Skip Navigation

Cardiovascular Research 2000 48(1):44-58; doi:10.1016/S0008-6363(00)00155-3
© 2000 by European Society of Cardiology
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow E-letters: Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when E-letters are posted
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Walker, B. D
Right arrow Articles by Campbell, T. J
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Walker, B. D
Right arrow Articles by Campbell, T. J
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

Copyright © 2000, European Society of Cardiology

Comparative effects of azimilide and ambasilide on the human ether-a-go-go-related gene (HERG) potassium channel

Bruce D Walkera, Cameron B Singletona, Hui Tiea, Jane A Bursilla, Ken R Wysea, Stella M Valenzuelab, Samuel N Breitb and Terence J Campbella,*

aDepartment of Medicine, University of New South Wales and Victor Chang Cardiac Research Institute, St Vincent's Hospital, Sydney, Australia
bCentre for Immunology, St. Vincent's Hospital, Sydney, Australia

* Corresponding author. Tel.: +61-2-8382-2352; Fax: +61-2-8382-2794 t.campbell{at}unsw.edu.au

Received 19 January 2000; accepted 29 May 2000


   Abstract
Top
 Abstract
1 Introduction
 2 Methods
 3 Results
 4 Discussion
 References
 
Objective: To evaluate the effects of azimilide and ambasilide on the biophysical properties of the human-ether-a-go-go-related (HERG) channel. Methods: HERG was stably transfected into Chinese hamster ovary (CHO-K1) cells and currents were measured using a whole cell, voltage-clamp technique. Results: Azimilide had a ‘dual effect’, inhibiting current at voltage steps above –40 mV and augmenting current at –40 and –50 mV. Tail current inhibition following a step to +30 mV did not vary with temperature (IC50 610 nM at 22°C and 560 nM at 37°C). The agonist effect at –50 mV was concentration-dependent and correlated with a hyperpolarizing shift in the V1/2 of activation (r = 0.98, P<0.05). Time constants of inactivation were faster and there was a –10 mV shift in the V1/2 of steady state inactivation suggestive of open and inactivated state binding. By comparison, ambasilide inhibited HERG channels with lower potency (IC50 3.6 µM), in a voltage- and time-dependent but frequency-independent manner (0.03–1 Hz). Ambasilide had no effect on activation or inactivation gating but prolonged both fast and slow components of deactivation consistent with unbinding from the open state. The net effect of both drugs was similar during a voltage ramp which simulated a cardiac action potential. Conclusions: Inhibition of HERG channels by azimilide and ambasilide exhibits a similar time and voltage-dependence. While both exhibit affinity for the open state, azimilide also binds to inactivated channels.

KEYWORDS Antiarrhythmic agents; Ion channels; K-channel; Long QT syndrome; Membrane currents; Repolarization


   1 Introduction
Top
 Abstract
 1 Introduction
2 Methods
 3 Results
 4 Discussion
 References
 
Class III antiarrhythmic drugs prolong repolarization principally via inhibition of the rapidly activating component of the delayed rectifier channel (IKr). The reverse frequency-dependent prolongation of action potential duration (APD) and effective refractory period (ERP), exhibited by a number of these agents may reduce efficacy at fast heart rates as well as predispose to QT prolongation and torsades de pointes at slow rates. There is no doubt that proarrhythmia is a major clinical concern with class III agents (as highlighted by the SWORD study [1]) and it has been postulated that multichannel blocking drugs (e.g. amiodarone, azimilide and ambasilide) may confer greater efficacy and safety [2,3].

Azimilide is an antiarrhythmic drug with predominant class III activity, which is chemically unrelated to the widely-studied methanesulphonanilide IK-blocking drugs. It is an equipotent inhibitor of IKs and IKr, and has significant activity against other cardiac ion channels (Table 1). In animal models, azimilide has well-established efficacy in the treatment of canine ischaemic and post-myocardial infarction ventricular tachycardia and fibrillation [4–6] as well as atrial flutter and vagally-induced atrial fibrillation [7,8]. Azimilide is now being evaluated in several large clinical trials for the treatment of supraventricular arrhythmias, as well as prevention of sudden cardiac death in post-MI patients with impaired left ventricular function [2,9,10].


