The anti-protozoal drug pentamidine blocks K_IR2.x-mediated inward rectifier current by entering the cytoplasmic pore region of the channel

Abstract

Background and purpose:

Pentamidine is a drug used in treatment of protozoal infections. Pentamidine treatment may cause sudden cardiac death by provoking cardiac arrhythmias associated with QTc prolongation and U-wave alterations. This proarrhythmic effect was linked to inhibition of hERG trafficking, but not to acute block of ion channels contributing to the action potential. Because the U-wave has been linked to the cardiac inward rectifier current (I_K1), we examined the action and mechanism of pentamidine-mediated I_K1 block.

Experimental approach:

Patch clamp measurements of I_K1 were made on cultured adult canine ventricular cardiomyocytes, K_IR2.1-HEK293 cells and K_IR2.x inside-out patches. Pentamidine binding to cytoplasmic amino acid residues of K_IR2.1 channels was studied by molecular modelling.

Key results:

Pentamidine application (24 h) decreased I_K1 in cultured canine cardiomyocytes and K_IR2.1-HEK293 cells under whole cell clamp conditions. Pentamidine inhibited I_K1 in K_IR2.1-HEK293 cells 10 min after application. When applied to the cytoplasmic side under inside-out patch clamp conditions, pentamidine block of I_K1 was acute (IC₅₀= 0.17 µM). Molecular modelling predicted pentamidine-channel interactions in the cytoplasmic pore region of K_IR2.1 at amino acids E224, D259 and E299. Mutation of these conserved residues to alanine reduced pentamidine block of I_K1. Block was independent of the presence of spermine. K_IR2.2, and K_IR2.3 based I_K1 was also sensitive to pentamidine blockade.

Conclusions and implications:

Pentamidine inhibits cardiac I_K1 by interacting with three negatively charged amino acids in the cytoplasmic pore region. Our findings may provide new insights for development of specific I_K1 blocking compounds.

Introduction

The ventricular cardiac action potential (AP) has a long plateau and repolarization phase, essential for contractility and recovery of excitability. Between successive APs, there is a constant resting membrane potential based on potent inward rectifier current (I_K1). In mammals, the KCNJ2 and KCNJ12 gene products K_IR2.1 and K_IR2.2 (nomenclature follows Alexander et al., 2009) constitute the main determinants of I_K1 in the ventricle (Dhamoon and Jalife, 2005). In addition, K_IR2.3 encoded by the KCNJ4 gene contributes to cardiac I_K1. Inhibition of endogenous I_K1 by dominant negative K_IR2.1 expression in the working ventricular myocardium of guinea pigs induces ectopic pacemaker activity (Miake et al., 2002). In humans, loss-of-function mutations in K_IR2.1 correlates with Andersen-Tawil syndrome (ATS1) characterized by ventricular arrhythmias, periodic paralysis and dysmorphic features (Plaster et al., 2001; Tristani-Firouzi et al., 2002). ATS1 is associated with the appearance of more prominent U waves (Zhang et al., 2005a; Nagase et al., 2007).

Acquired ion channel dysfunction, either by loss or gain of function, may result in a wide variety of cardiac arrhythmias (Kannankeril and Roden, 2007; Murphy and Dargie, 2007). The ion current involved is often the K_v11.1 (human ether-a-go-go related gene; hERG)-based, rapid component of the delayed rectifier current (I_Kr). Unfortunately, non-cardiac drugs may affect this current, but many other currents as well. Recently, the widely used antimalarial drug chloroquine and the oestrogen receptor antagonist tamoxifen were demonstrated to inhibit K_IR2.1-based I_K1 by blocking the ion channel pore from the cytoplasmic side or by interfering in the K_IR2.1–PIP₂ interaction, respectively (Rodríguez-Menchaca et al., 2008; Ponce-Balbuena et al., 2009).

