Next, carvedilol block was investigated using extended voltage pulse durations. K_2P2.1 currents were recorded during a single depolarizing step to +20 mV for 7.5 s under control conditions and after application of 100 μM carvedilol for 20 min while keeping the membrane potential at −80 mV (Figure 2B). Current inhibition during the first 400 ms of the protocol is displayed with linear and logarithmic time scales in Figure 2C and D. Analysis of channel block after carvedilol administration revealed that significant inhibition of hK_2P2.1 channels had occurred at −80 mV, indicated by the level of block at the beginning of the depolarizing pulse (60.2 ± 2.3%; n = 8; P < 0.0001) (Figure 2C,D). A second, time-dependent component of block was observed during the +20 mV-pulse (block at 400 ms: 72.2 ± 1.0%; n = 8; P < 0.0001), similar to ∼76% inhibition observed above with 500 ms-voltage steps.

To study frequency-dependence of block, human K_2P2.1 channels were rapidly activated by a depolarizing step to +20 mV (500 ms) at intervals of 2 or 10 s, respectively. Current reduction by 30 μM carvedilol was plotted versus time (Figure 2E). The degree of inhibition after 20 min was not significantly different (P = 0.079) between 0.5 Hz (53.0 ± 0.9%; n = 7; P = 0.002) and 0.1 Hz stimulation rate (58.2 ± 2.6%; n = 7; P = 0.001). Finally, the specificity of K_2P2.1 current block was investigated. Compared with carvedilol, K_2P2.1 displayed reduced affinity to the selective β-adrenoceptor antagonist metoprolol (100 μM; 20 min) and to the non-selective β₁- and β₂- adrenoceptor blocker propranolol (100 μM; 20 min) under experimental conditions similar to those described above (Figure 1). Metoprolol reduced hK_2P2.1 currents by 18.8 ± 2.5% (n = 13; P = 0.0003) and propranolol blocked the channels by 31.2 ± 3.7% (n = 14; P = 0.001) (Figure 2F).

Carvedilol targets hK_2P10.1 (TREK2) channels

To further elucidate the pharmacological profile of carvedilol, its effects on human K_2P10.1 channels were analysed in oocytes. Concentration-dependent inhibition of K_2P10.1 current was revealed using the experimental approach described above (Figure 3A–D) and the derived IC₅₀ value (Figure 3C), with n_H of 2.2 ± 1.0, from a series of 6–7 cells is shown in Table 1. Currents were stable during a 20-min control period (I_0 min, 1.73 ± 0.23 μA; I_20 min, 1.65 ± 0.2 μA = 12; P = 0.12). During carvedilol application (100 μM), K_2P10.1 currents decreased by 67.5 ± 3.3% (n = 7; P = 0.0007) (Figure 3D). Current inhibition was reduced to 40.5 ± 10.0% during washout, indicating partial reversibility of block. Analysis of K_2P10.1 current-voltage (I–V) relationships yielded a negligible change of 1.1 mV towards more negative potentials in the presence of carvedilol (V_{50%,K2P,control} = 25.5 ± 1.2 mV; V_{50%,K2P,carvedilol} = 24.4 ± 1.0 mV; n = 7; P = 0.02) (Figure 3E,F).

Figure 3.

Human K_2P10.1 (TREK2) channels in oocytes are sensitive to carvedilol. Analyses performed as described in Figure 1 illustrate concentration-dependent K_2P10.1 current inhibition (A–C; n = 6–7 cells). Zero current levels are indicated by dashed lines. (D) Time course of blockade by 100 μM carvedilol (n = 7). Open rectification (E, F) and current-voltage relationships of K_2P10.1 channels were not markedly affected by carvedilol (100 μM; n = 7).

Concentration-dependent blockade of hK_2P2.1 and hK_2P10.1 channels in mammalian cells

To evaluate the physiological significance of drug-receptor interactions, concentration-response relationships obtained from mammalian expression systems are required. K_2P2.1 and K_2P10.1 potassium channels were expressed in HEK 293 cells to analyse channel sensitivities to carvedilol (Figure 4). From a holding potential of −80 mV, pulses were applied for 500 ms to voltages between −120 and +80 mV in 20 mV increments (0.2 Hz). Following current equilibration (10 min), the degree of inhibition was determined at +60 mV 10 min after carvedilol application (Figure 4A,B,D,E). For HEK 293 cells, the IC₅₀ values for blockade of hK_2P2.1 channels (Hill coefficient n_H = 0.61 ± 0.20; Figure 4C) and for hK_2P10.1 background channels (Hill coefficient n_H = 1.26 ± 0.06; Figure 4F) are given in Table 1. Four to eight cells were investigated at each concentration. Steady-state outward currents recorded in the presence of 100 μM carvedilol were blocked by 89.4 ± 4.0% (hK_2P2.1; n = 4; P = 0.17) and by 96.6 ± 1.5% (hK_2P10.1; n = 4; P = 0.04).