View this table:
[in this window]
[in a new window]

  		Table 1 Inhibition of cardiac ion channels by azimilide and ambasilidea

 
Ambasilide is an aminobenzoic acid derivative with a chemical structure similar to E-4031, which also lacks a methanesulphonamide group [11,12]. It prolongs APD and ERP in the canine, rabbit and human atrium [11,13,14] as well as in the canine, guinea pig and human ventricle [15–18]. These effects are independent of stimulation frequency and extracellular [K+] [11,13,15–17]. The cellular mechanism for this activity is block of IKr and IKs, as well as less potent inhibition of other cardiac K+ channels (Table 1).

Human IKr is encoded by the human ether-a-go-go-related gene (HERG) [19] and mutations of this gene may cause the congenital long QT syndrome [20]. In this study, we compared the effects of these newer class III antiarrhythmic agents on the HERG channel. Initial experiments focused on the effects of azimilide at room and physiological temperature, with particular attention to the kinetics of channel binding, modulation of channel gating and of the inactivated state. Since the effects of ambasilide on human IKr have not previously been described, we next examined the frequency and state-dependence of HERG channel inhibition by ambasilide.


   2 Methods
Top
 Abstract
 1 Introduction
 2 Methods
3 Results
 4 Discussion
 References
 
This investigation conforms with 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). A detailed summary of experimental methods has previously been published [21].

2.1 Molecular biology
Eukaryotic expression of HERG (sequence as per [19]) was performed by directionally cloning the coding region of the HERG gene (gift from Dr. G. Robertson, Department of Physiology, University of Wisconsin Medical School, Madison, WI, USA) into the expression vector pRc/CMV (Invitrogen, San Diego, CA, USA), which also carries the G418 resistance gene. This construct was then transfected into Chinese hamster ovary (CHO-K1) cells using Lipofectamine Reagent (Gibco). Stably transfected cells were then selected with 1000 µg/ml G418 (Boehringer, Mannheim) and subcloned. Individual subclones were maintained long-term in tissue culture and used for the patch clamping experiments to be described below.

2.2 Electrophysiology
Currents were recorded from CHO-K1 cells at room temperature (22°C) and at 37°C, using a standard whole-cell patch-clamp technique. CHO-K1 cells plated on coverslips were placed at the bottom of a 0.5-ml perfusion chamber mounted on the stage of an inverted phase contrast microscope (Nikon Diaphot, Tokyo, Japan). Cells were patched using micropipettes fabricated from thin-walled borosilicate glass (Vitrex Microhematocrit Tubes, Modulohm I/S, Denmark) with a vertical pipette puller (Model 720, David Kopf Instruments, California, USA). Currents were amplified and filtered at 2 kHz with a 4-pole Bessel filter (–3dB point) using an Axopatch 1D amplifier (Axon Instruments, Foster City, CA, USA). Stimulation protocols and data acquisition were carried out using a microcomputer (IBM pentium), running commercial software and hardware (PCLAMP 6.0/DIGIDATA 1200, Axon Instruments and Scientific Solutions). Whole-cell capacitance was determined from capacitative transient decay in current recordings following voltage steps of ±10 mV from the holding potential. At least 80% series resistance compensation was achieved in all reported experiments.

2.3 Solutions and drugs
The intracellular pipette solution contained (mM): K gluconate 120, KCl 20, MgATP 1.5, EGTA 5, N-2-hydroxylethylpiperazine-N'-2-ethanesulphonic acid (HEPES) 10, adjusted to pH 7.3. The superfusion solution contained (mM): NaCl 130, KCl 4.8, MgCl2 1.2, NaH2PO4 1.2, HEPES 10, glucose 12.5, CaCl2 1.0, adjusted to a pH of 7.4. Azimilide was obtained from Procter and Gamble (Cincinatti, OH, USA) and prepared as stock solution using the above superfusion solution. Ambasilide was kindly provided by Knoll (Ludwigshafen, Germany) and prepared as stock solution in dimethylsulfoxide (DMSO), then subsequently diluted as required with superfusate (maximum final DMSO concentration of 0.1%, v/v). In preliminary experiments, we determined that 0.1% DMSO (v/v) had no effect on the parameters under study

2.4 Statistics and analysis
Current analysis was performed using the CLAMPFIT module of the PCLAMP software. Statistical analyses were performed using PRISM 2.0 (Graphpad Software, San Diego, CA, USA). Data are expressed as mean±standard error for n experiments. Unpaired t-tests were used for comparisons of two groups and a repeated measures ANOVA with post-hoc comparison of means using Bonferroni's test was used for multiple group comparisons. A P value <0.05 was considered significant.