Pentamidine belongs to the diamine family used for treatment of pathogenic protozoal infections causing human African trypanosomis (sleeping sickness), visceral Leishmaniasis (Bray et al., 2003), and opportunistic pathogenic infections causing candidiasis and pneumonia in immunocompromised individuals, such as HIV-positive patients (Sands et al., 1985; Goa and Campoli-Richards, 1987). Clinically, pentamidine has been associated with QTc prolongation, U-wave amplitude increase, U-wave alternans and Torsades de Pointes arrhythmias with and without electrolyte abnormalities (Bibler et al., 1988; Mitchell et al., 1989; Gonzalez et al., 1991; Quadrel et al., 1992; Eisenhauer et al., 1994; Girgis et al., 1997). Recently, pentamidine has been shown to reduce I_Kr and to prolong the AP, because it hampers transport of the K_v11.1 ion channel protein towards the sarcolemma (Cordes et al., 2005; Kuryshev et al., 2005). Although this might explain QTc prolongation, U-wave alterations remain unexplained. Given the similarity between alterations in U-wave morphology under conditions of KCNJ2 mutations and pentamidine administration, we have assessed whether pentamidine affects the inward rectifier current and its mode of action at the molecular level.

We demonstrate that (i) pentamidine inhibits I_K1 in isolated adult ventricular cardiac myocytes at clinical concentrations; that (ii) it produces acute block of K_IR2.x (inward rectifier channel subunit)-based I_K1 by plugging the pore region from the cytoplasmic side; that (iii) the acidic amino acid residues E224, D259 and E299 of K_IR2.1 are directly involved in pentamidine-mediated block of the outward I_K1.

Methods

Cell culture

All animal care and experimental procedures conformed to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication no. 85-23, revised 1996) and was approved by the institutional committee for animal experiments. Left ventricular myocytes from adult dogs were isolated and GFP-tagged K_IR2.1 HEK293 (human embryonic kidney cell line stably expressing GFP-tagged inward rectifier channel protein K_IR2.1; HEK-KWGF) cells were generated and cultured as described previously (De Boer et al., 2006a,b;). In inside-out experiments, HEK293T cells were transfected with 20 ng KCNJx cDNA and 6 ng EGFP per cm² and used within 48 h. Mouse KCNJ2, KCNJ12 and KCNJ4 pCXN2 expression constructs and human wild type E224A, F254A, D259A and E299A mutant KCNJ2 constructs have been described previously (Ishihara and Ehara, 2004; Rodríguez-Menchaca et al., 2008).

Pentamidine

Pentamidine-isethionate (Pentacarinat® 300, Sanofi Aventis, Gouda, The Netherlands) was dissolved in water to provide a stock solution of 0.1 M, sterilized by filtration (0.22 µm), aliquoted and stored at −20°C until use.

Western blotting

Cells were lysed in RIPA buffer [20 mM Tris-HCl, pH 7.4, 150 mM NaCl, 10 mM Na₂HPO₄, 1% (v/v) Triton X-100, 1% (w/v) Na-deoxycholate, 0.1% (w/v) SDS, 1 mM EDTA, 50 mM NaF, 2 mM PMSF and 14 µg·mL⁻¹ aprotinin]. Lysates were clarified by centrifugation at 14 000× g for 10’ at 4°C and mixed with loading buffer. Twenty-five micrograms of proteins were separated by 10% SDS-PAGE and blotted onto nitrocellulose membrane (Bio-Rad, Veenendaal, Netherlands). Protein transfer was assessed by Ponceau S staining (Sigma, St Louis, MO, USA). After blocking with 5% non-fat milk, blots were incubated with K_IR2.1 (Cat. no. sc-28633; Santa Cruz Biotechnology, Santa Cruz, CA, USA) and Cadherin (Cat. no. C1821; Sigma) antibodies for adult cardiomyocytes, or GFP antibody (Cat. no. sc-9996; Santa Cruz Biotechnology) for HEK-KWGF cells. Finally, peroxidase-conjugated secondary antibody (Jackson ImmunoResearch, West Grove, PA, USA) was applied. Standard ECL procedure was used as final detection (Amersham Bioscience, Buckinghamshire, UK). ImageQuant TL software (Amersham) was used for signal quantification.