Figure 4.

Inhibition of hK_2P2.1 and hK_2P10.1 currents expressed in HEK293 cells. Representative K_2P2.1 and K_2P10.1 measurements before (A, D) and after exposure to 10 μM carvedilol (B, E) are displayed. (C, F) Concentration-response relationships indicating IC₅₀ values of respective K_2P channels (n = 4–8 cells).

Alternative translation initiation regulates K_2P channel sensitivity to carvedilol

Alternative translation initiation generates two (K_2P2.1) and three (K_2P10.1) functionally different K_2P ion channel subunits respectively (Figure 5A–E) (Simkin et al., 2008; Thomas et al., 2008; Staudacher et al., 2011a). Immunoblot analysis of tissue-specific total protein samples revealed significant K_2P2.1 protein expression in adult human heart in addition to brain, aorta, skeletal muscle, stomach, small intestine, uterus and prostate (Figure 5C). Furthermore, the presence of a smaller K_2P2.1 band (which may correspond to truncated K_2P2.1 channels produced by ATI) was detected in the heart and other tissues (Figure 5C). K_2P10.1 protein analysis was not performed as reliable cardiac immunoblot protocols and antibodies have not yet been established. To study differential pharmacological effects of carvedilol on K_2P isoforms, mutant subunits that specifically produce full-length channels by elimination of downstream start codons were used (K_2P2.1 M57I; K_2P10.1 M60I M72I). Drug effects on these channels were compared with truncated subunits associated with ATI (K_2P2.1 Δ1–56; K_2P10.1 Δ1–59 M72I; K_2P10.1 Δ1–71). Figures 5B and E provide an overview of the channels investigated. Using the experimental approach described in Figures 1 and 3, marked differences in drug affinity were detected. Carvedilol (100 μM) reduced full-length hK_2P2.1 M57I currents by 72.6 ± 3.3% (n = 10; P = 0.0002; Figure 6A,B) compared with 40.7 ± 3.7% inhibition of truncated hK_2P2.1 Δ1–56 channels (n = 10; P = 0.0001; Figure 6F,G). The corresponding IC₅₀ values for blockade of hK_2P2.1 M57I channels (Hill coefficient n_H = 1.4 ± 0.5; Figure 6C) and of hK_2P2.1 Δ1–56 channels (Hill coefficient n_H = 1.5 ± 0.6; Figure 6H) are shown in Table 1. Four to 10 cells were investigated at each concentration. Carvedilol inhibition of full-length and truncated K_2P2.1 variants (K_2P2.1 M57I and K_2P2.1 Δ1–56) was statistically compared. The resulting P-values (Table 2) illustrate statistically significant differences in the degree of blockade at concentrations >1 μM, providing statistical support for different drug affinities between full-length and truncated K_2P2.1 channel variants. Please note that significant differences at 0.1 and 1 μM carvedilol are not expected as there was virtually no inhibition of either subunit at these concentrations (Figure 6C,H). Activation voltages of both K_2P2.1 isoforms were virtually unchanged in the presence of 100 μM carvedilol (M57I; V_{50%,K2P,control} = 28.1 ± 0.3 mV; V_{50%,K2P,carvedilol} = 26.0 ± 0.3 mV; n = 10; P < 0.001; Δ1–56; V_{50%,K2P,control} = 31.2 ± 0.4 mV; V_{50%,K2P,carvedilol} = 29.7 ± 0.7 mV; n = 10; P = 0.058).

Figure 5.