The amplitude of the activating current was calculated as the difference between the initial current recorded just after the step depolarization and the maximum reached at the end of the step. Similarly, the tail current amplitude was recorded as the difference between the peak and steady state current after repolarization to –60 mV. In experiments where tail currents did not fully relax to steady state, the peak tail current amplitude was used in the analysis.

The voltage-dependence of current activation was determined by fitting the values of the normalised tail currents to a Boltzmann function I=1/(1+exp [(V1/2–Vt)/k]) where I represents the tail current, V1/2, the voltage at which the current was half activated, Vt, the test potential and k, the slope factor. The relationship between drug concentration and current blockade was determined by fitting values to a Hill equation after normalisation of post-drug current to control current Idrug/Icontrol=1/[1+10log (IC50–D)n] where I represents the tail current, IC50 the concentration required for 50% channel blockade, D the drug concentration and n the Hill coefficient.


   3 Results
Top
 Abstract
 1 Introduction
 2 Methods
 3 Results
4 Discussion
 References
 
3.1 Concentration- and voltage-dependence of HERG channel inhibition
We have previously reported that HERG channels stably transfected into a Chinese hamster ovary (CHO-K1) cell line produce currents with similar biophysical characteristics to HERG expressed in Xenopus oocytes and other mammalian cell lines [19,21–23]. Currents were elicited using step depolarizations to potentials between –50 and +30 mV from a holding potential of –80 mV at room temperature (22°C). In Fig. 1A and B, representative currents are demonstrated at baseline, then during steady state blockade maintained by stimulation at 0.1 Hz in the presence of 1 µM azimilide or 3 µM ambasilide. In experiments using 1 µM azimilide, current returned to 92±10% of baseline after a 10-min washout period (n = 6), whereas washout of 3 µM ambasilide was less complete (77±2% of baseline current, n = 5). Azimilide inhibited tail currents at –60 mV following a voltage step to +30 mV with an IC50 of 610 nM (95%CI 510–730 nM) and Hill coefficient of –1.0±0.1 (n = 4–10, Fig. 1C). There was no difference in the potency of current inhibition at 37°C (IC50 560 nM, 95%CI 470–670 nM, Hill coefficient –0.9±0.1, n = 4–7), hence all remaining experiments were performed at room temperature. Using an identical protocol, ambasilide inhibited tail currents less potently with an IC50 of 3.6 µM (95%CI 3.1–4.3 µM) and Hill coefficient of –0.8±0.02 (n = 4–11, Fig. 1D). The potency of block by ambasilide was voltage-dependent, since the IC50 for tail current block at –30 mV was 16.1 µM (95%CI 9.0–28.8 µM) with a Hill coefficient of –0.8±0.2 (n = 4–11).


Figure 1
View larger version (28K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 1 Currents elicited by incremental voltage steps to between –50 and +30 mV from a holding potential of –80 mV (see inset). Recordings were made at baseline and during steady state blockade with 1 µM azimilide (A) or 3 µM ambasilide (B). (C) Concentration-dependence of current inhibition by azimilide. Relative tail currents (Iaz/Icon) after a step to +30 mV were fitted with a Hill function with no difference in potency at 22°C versus 37°C (dotted line) (n = 4–10). (D) Concentration–response relationship for inhibition by ambasilide. Relative tail currents (Iamb/Icon) after a step to +30 mV were fitted with a Hill function (n = 4–11).

 
Currents elicited during step depolarizations (‘activating currents’) were inhibited by 1 µM azimilide at potentials greater than –40 mV, but were augmented at –40 and –50 mV (n = 12, Figs. 1A and 2AGo). Similarly, azimilide inhibited tail currents at voltages above –40 mV, but augmented current at more negative potentials (Fig. 2B). This ‘agonist effect’ was use-dependent, since it was only observed when cells were stimulated with 500 ms prepulses to +30 mV after application of drug. Blockade of activating and tail currents was not voltage-dependent between –30 mV and +30 mV, although there was a trend towards increased block at positive potentials (n = 12, Fig. 2C). The voltage of half-maximal activation (V1/2) was shifted from –18.7±0.1 mV to –25.6±1.5 mV (P<0.01, n = 12), whereas the slope factor did not change (8.2±0.1 in controls and 10.2±1.2 after azimilide, P = ns). Current augmentation at –50 mV was concentration-dependent, increasing by a factor of 1.9±0.2 in 100 nM azimilide, compared with 3.6±0.5 in 1 µM azimilide (P<0.05, n = 9). Fig. 2D demonstrates that there was also a strong correlation between the shift in voltage-dependence of activation and the degree of current augmentation for a given drug concentration (V1/2 –21.0±0.3 mV with 100 nM versus –25.6±1.5 mV with 1 µM azimilide, P<0.05, n = 9).