Electrophysiology

Patch clamp measurements were made using a HEKA EPC-10 Double Plus amplifier controlled by PatchMaster 2.10 software (HEKA, Lambrecht/Pfalz, Germany). Voltage clamp measurements of whole cell I_K1 were made by applying 1 s test pulses ranging between −120 and +40 mV, with 10 mV increments, from a holding potential of −40 mV and with minimal series resistance compensation of 70%. Steady-state current at the end of the pulse was normalized to cell capacitance and plotted versus test potential. Inside-out patch-clamp measurements of I_K1 were made in absence of Mg²⁺ as described previously (Ishihara and Ehara, 2004). Excised membrane patches were placed close to the inflow region of the recording chamber, and experiments were started after inward rectification due to endogenous polyamines had disappeared as completely as possible. Currents were elicited using a ramp protocol, from −100 to 100 mV in 5 s, starting at a holding potential of −40 mV. Recorded current traces were normalized to the holding current at −40 mV obtained from the control traces. After recording control traces, excised membrane patches were exposed to bath solution containing spermine and/or pentamidine. For determination of IC₅₀, each excised membrane patch was exposed to a series of pentamidine concentrations. Fractional block of outward I_K1 was calculated using current level at +50 mV, and values obtained with experimental solutions were divided by corresponding values from the previously recorded control trace. Patch pipettes were made with a Sutter P-2000 puller (Sutter Instrument, Novato, CA, USA), and had resistances of 2–3 MΩ. Extracellular solution for whole cell I_K1 measurements contained (in mmol·L⁻¹): NaCl 140, KCl 5.4, CaCl₂ 1, MgCl₂ 1, glucose 6, NaHCO₃ 17.5, HEPES 15, pH 7.4/NaOH. The pipette solution contained potassium gluconate 125, KCl 10, HEPES 5, EGTA 5, MgCl₂ 2, CaCl₂ 0.6, Na₂ATP 4, pH 7.20/KOH. Inside-out experiments were done with a bath solution containing: KCl 125, EDTA (2K) 4, K₂HPO₄ 7.2, KH₂PO₄ 2.8, pH 7.20/KOH. The pipette solution contained KCl 145, CaCl₂ 1, HEPES 5, pH 7.40/KOH.

Molecular modeling and ligand docking

Pentamidine was docked into the crystal structure of the cytoplasmic K_IR2.1 domain (pdb identifier: 1U4F) (Pegan et al., 2005). Drug coordinates were obtained from the ZINC database (Irwin and Shoichet, 2005). FlexX (version 1.20.1) in Sybyl8.0 (Tripos International, St. Louis, MI, USA) was used to randomly dock pentamidine (n= 100) into the cytoplasmic cavity of K_IR2.1 using default parameters. Ligand partial charges were calculated with the Gasteiger–Hückel method. Modelling was visualized with Pymol software (Delano Scientific, Palo Alto, CA, USA). Modelling, docking and visualization were performed on a Linux 4 Intel Core2 Quad workstation, SUSE Linux 10.2 operating system (Novell Inc., Waltham, MA, USA).

Statistics

Group averages are presented as mean ± standard error of mean, and were tested for significance using Student's t-test (two groups) or an anova test with a Holm's post hoc test when more than two groups were involved. Analysis was done using Kaleidagraph 4.0 (Synergy Software, Reading, PA, USA), with the significance level at P < 0.05.

Results

Pentamidine inhibits cardiac I_K1

Based on the findings of pentamidine-associated U-wave alterations in patients and a correlation between the U-wave and I_K1, we hypothesized that pentamidine might inhibit I_K1 in cardiac myocytes. Figure 1 and Table 1 show that pentamidine treatment at 10 µM for a period of 24 h inhibited the inward and outward components of I_K1 by 62 and 73%, respectively, in adult canine ventricular myocytes (Figure 1A and C, Table 1). One micromolar pentamidine appeared to yield mild inhibition, but this inhibitory effect did not reach significance, even when the period of application was prolonged from 24 to 48 h. Furthermore, pentamidine substantially changed the kinetics of the inward component; the time to 90% of maximal activity increased almost 30-fold (Table 1). K_IR2.1 protein expression levels as assessed by Western blot revealed no decrease at 24 h and 18% decrease at 48 h of 10 µM pentamidine incubation (Figure 1D).

Figure 1.

Pentamidine inhibits I_K1 in cultured cardiomyocytes. Treatment for 24 h with pentamidine decreased I_K1 densities in cultured adult dog cardiomyocytes (A, C) and HEK-KWGF cells (B, C). Besides reduced current levels, also inward current activation was slowed, as can be seen with 10 µM pentamidine in dog cardiomyocytes, and more pronounced with HEK-KWGF cells (see insets depicting currents evoked by a step to −120 mV) (A, B). Representative image (n= 2) of K_IR2.1 expression in cultured adult cardiomyocytes in time (h) following pentamidine treatment (10 µM) as determined by Western blot, N-cadherin and Ponceau staining serve as loading control for quantification (D, upper panel). Representative image (n= 3) of K_IR2.1-GFP expression in HEK-KWGF cells after pentamidine treatment (10 µM) as determined by Western blot, Ponceau staining serves as loading control for quantification (D, lower panel). HEK-KWGF, human embryonic kidney cell line stably expressing GFP-tagged inward rectifier channel protein K_IR2.1.