Alternative translation initiation of hK_2P2.1 and hK_2P10.1 mRNA produces multiple channel isoforms. (A, D) N terminal amino acid sequence alignments of wild-type (WT) K_2P2.1 (A) and K_2P10.1 channels (D). Alternative translation initiation sites are indicated by arrows. In addition, alignments of mutated channels used to specifically produce full-length or truncated K_2P proteins in isolation are shown. (B, E) The topologies of K_2P2.1 (B) and K_2P10.1 (E) WT channels (left) and of protein variants arising from alternative mRNA translation initiation studied in this work (right) are displayed. (C) Expression of two K_2P2.1 protein isoforms in indicated human tissues (skeletal m., skeletal muscle; small int., small intestine) visualized by Western blot analysis using an antibody to the K_2P2.1 C terminus. Markers correspond to apparent molecular weights: 50, 50 kDa; 37, 37 kDa.

Figure 6.

Human K_2P2.1 variants produced by ATI exhibit different carvedilol affinity. Channels were studied in Xenopus oocytes using indicated voltage protocols. (A, B, F, G) Representative current recordings under control conditions (A, F) and after 20 min perfusion with 100 μM carvedilol (B, G) are illustrated for indicated isoforms. Concentration-response relationships show a 4.8-fold higher IC₅₀ value for K_2P2.1 Δ1–56 (H) compared with K_2P2.1 M57I (C) (n = 6 to 10 cells analysed at each concentration). (D, E, I, J) Voltage-dependence of activation before and after carvedilol application (D, I, original current amplitudes; E, J, values normalized to maximum currents; n = 10). Dashed lines indicate zero current level.

Table 2.

Statistical comparisons of relative current inhibition by a range of carvedilol concentrations between full-length (K_2P2.1 M57I) and truncated (K_2P2.1 Δ1–56) hK_2P2.1 channel variants

	0.1 μM	1 μM	10 μM	100 μM	300 μM
P-value (n, hK2P2.1 M57I; n, hK2P2.1 Δ1–56)	0.19ns (10; 8)	0.21ns (10; 6)	0.002** (9; 8)	<0.0001*** (10; 10)	0.014* (10; 4)

^*P < 0.05;

^**P < 0.01;

^***P < 0.001; ns, not statistically significant.

ATI-derived K_2P10.1 channel isoforms were differentially regulated by carvedilol as well (Figure 7). Representative current traces are shown for control conditions (Figure 7A,F,K) and after application of 100 μM carvedilol (Figure 7B,G,L) for indicated subunits. Currents were blocked by 69.7 ± 3.5% (K_2P10.1 M60I M72I; n = 7; P = 0.001), by 45.3 ± 4.6% (K_2P10.1 Δ1–59 M72I; n = 8; P = 0.0002) and by 26.5 ± 1.1% (K_2P10.1 Δ1–71; n = 7; P = 0.002) respectively. The IC₅₀ values for block of hK_2P10.1 M60I M72I (n_H = 1.3 ± 0.4; n = 6–7) and of hK_2P10.1 Δ1–59 M72I (n_H = 0.97 ± 0.3; n = 6–8) and of hK_2P10.1 Δ1–71 (n_H = 0.8 ± 0.3; n = 6–8) channels are shown in Table 1. Statistically significant differences between inhibition of full-length channels (K_2P10.1 M60I M72I) and both truncated variants (K2P10.1 Δ1–59 M72I, K_2P10.1 Δ1–71) were observed at 100 μM carvedilol (Table 3), confirming different drug sensitivity at concentrations close to the respective IC₅₀ values. I–V relationships of all K_2P10.1 variants tested were not markedly modified by 100 μM carvedilol (K_2P10.1 M60I M72I; V_{50%,K2P,control} = 25.9 ± 0.6 mV; V_{50%,K2P,carvedilol} = 25.0 ± 0.5 mV; n = 7; P = 0.04; K_2P10.1 Δ1–59 M72I; V_{50%,K2P,control} = 30.3 ± 0.4 mV; V_{50%,K2P,carvedilol} = 27.8 ± 0.4 mV; n = 8; P = 0.001; K_2P10.1 Δ1–71; V_{50%,K2P,control} = 30.7 ± 0.4 mV; V_{50%,K2P,carvedilol} = 29.3 ± 0.5 mV; n = 7; P = 0.005; Figure 7D,I,N).

Figure 7.