Figure 2
View larger version (24K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 2 Voltage-dependence of HERG channel modulation by azimilide. (A) I–V relationship for activating current in the presence of 1 µM azimilide. Inhibition occurred at potentials greater than –40 mV, but an agonist effect was apparent at –40 and –50 mV (n = 12). (B) Tail currents were normalised to the amplitude of the maximal control current and fitted with Boltzmann functions (–7 mV shift with azimilide, n = 12). (C) Voltage-dependence of tail current blockade. Relative current was plotted for each potential in (B). Inhibition by azimilide was independent of voltage between –30 mV and +30 mV, but an agonist effect was apparent at –40 and –50 mV compared with +30 mV (* P<0.01, n = 12). Activating currents displayed an identical voltage-dependence. (D) Correlation of V1/2 and agonist effect at –50 mV. Points fitted with a linear regression equation (V1/2=a(Iaz/Icon)+b, where a = –1.4±0.2 and b = –16.3±0.4). Numbers in brackets represent azimilide concentration (µM).

 
The I–V relationships for both activating and tail currents at baseline and after ambasilide 3 µM are shown in Fig. 3A and B. There was no shift in the voltage-dependence of activation in the presence of ambasilide (V1/2 –22.8±0.2 mV in controls, versus –23.9±0.6 mV after ambasilide, respectively, n = 8), although there was a small difference in the slope factors (9.7±0.2 versus 12.2±0.5 mV, P<0.01, n = 8). We further assessed the voltage-dependence of block by plotting relative tail current (Iamb/Icon) at each test potential. Current inhibition increased from 20±6% at –40 mV to 49±4% at +30 mV (P<0.05, n = 8, Fig. 3C).


Figure 3
View larger version (20K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 3 Voltage-dependence of channel inhibition by ambasilide. (A, B) I–V relationships for activating and tail currents at baseline and after 3 µM ambasilide. Tail currents were fitted with a Boltzmann function and there was no shift in the voltage of half-maximal activation in the presence of ambasilide (n = 8), although the slope factor increased from 9.7±0.2 to 12.2±0.5 (P<0.01, n = 8). (C) Voltage-dependence of block. Relative tail current plotted at each test potential from points in (B). Current inhibition decreased at near-threshold potentials (*, P<0.05 compared with +30 mV, n = 8).

 
3.2 Effect of pulse duration on inhibition
Using an envelope of tails protocol, peak tail currents were recorded after steps to +30 mV of increasing duration (20–400 ms) before and during exposure to azimilide 1 µM (Fig. 4A). We used small incremental time-steps based on a previous report that binding reaches steady state relatively rapidly after channel activation [24]. HERG tail currents were normalised to the maximum control current and fitted with exponential functions. The time constant of channel activation determined using this method was shortened by azimilide ({tau} 96±8 ms in controls, versus 59±7 ms after azimilide, P<0.05, n = 9, Fig. 4B) but not changed by ambasilide ({tau} 144±18 ms for controls and 103±16 ms after ambasilide, P = ns, n = 12, Fig. 4D). Inhibition by both drugs increased with pulse duration and the development of block was fitted with a single exponential function yielding time constants of 144±10 ms (n = 9, Fig. 4C) for azimilide and 79±4 ms for ambasilide (n = 6, Fig. 4E).


Figure 4
View larger version (33K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 4 Envelope of tails protocol. (A) Currents were recorded after steps to +30 mV of increasing duration before and after superfusion with 1 µM azimilide (see inset). Although not shown, an identical protocol was used in experiments with 3 µM ambasilide. (B) Peak tail currents were normalised to the maximum control current and fitted with either a single (control) or double (azimilide) exponential function. Activation was faster in the presence of azimilide (n = 9). (C) Relative current plotted as a function of pulse duration. Block increased monoexponentially with pulse duration (n = 9). (D) Normalised peak tail currents were plotted for each pulse and fitted with a single exponential function. There was no change in the time course of activation with ambasilide (n = 12). (E) Block by ambasilide increased monoexponentially with pulse duration (n = 12).

 
Since block did not reach steady state in the envelope of tails protocol, we studied channel inhibition during a prolonged (15 s) depolarization to +10 mV in the presence of drug (Fig. 5A and B). When relative current (Iaz/Icon) in the representative experiment was plotted on an expanded time scale (Fig. 5C and D), it was apparent that inhibition by azimilide and ambasilide occurred rapidly after channel opening ({tau}, 258 and 140 ms, respectively). Similar results were obtained in three different cells.