Table 1.

Twenty-four-hour pentamidine treatment alters I_K1 characteristics

	Adult dog cardiomyocytes		HEK-KWGF
	Control	10 µM	Control	1 µM	10 µM
Max. outward current (pA/pF)	0.30 ± 0.04 (6)	0.08 ± 0.02 (6)	29.0 ± 8.0 (6)	1.6 ± 1.1 (9)	−9.4 ± 7.2 (6)
Max. inward current (pA/pF)	−13.3 ± 1.2 (6)	−5.1 ± 0.1 (6)	−482 ± 128 (6)	−162.9 ± 59.0 (6)	42.0 ± 19.0 (6)
90% max. current rise time (ms)	22.0 ± 0.4 (6)	541 ± 85 (6)	2.6 ± 0.2 (6)	ND	727 ± 36 (5)

P < 0.05 for all treatments versus control. Number of experiments are indicated between brackets.

HEK-KWGF, human embryonic kidney cell line stably expressing GFP-tagged inward rectifier channel protein K_IR2.1; ND, not determined.

Cardiac I_K1 is mainly determined by K_IR2.1 protein. Therefore, we analyzed whether K_IR2.1 based I_K1 density is sensitive to pentamidine in HEK293 cells stably expressing GFP-tagged K_IR2.1 (HEK-KWGF). Under these conditions, the effects of pentamidine were even stronger. Even 1 µM pentamidine applied for 24 h reduced the inward (66%) and outward component of I_K1 (94%) with virtually complete blockade at 10 µM pentamidine (Figure 1B and C, Table 1). At the protein level, a 17, 35 and 50% inhibition of expression levels was found at 16, 24 and 48 h of incubation with 10 µM pentamidine, respectively.

Pentamidine acutely blocks I_K1 channels when applied from the cytoplasmic side

To further explore the time relationship between pentamidine application and I_K1 block, the drug was infused during continuous I_K1 measurement. Figure 2A shows that acute application of 10 µM pentamidine decreased I_K1 after 10 min in HEK-KWGF cells in the whole cell patch mode. Altered opening kinetics of the inward component was seen from that time onwards (inset). This effect was slow, because pentamidine acts from the cytoplasmic side of the ion channel as can be appreciated from the data in the inside-out patch configuration (Figure 2A), resulting in an IC₅₀ of 0.17 ± 0.04 µM and a Hill coefficient of −0.87 ± 0.10 (n= 11) (Figure 2B and C).

Figure 2.

Pentamidine dose-response relationship for K_IR2.1 mediated I_K1. Typical recordings demonstrating a much slower developing pentamidine (10 µM)-induced I_K1 block in whole cells, compared with inside-out recordings. Inset: Current traces during drug application in the whole cell configuration display altered channel kinetics (arrows). Control cells displayed no rundown for at least 20 min (A). Typical inside-out recording depicting inhibition of I_K1 with increasing concentrations of pentamidine, demonstrating strongest effects on outward current. Dashed trace is the control recording (B). Average concentration-response curve for I_K1 recorded in the inside-out configuration (n= 11) (C). Comparison of the inhibitory effect of chloroquine and pentamidine (both 1 µM) measured in inside-out experiments. Pentamidine has a significantly stronger blocking effect at this concentration, particularly for outward currents (D).

We next compared pentamidine-mediated block with that of chloroquine, which is known to inhibit the I_K1 ion channel acutely with an IC₅₀ of 1.1 ± 0.2 µM (Rodríguez-Menchaca et al., 2008). Figure 2D shows the block of I_K1 by 1 µM pentamidine and by 1 µM chloroquine. Chloroquine outward current block was 70%, pentamidine caused a 95% block. Although inward current was less affected by both drugs, pentamidine was again more potent than chloroquine. No additive effects were seen when applying the two drugs together (not shown).