ATI affects inhibition of hK_2P10.1 channels by carvedilol. Representative macroscopic recordings from oocytes and activation curves reflect blockade of full-length (K_2P10.1 M60I M72I; A, B, D, E) and of two truncated isoforms (K_2P10.1 Δ1–59 M72I; F, G, I, J; K_2P10.1 Δ1–71; K, L, N, O) by 100 μM carvedilol. Zero current levels are marked by dashed lines. Corresponding concentration-response curves demonstrate decreased carvedilol sensitivity of truncated isoforms (H, M) compared with full-length channels (C) (n = 7 to 8 cells were studied).

Discussion

Carvedilol inhibits hK_2P2.1 (TREK1) and hK_2P10.1 (TREK2) K_2P channels

We report concentration-dependent inhibition of hK_2P2.1 and hK_2P10.1 currents by carvedilol with IC₅₀ values of 20.3 and 24.0 μM obtained from Xenopus laevis oocytes (Table 1). An extension of these findings to I_TREK1 and I_TREK2 recorded from mammalian HEK 293 cells revealed IC₅₀ values of 1.6 and 7.6 μM respectively. Differences in IC₅₀ values between mammalian cells and oocytes are commonly observed and attributed to specific properties of oocytes (e.g. the vitelline membrane and the yolk) that reduce drug affinities. Different IC_50,oocytes/IC_{50,mammalian cells} ratios are primarily determined by physicochemical drug characteristics. Here, these ratios yield 12.7 (hK_2P2.1) and 3.2 (hK_2P10.1) for the same drug, carvedilol, suggesting that channel-specific mechanisms of drug access to the putative receptor site within the pore are affected by oocyte properties as well. During therapeutic application of the drug, total plasma levels between 47 and 615 nM have been measured (McPhillips et al., 1988; Morgan et al., 1990; Gehr et al., 1999; Karle et al., 2001). Carvedilol is approximately 98% bound to plasma proteins, reducing effective free drug concentrations to 0.94–12.3 nM, which would suggest a minimal effect of K_2P channel blockade on cardiac electrophysiology under normal pharmacokinetic conditions.

Biophysical properties of hK_2P2.1 blockade by carvedilol

The biophysical mechanism of K_2P2.1 inhibition by carvedilol was analysed in detail. A rapid onset of block supports a direct drug-channel interaction as opposed to increased protein turnover or accelerated channel degradation as a molecular mechanism of action. In addition, carvedilol might indirectly affect K_2P2.1 or K_2P10.1 function through inhibition of β-adrenoreceptor signalling in oocytes or HEK 293 cells. However, this mechanism appears unlikely as activation of the cAMP/PKA system caused K_2P2.1 and K_2P10.1 current reduction, rather than enhancement (Gu et al., 2002; Honoré et al., 2002; Murbartián et al., 2005). Furthermore, functional data suggest that levels of endogenous β-adrenoreceptors are negligible (Kathöfer et al., 2000). The lack of complete K_2P2.1 and K_2P10.1 current inhibition in oocytes compared with mammalian cells may be attributed to specific lipophilic properties of the oocyte expression system. Open rectification that is characteristic to K_2P channel function in physiological ionic conditions was observed before and during drug block. Finally, the lack of frequency-dependence of K_2P2.1 channel blockade supports continuous accessibility of the putative drug binding site inside the channel. It is likely that kinetics of drug block (Figure 2) were largely independent of channel opening as the cytosolic channel gate is constitutively open (Piechotta et al., 2011; Brohawn et al., 2012; Miller and Long, 2012), allowing for drug binding within the channel pore in state-independent manner.

ATI determines drug sensitivity of K_2P channels

This work identifies ATI as a genetic mechanism regulating sensitivity of K_2P2.1 and K_2P10.1 channels to carvedilol. Compared with full-length K_2P channels, truncated isoforms produced by ATI were 2.6 to 5.0-fold less sensitive to the drug (Table 1). These findings are in line with a previous report describing decreased sensitivity (by 70%) of truncated K_2P2.1 channels to the antidepressant drug fluoxetine (Eckert et al., 2011). The present study extends regulation of K⁺ channel drug sensitivity by ATI to cardiac K_2P channels and to a cardiovascular drug, suggesting a broader role of this pharmacological mechanism. Several studies have identified the C-terminus of K_2P2.1 channels as a critical region for polymodal channel regulation (Honoré, 2007). Reduced carvedilol affinity observed in N-terminally truncated isoforms now indicates a significant contribution of the N-terminus of K_2P2.1 and K_2P10.1 channels. Truncated K_2P2.1 ATI variants may exhibit reduced carvedilol sensitivity, because of the additional Na⁺ conductance of the channel pore (Thomas et al., 2008). However, this mechanism could not explain reduced sensitivity among K_2P10.1 ATI variants, as loss of K⁺ selectivity has not been demonstrated for this channel.