Figure 5
View larger version (22K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 5 Inhibition during a prolonged depolarization. (A) Representative traces elicited by a 15-s depolarization to +10 mV before and after addition of 1 µM azimilide. Inhibition by both drugs was maximal within the first 2 s of the depolarization with little further time-dependent decay. (B) Time course of block development by azimilide over initial 2 s (boxed area in (A)). Relative current (Iaz/Icon) was fitted by a single exponential function (dotted line). (C) Currents elicited by identical protocol in the presence of ambasilide 3 µM. (D) Block development by ambasilide occurred monoexponentially (dotted line).

 
3.3 Frequency-dependence of HERG channel block
A previous study showed that HERG channel inhibition by azimilide exhibited reverse frequency-dependence for stimulation rates between 0.04 and 1 Hz [24]. The frequency-dependence of channel blockade by 3 µM ambasilide was assessed using a protocol in which peak tail currents at –60 mV were recorded after each of a train of 0.5 s voltage steps to +30 mV, delivered at frequencies between 0.03 and 1 Hz. There was no significant change in the control current during this protocol and inhibition by ambasilide 3 µM did not differ after 2.5 min of stimulation (37±1, 36±4, 33±3 and 41±5% block with 0.03, 0.1, 0.5 and 1 Hz stimulation, respectively, n = 4– 12).

3.4 Modulation of activation and deactivation kinetics
Time constants of activation were calculated by fitting a single exponential function to activating currents in Fig. 1A. In agreement with the findings in Fig. 4, there was a significant increase in the rate of channel activation following azimilide at most potentials (Fig. 6A), whereas there was no change following ambasilide (Fig. 6B).


Figure 6
View larger version (13K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 6 Modulation of activation kinetics. (A) Time constants of activation obtained by single exponential fitting of activating currents from the protocol in Fig. 1A (steps to –40 and –50 mV excluded due to inaccurate fitting). Time constants were significantly accelerated after addition of azimilide at voltages between between –50 and +10 mV (n = 6). (B) Time constants of activation were unchanged by ambasilide (n = 7).

 
We studied the deactivation of tail currents produced by a 3.9 s pulse to +30 mV followed by a 10-s repolarization to –60 mV (Fig. 7). While azimilide did not affect tail current kinetics, both fast and slow components of deactivation were significantly slower with ambasilide ({tau} 889±50 ms and 4257±304 ms for controls versus {tau} 1437±121 ms and 7120±520 ms with ambasilide, P<0.01, n = 6, Fig. 7B). The relative contribution of the fast component of deactivation was not modified by ambasilide.


Figure 7
View larger version (12K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 7 Modification of tail current deactivation by ambasilide. (A) Tail currents at –60 mV following a step to +30 mV were fitted with double exponential functions. (B) Both fast and slow components of deactivation were significantly slower in the presence of ambasilide. (**, P<0.01, n = 6).

 
3.5 Modulation of channel inactivation
Mutations at the S631 residue significantly weaken inactivation and abolish current augmentation by azimilide [24]. We therefore expected that azimilide itself may have effects on inactivation gating and voltage-dependence. A dual pulse protocol was used to assess the time course of channel inactivation (Fig. 8A). A 500-ms depolarization to +30 mV was followed by a 40-ms hyperpolarizing step to –80 mV to relieve rapid inactivation. A second pulse to potentials between –100 and +40 mV was then used to elicit large currents which underwent rapid reinactivation. Time constants of inactivation were determined by fitting a single exponential function to these currents. Fig. 8B demonstrates that inactivation was accelerated in the presence of 1 µM azimilide at most potentials (e.g. {tau} 14.2±1.1 ms before and 11.2±0.7 ms after azimilide at –40 mV, P<0.05, n = 7). Time constants of inactivation were modestly accelerated by ambasilide at potentials between –40 and –20 mV, and not significantly affected at other potentials (data not shown, n = 9).