Direct pentamidine block depends on negatively charged residues in the K_IR2.1 cytoplasmic pore region

As the direct effects of pentamidine on K_IR2.1-based I_K1 reflect those observed for chloroquine (Rodríguez-Menchaca et al., 2008), we next assessed whether pentamidine is also able to enter the cytoplasmic pore region and with which amino acids it interacts. Molecular modelling was used to examine potential physical interaction sites of the K_IR2.1 channel with pentamidine (structure shown at top of Figure 3). Figure 3B and C show the preferential orientation of pentamidine within the cytoplasmic cavity. Pentamidine adopts a U-shaped binding conformation, and favourable interactions with residues E224, D259 and E299 are predicted. The carboxylate oxygen atoms of E224 from two opposing subunits form hydrogen bonds with the protons of the amino moiety of the benzamidine rings from pentamidine. Hydrogen bonds are also formed between the backbone oxygen of E299 and two nitrogen atoms of pentamidine. Furthermore, the carboxylate oxygen of D259 from subunit B interacts favourably with the amino group of one of the benzamidine rings.

Figure 3.

Docking of pentamidine within the K_IR2.1 cytoplasmic domain. Side view of 2 K_IR2.1 domains. The transmembrane regions (homology model of K_IR2.1, unpublished) and cytoplasmic domain (1U4F, crystal structure) with acidic residues E224, D259, E299 important for pentamidine block, highlighted (A). Bottom view of the preferential orientation of pentamidine in the cytoplasmic pore domain (B). Stereo view of the cytoplasmic pore domain with pentamidine plugging the pore. Residues are coloured according to domains (A = blue, B = green, C = orange, D = red) (C). All figures were prepared using the program PyMol (Delano Scientific).

Using inside-out patches, Figure 4A (‘wild type’) and B (‘K_IR2.1’) show that wild-type human K_IR2.1 I_K1 channels are very effectively blocked (95%) by 1 µM pentamidine. The mutations E224A, D259A and E299A all significantly diminish this outward current block to about 50%, whereas the F254A mutation did not affect outward block, in contrast to its effect on chloroquine-mediated block (Rodríguez-Menchaca et al., 2008).

Figure 4.

Pentamidine interacts with specific residues lining the K_IR2.1 channels pore region. Inside-out recordings of I_K1 based on human K_IR2.1 in absence and presence of 1 µM pentamidine. In panel A, are recordings from alanine substitution mutants that demonstrate loss of normal pentamidine block of outward current when residues E224, D259 and E299 are substituted with alanine. In panel B, plot of average residual I_K1 (at 50 mV) for the mutants depicted in panel A, normalized to control current levels. In panels C and D, inside-out recordings of I_K1 based on mouse K_IR2.1, −2.2 and −2.3 in the absence and presence of 1 µM pentamidine, showing reduced pentamidine sensitivity of channels composed of K_IR2.2 and K_IR2.3 tetramers. In panel E, the amino acid alignment of the cytoplasmic pore region of human K_IR2.1, −2.2 and −2.3. Amino acid residues mutations tested for pentamidine block are indicated by asterisks. Murine sequences in this region are identical to the human forms (not shown). F254 is not conserved in K_IR2.3. Identical residues are indicated in white lettering and black shading.

In the heart, I_K1 ion channels are formed by either homo- or heterotetramers of K_IR2.1, K_IR2.2 and K_IR2.3. We therefore assessed the effect of acute pentamidine block on homotetramers of murine K_IR2.2 and K_IR2.3 using the inside-out voltage clamp configuration. Figure 4C and D shows that pentamidine sensitivity was significantly reduced in homomers of K_IR2.3 compared with that observed for homomers of K_IR2.1. K_IR2.2 sensitivity for pentamidine block is in between that of K_IR2.1 and K_IR2.3 channels. Amino-acid alignment of the cytoplasmic pore region from K_IR2.1, 2.2 and 2.3 reveals complete identity for E224, D259 and E299 (Figure 4E). Overall, we conclude that cardiac I_K1, irrespective of its underlying molecular determinants, was susceptible to pentamidine-mediated acute block.