The apparent role of the N-terminus in K_2P channel regulation is in line with data obtained from human ether-a-go-go-related gene (hERG) K⁺ channels. The length of the hERG channel N-terminus is determined by mRNA splicing (as opposed to ATI-dependent regulation of the K_2P2.1 N-terminus reported in the present work). The transcript variant hERG 1b is characterized by a shortened N-terminus and displays reduced affinity to the class III compounds dofetilide and E-4031 (Abi-Gerges et al., 2011). It is reasonable to assume that reduced length of the N-terminus modulates drug sensitivity through alterations of the channel pore structure that contains the putative binding site. As the N-terminus is not detailed in the crystal structures of K_2P1.1 and mechanosensitive K_2P4.1 channels, the precise mechanism by which the N-terminus regulates K_2P channel pharmacology remains to be elucidated (Brohawn et al., 2012; Miller and Long, 2012).

Clinical implications

Class III antiarrhythmic drugs suppress atrial fibrillation and ventricular tachyarrhythmia through K⁺ channel block, resulting in prevention of electrical reentry via prolongation of action potential duration and cardiac refractoriness. Outward potassium currents mediated by K_2P channels contribute to cardiac repolarization. Thus, inhibition of cardiac K_2P channels represents a novel concept in antiarrhythmic therapy, with carvedilol blockade of hK_2P2.1 and hK_2P10.1 K⁺ currents observed in the current study and inhibition of hK_2P3.1 channels reported previously (Staudacher et al., 2011b) serving as proof of this principle. The specific contribution of K_2P channel blockade to the antiarrhythmic efficacy of carvedilol in heart failure and atrial fibrillation patients (Senior et al., 1992; Brunvand et al., 1996; Cice et al., 2000; Takusagawa et al., 2000; Katritsis et al., 2003; Merritt et al., 2003; Ramaswamy, 2003; Haghjoo et al., 2007; Acikel et al., 2008), however, is expected to be low, as the IC₅₀ values determined in vitro in the low micromolar range were much higher than the nanomolar free therapeutic drug concentrations found in clinical practice. Compared with K_2P2.1 (IC₅₀ = 1.6 μM) and K_2P10.1 channels (IC₅₀ = 7.6 μM), higher sensitivity implying greater clinical effect was demonstrated for the K_2P3.1 channels (IC₅₀ = 0.8 μM; Staudacher et al., 2011b). The overall antiarrhythmic efficacy of the drug is likely to result primarily from its anti-adrenergic properties, supported by multichannel blocking effects on potassium currents (I_Kr, I_Ks, I_to, I_Kur, I_TASK1, I_TREK1, I_TREK2, I_K,ATP), pacemaker current (I_f), L-type calcium current and the cardiac ryanodine receptor with different affinities (Cheng et al., 1999; Kawakami et al., 2006; Kikuta et al., 2006; Deng et al., 2007; Yokoyama et al., 2007; Staudacher et al., 2011b; Zhou et al., 2011).

In conclusion, this investigation showed inhibition of K_2P2.1 and K_2P10.1 channels by carvedilol, providing proof-of-concept to support a pharmacological role of cardiac K_2P channels as antiarrhythmic drug targets. The therapeutic significance of K_2P channel blockade in heart rhythm disorders requires validation in translational and clinical studies. At the molecular level, antiarrhythmic drug sensitivity is determined by alternative translation initiation of K_2P2.1 and K_2P10.1 mRNA that produces subunits with different carvedilol affinity. Differential expression of ATI-dependent K_2P channel variants may result in spatiotemporal variation of electrophysiological carvedilol effects in the heart. In addition, the effect of ATI on antiarrhythmic drug efficacy has broad and general implications for cardiac drug discovery.

Glossary

ATI

alternative translation initiation

K_2P

two-pore-domain K⁺ channel

TREK

TWIK-related K⁺ channel

TWIK

tandem of P domains in a weak inward rectifying K⁺ channel

Conflict of interest

None declared.