Figure 8
Figure 8
View larger version (49K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 8 Effect of azimilide on channel inactivation. (A) A dual-pulse protocol was used to assess the time course of inactivation (see inset). Current traces from time period represented by boxed area on inset. (B) Inactivation time constants plotted as a function of membrane potential. Inactivation was accelerated in the presence of 1 µM azimilide at voltages between –40 and 0 mV (n = 7). (C) Recovery from inactivation was determined by fitting a single exponential function to the initial ‘hook’ preceding slower deactivation of tail currents elicited by a 500 ms depolarization to +30 mV. Time constants of recovery from inactivation were only different between –50 and –30 mV (n = 12). (D) Voltage dependence of steady state inactivation was assessed using a different dual pulse protocol (see text). Current traces show time period during and shortly after the brief hyperpolarizing step. (E) Peak currents elicited by the second step to +30 mV were plotted as a function of the potential of the preceding step and fitted with Boltzmann functions. After 1 µM azimilide, V1/2 was shifted by approximately –10 mV (P<0.01, n = 9).

 
Recovery from inactivation was determined by fitting a single exponential function to the initial ‘hook’ preceding slower deactivation of tail currents at potentials between –120 and 0 mV (elicited by a 500-ms step to +30 mV). Azimilide (n = 12, Fig. 8C) had minimal, and ambasilide (n = 5, data not shown) had no discernible effect on the time constants of recovery from inactivation.

The voltage dependence of steady-state inactivation was assessed using the following protocol. Following a 1-s step to +30 mV, 20 ms pulses were applied to potentials between –120 and +40 mV, followed by a second step to +30 mV (Fig. 8D). Peak currents (measured 3 ms after the onset of the pulse to allow for capacitance artefact) elicited by the second step to +30 mV were plotted as a function of the potential of the preceding 20 ms step (Fig. 8E). After 1 µM azimilide, V1/2 shifted from –60.4±1.1 to –70.3±1.6 mV (P<0.01, n = 9), while the slope factor remained unchanged. There was no change in the voltage dependence of inactivation in the presence of ambasilide 3 µM (V1/2 and slope –58.5±1.0 mV and –18.9±0.9 in controls versus –57.3±1.3 mV and –20.5±1.1 after ambasilide, n = 11).

3.6 Voltage ramp protocol
The use of a ventricular action potential as the voltage clamp command protocol has shed more light on the in vivo characteristics of the HERG channel [25,26]. It has been shown that very similar current traces and I–V relationships are produced by a 0.2–0.5 s descending voltage ramp [25], and, for convenience, we have used this approach here. A 500-ms voltage ramp from +40 to –80 mV was used to elicit currents such as those demonstrated in Fig. 9A. At baseline, the HERG current peaked at 370 ms into the ramp, corresponding to a membrane potential of –48 mV (n = 10). After exposure to 1 µM azimilide, the current peaked at 350 ms (membrane potential –44 mV, n = 6, Fig. 9B). When relative current (derived from the data in Fig. 9B) was plotted as a function of membrane potential there was a non-significant trend towards increased block at negative membrane potentials (P = ns, Fig. 9C).


Figure 9
View larger version (27K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 9 Voltage ramp protocol. (A) A 500-ms voltage ramp was used to simulate the repolarization during a cardiac action potential. Currents were recorded at baseline and during superfusion with 1 µM azimilide. Both increased steadily reaching a peak at 370 ms (control) and 350 ms (azimilide) and then declined more rapidly at more negative voltages due to deactivation. (B) I–V relationship for ramp experiments (n = 6). Currents were corrected for cell capacitance and every tenth point was plotted for clarity. After azimilide there was a small (+4 mV) shift in peak current. (C) Relative current during voltage ramp protocol. Block was not strictly voltage-dependent, although there was a trend towards increased block at negative membrane potentials. (D) An identical voltage ramp was applied in experiments with 3 µM ambasilide. (E) I–V relationship for capacitance-corrected currents (n = 6). After the addition of 3 µM ambasilide, there was a modest (–3 mV) shift in peak current. (F) Block increased during the ramp from 7±4% at the onset of the ramp to 41±9% at the end of the ramp (P<0.01, n = 6).

 
The same ramp protocol was applied to cells at baseline and in the presence of ambasilide (Fig. 9D). There was a modest shift in peak current from –44 mV at baseline to –47 mV after 3 µM ambasilide (n = 6, Fig. 9E). As with azimilide, block increased as the ramp potential became more negative from 7±4% at the onset of the ramp to 41±9% at the end of the ramp (P<0.01, Fig. 9F).