Inward rectification of I_K1 ion channels is accomplished by the binding of endogenous polyamines inside the cytoplasmic and transmembrane regions of the pore (Lu, 2004). The negatively charged amino acids E224, D259 and E299 in the cytoplasmic region of the pore are likely to be involved in the polyamine-induced rectification (Fujiwara and Kubo, 2006; Tai et al., 2009). To assess whether pentamidine block is affected by the presence of polyamines, pentamidine was applied in the presence of spermine. With low concentrations of spermine (0.1 µM), the outward component of I_K1 displayed in a bi-phasic block (Figure 5A and B), presumably due to interactions of spermine with the transmembrane region (peak 1) and with the cytoplasmic region (peak 2) (Ishihara and Ehara, 2004; Ishihara and Yan, 2007). Co-application of pentamidine largely inhibited the cytoplasmic spermine binding region (peak 2), leaving the transmembrane region relatively unaffected (Figure 5). These results indicate that pentamidine enters and blocks the I_K1 channel pore, but most likely does not penetrate as far as spermine. Thus, pentamidine-mediated I_K1 block is not dependent on the presence of the polyamine spermine.

Figure 5.

Pentamidine block is independent of spermine. Inside-out recording of mouse K_IR2.1 based I_K1 in the absence of spermine and pentamidine (control), in the presence of 0.1 µM spermine (SPM) or with the combination of 1 µM pentamidine and spermine (SPM + penta) (A). Enlargement of panel A illustrates that outward I_K1 block by 1 µM pentamidine is not different when pentamidine is added in the presence or absence of 0.1 µM spermine (averaged traces of 3, 3 and 4 patches, respectively) (B, left panel). The numbered peaks correspond with mode 1 and 2 block of I_K1 channels by spermine (Ishihara and Ehara, 2004). Conductance of these peaks was significantly diminished by pentamidine; mode 2 block seems especially sensitive to pentamidine. Values were normalized using the mode 1 conductance recorded in presence of spermine for each recording (B, right panel).

Discussion and conclusions

K_IR2.x ion channel blockers

Very few drugs that block acutely the K_IR2.1 ion channels have had their mode of action resolved so far. Block mediated by tamoxifen, 4-hydroxy-tamoxifen and raloxifene is voltage independent and has been suggested to act by interfering with the K_IR2.1-PIP₂ interaction, outside the pore region (Ponce-Balbuena et al., 2009). Chloroquine was found to inhibit the K_IR2.1 based I_K1 channel in a voltage dependent mode by plugging the cytoplasmic pore region (Rodríguez-Menchaca et al., 2008). We observed voltage-dependent pentamidine block, that is the outward component of I_K1 was more strongly affected than the inward component, and altered kinetics with respect to maximum current rise time upon hyperpolarization. In addition, our molecular docking simulations identified a binding site for pentamidine in the cytoplasmic pore domain of K_IR2.1. The binding site is flanked by negatively charged residues E224, D259 and E229, which is in agreement with our alanine substitution results. Therefore, our data suggest that pentamidine acts by plugging the conduction pore, as has already been suggested for chloroquine (Rodríguez-Menchaca et al., 2008). Both drugs seem to interact with K_IR2.1 primarily by electrostatic interactions. Our modelling on the closed state K_IR2.1 model did not provide evidence for drug interaction in the transmembrane cavity. Due to the lack of a working open state model, we cannot rule out the possibility of direct interaction of pentamidine with the transmembrane cavity for an open channel, which would result in drug-channel interaction at additional amino acid residues. The somewhat different behaviour of the inward component compared with the outward component in the different mutants with respect to pentamidine block would be compatible with this suggestion.

Of the above-mentioned drugs, only pentamidine displays specificity towards I_K1, while chloroquine, tamoxifen and raloxifene block additional cardiac ion channels (Liu et al., 1998; He et al., 2003; Liew et al., 2004; Liu et al., 2007; White, 2007). However, pentamidine is known to affect cardiac I_Kr in a ‘non-acute’ mode by interfering in forward trafficking of hERG mediated channels (Cordes et al., 2005; Kuryshev et al., 2005). With respect to K_IR2.1, we observed a decrease in protein expression levels over a relatively long timescale, and after the time at which direct channel block was evident. Expression level decreased faster in HEK-KWGF cells compared with adult cardiomyocytes, which may be explained by differences in K_IR2.1 half-life in the ectopic expression system compared with native K_IR2.1-expressing cells. However, a difference in pentamidine uptake, which is transporter dependent (Ming et al., 2009), could also account for the observed temporal differences.