   4 Discussion
Top
 Abstract
 1 Introduction
 2 Methods
 3 Results
 4 Discussion
References
 
Selective IKr blockade is associated with potent antifibrillatory effects in the atria, but has not proven beneficial in the prevention of sudden death in post-myocardial infarction patients [1,27,28]. Interest in multichannel blocking drugs such as azimilide and ambasilide stems from clinical studies showing that amiodarone has at least equal or greater efficacy than other antiarrhythmic drugs with a low incidence of torsades de pointes [29,30]. Preliminary studies with azimilide suggest at least equivalent efficacy to other class III antiarrhythmic drugs with a low incidence of adverse cardiac events [2,9,10]. Although not in clinical use, ambasilide is a useful investigational agent since it blocks both components of IK and has in vivo effects similar to amiodarone [13,31].

Inhibition of IKr by azimilide (IC50 0.2–1.0 µM) is more potent than block of HERG channels in Xenopus oocytes (IC50 5.2 µM), but equivalent to our findings (IC50 610 nM) [24,32–34]. Differences between the properties of the oocyte membrane and mammalian cells may account for this discrepancy [26,35]. We found that extra-cellular application of ambasilide blocked HERG channels with an IC50 of 3.6 µM, which is similar to the potency of IKr block reported in guinea pig ventricular myocytes (IC50 of 2–5.6 µM) [12,14].

Changes in temperature are known to affect the ERP and frequency-dependence of class III antiarrhythmic activity in papillary muscle preparations [36]. Azimilide clearly exhibits forward frequency-dependence at 37°C (1–4 Hz), but this effect is much less apparent at 32°C [36]. Since the temperature at which recordings are made significantly modifies HERG channel properties and the blocking potency of class III agents [26], we repeated experiments at 37°C, finding little difference in the IC50 value. As expected, drug binding by azimilide was faster at 37°C, but unbinding was slow and difficult to evaluate. We performed experiments with ambasilide at room temperature and it is possible that the inhibitory potency, frequency-dependence and binding kinetics may be different at 37°C, since the greatest effect of temperature is on HERG activation kinetics followed by inactivation and deactivation [26].

4.1 Mechanism for ‘agonist effect’ of azimilide
Azimilide augments IKs in guinea pig ventricular myocytes at low concentrations (50–100 nM), in particular at near threshold potentials [37]. A recent report described a use-dependent HERG channel agonist effect at –50 mV in the presence of 5 µM azimilide [38]. This ‘prepulse potentiation’ was lost when the same protocol was applied to HERG channels with mutations at the outer mouth (S631C) or S5-P loop regions (H587P), suggesting that the agonist site is in the extracellular domain [38]. Our studies confirmed that the agonist effect is present at membrane potentials negative to –30 mV. The mechanism appears to be a concentration-dependent, hyperpolarizing shift in the voltage-dependence of channel activation induced by azimilide. The changes in activation kinetics may also be explained by this shift in activation gating, since after correction for the shift in the activation curve, we found that only the time constants at –20 and –30 mV were significantly different from controls. Similar findings have been reported for the experimental class III agent almokalant, in which augmentation of IKr at near threshold potentials was attributed to a hyperpolarizing shift in the voltage-dependence of activation [39]. These findings provide further evidence for an interaction between the outer mouth and nearby extracellular domain and activation gating (S4 region) [38].

4.2 State-dependence of channel block
HERG channel block by azimilide was not voltage-dependent between –30 and +30 mV, consistent with previous reports of HERG and IKr inhibition in canine ventricular myocytes [24,34]. Figs. 4 and 5Go show that inhibition required channel activation and exhibited time-dependence ({tau} 258 ms). Azimilide is known to increase the rate of IKr activation in AT-1 cells [24]. In agreement, we found that time constants of channel activation obtained using the envelope of tails protocol were significantly accelerated in the presence of 1 µM azimilide, as were activation time constants obtained by direct exponential fitting of HERG currents elicited by a series of membrane depolarizations.

Since recent data suggest that the HERG channel agonist effect correlates with the degree of inward rectification [38,40], we examined the effect of azimilide on the time course and voltage-dependence of channel inactivation. There was a significant acceleration in channel inactivation at most potentials in the presence of 1 µM azimilide, but recovery from inactivation was minimally affected, and, in addition, there was a ~10-mV hyperpolarizing shift in the voltage-dependence of inactivation. This shift represents a reduction in channel availability although such an effect would be significantly attenuated by the hyperpolarizing shift in the voltage-dependence of activation. It also suggests that azimilide has affinity for the inactivated channel state. In this regard, our findings contrast with those of Busch et al. and Jiang et al., who found that azimilide did not alter the rate of channel inactivation, and that affinity for the S631A mutant (a point mutation which greatly reduces HERG channel inactivation), did not differ from that for wild-type HERG channels [24,38]. The reason for this disparity is not clear but may relate to differences in methodology or properties of the expression systems. Finally, elevation of [K+]o is known to slow inactivation kinetics and shift the steady state inactivation curve to more positive potentials, thus considerably weakening inactivation [41]. Block of HERG channels and IKr in AT-1 cells is much less potent in the presence of high [K+]o (10 mM) [24], which supports a conclusion that azimilide binds to both open and inactivated states.