Clinical use and pharmacology of pentamidine

The second-line drug pentamidine is widely used and advocated in treatment of diverse forms of leishmaniasis (WHO estimation of 12 million infections worldwide, 1.5–2 million new cases each year) and human African trypanosomiasis (WHO estimation 70 thousand infections) (Lai A Fat et al., 2002; Werbovetz, 2006). In the Western world, it has been used for the treatment of P. carinii pneumonia in HIV-positive patients, but its use is decreasing due to improved control of the primary infection in this patient group. Nevertheless, pentamidine use may increase due to the import (Zeegelaar et al., 2005) and the northward spread (Dujardin et al., 2008) of leishmaniasis in the Western world and in developing countries, as leishmaniasis becomes resistant to first-line medication (Croft et al., 2006). Intravenous application of clinical doses (4 mg·kg⁻¹) results in plasma levels of 1.5–5 µM (Sands et al., 1985; Goa and Campoli-Richards, 1987; Lidman et al., 1994), comparable with the concentrations used in our in vitro experiments and in studies on K_v11.1 (hERG) (Cordes et al., 2005; Kuryshev et al., 2005). Although pentamidine directly interacts with the I_K1 channel, block in the whole cell configuration occurs only after a short lag-time of 10 min, suggesting a relatively slow entrance of pentamidine into the cell. Indeed, pentamidine uptake in mammalian cells was defined recently as an organic cation transporter dependent process (Ming et al., 2009). In mice, rats and humans, pentamidine accumulates in tissues with a strong preference for liver and kidney, while brain tissue remains virtually devoid of the drug (Waalkes et al., 1970; Waldman et al., 1973; Donnelly et al., 1988). Neurons, like cardiac myocytes, express high levels of functional K_IR2.x ion channels. The absence of pentamidine in brain tissue may explain the absence of strong neurological disorders during pentamidine therapy.

Experimental use of pentamidine and future perspectives

In contrast to several other ion currents like I_Kr and I_Ks, the contribution of I_K1 to repolarization redundancy, known as repolarization reserve (Roden, 1998), has not been studied extensively in vivo. This is mainly due to the limited availability of specific I_K1 inhibitors applicable to large animal models. The specific I_K1 inhibitor Ba²⁺ is of limited use in in vivo studies due to reported lethality mainly caused by respiratory arrest (Roza and Berman, 1971). The I_K1 pore blockers chloroquine (Rodríguez-Menchaca et al., 2008) and pentamidine may, in principle, serve as an alternative for Ba²⁺ in experimental use. The disadvantages of chloroquine are its effects on protein degradation (Jansen et al., 2008) and direct block of other cardiac ion channels (White, 2007). Pentamidine, on the other hand, may be used as a direct I_K1 blocker, but long-term application may lead to ‘trafficking’ defects of K_v11.1 and K_IR2.1 channel proteins also. Recently, pentamidine infusion in dogs was reported to induce mild QT(c) prolongation, which was suggested to be caused by unexpectedly rapid I_Kr channel trafficking defects (Yokoyama et al., 2009). Based on our results described here, we would propose that the mild QT(c) prolongation in sinus rhythm (Yokoyama et al., 2009) and chronic AV block (our unpublished observations) dogs is caused by acute I_K1 block.

A potential clinical use for specific I_K1 blockers can be found in gain-of-function mutations in the KCNJ2 gene resulting in exacerbation of K_IR2.1-based I_K1 associated with atrial fibrillation (Xia et al., 2005; Zhang et al., 2005b; Ehrlich, 2008). As elegantly shown by El Harchi et al. (2009), mutant channels are susceptible to chloroquine inhibition, presenting new opportunities for drug-based treatments.

Irrespective of the clinical utility of I_K1 blocking drugs, target specificity is crucially required. Therefore, the molecular modelling studies on the interactions of chloroquine (Rodríguez-Menchaca et al., 2008) and pentamidine (this study) with the K_IR2.1-based I_K1 channel may promote further development of more specific I_K1 channel blockers, applicable for both in vivo experimentation and therapeutic use.

Glossary

Abbreviations:

AP

action potential

ATS1

Andersen-Tawil syndrome

HEK-KWGF

human embryonic kidney cell line stably expressing GFP-tagged inward rectifier channel protein K_IR2.1

hERG

human ether-a-go-go related gene

I_K1

inward rectifier current

I_Kr

delayed rectifier current

K_IR2.x

inward rectifier channel subunit

QT(c)

Q-T time interval on electrocardiogram corrected for heart rate

Conflict of interest

The authors state no conflict of interest.