Inhibition of HERG channels by ambasilide exhibited voltage-and time-dependence similar to azimilide, but was not frequency-dependent. Block was greater at positive potentials consistent with binding of a positively charged component of the drug to a site within the transmembrane electrical field. The pKa of ambasilide is 6.5 (personal communication: E. Schneider, Knoll, Milan, Italy) which means that ~13% is positively-charged at pH 7.4. The time course of HERG channel activation was not modified by ambasilide, however, time constants of deactivation were significantly prolonged in comparison with controls. This is also observed with IKr inhibition by ibutilide and flecainide as well as block of IK and the cloned human K+ channel, HK2, by quinidine [42–45]. It is thought to represent partial unbinding of drug from open channel states, allowing brief conduction before channel closure [45]. Since the kinetics and voltage dependence of inactivation were not modified, these findings suggest that ambasilide has predominant affinity for the open state.

4.3 Effect of azimilide and ambasilide during a voltage ramp
Previous studies have shown that HERG current during a voltage ramp closely resembles that produced when a ventricular action potential is used as the voltage clamp protocol [25,26,46]. Azimilide caused only a small (+4 mV) shift in the voltage of peak current, with no significant variation in the degree of blockade throughout the ramp. This 4-mV shift was probably due to the change in the voltage-dependence of steady state inactivation (–10 mV). Block developed quickly during the ramp, reaching steady state before negative potentials were reached, as predicted from the time constants for block in Figs. 4 and 5Go. Current augmentation was not observed at potentials negative to –30 mV, suggesting that this phenomenon may not be clinically relevant.

After application of ambasilide, the current was simply scaled down, but with a modest (–3 mV) shift in peak current, which is probably due to slower deactivation in the presence of drug. Block increased as the ramp progressed, consistent with the time course of open channel block described in the envelope of tails test.

4.4 Clinical relevance of HERG channel blockade
Following standard dosing, the human plasma concentration of azimilide is approximately 1 µM (personal communication: Dr. R.R. Brooks, Procter and Gamble Pharmaceuticals, [2]). At this concentration one would expect approximately 60% suppression of the HERG current, however, since the drug is 94% plasma protein bound, the concentration available to interact with the channel would be much lower. While azimilide causes QT prolongation in animal models to an equivalent or greater extent than other class III antiarrhythmic drugs [5,47], its propensity to cause QT prolongation and torsades de pointes in humans remains undefined. While an agonist effect was not apparent during a single voltage ramp it may have significance in the setting of a delayed afterdepolarization or in partially depolarized tissue (e.g. ischaemia) in which IKr potentiation may occur [38].

The human therapeutic plasma concentration of ambasilide has not been reported, however, in a dog model, it was approximately 15 µM [13]. This concentration would clearly inhibit IKr, but would also have significant effects on other cardiac K+ channels (see Table 1). The complex electrophysiological effects of ambasilide in the human atrium and ventricle are thus explained by its inhibitory effects on multiple cardiac ion channels.

In conclusion, evidence presented in this study suggests that azimilide binds to both open and inactivated states. Current augmentation at near-threshold potentials is due to a concentration-dependent shift in the voltage-dependence of activation, but does not influence the pattern of channel block during a single action potential. Like other class III agents, ambasilide binds with greatest affinity to the open channel state but does not exhibit steady-state frequency-dependence. Both drugs have similar binding kinetics and inhibitory effects during a voltage ramp.

Time for primary review 34 days.


   Acknowledgements
 
This work was supported by research grants from the National Health and Medical Research Council of Australia, National Heart Foundation of Australia, St. Vincent's Clinic Foundation and The Clive and Vera Ramaciotti Foundation. We thank Dr. Robert Brooks for his review of the manuscript prior to submission.


   References
Top
 Abstract
 1 Introduction
 2 Methods
 3 Results
 4 Discussion
 References
 

  1. Waldo A.L, Camm A.J, deRuyter H, et al. Effects of D-sotalol on mortality in patients with left ventricular dysfunction after recent and remote myocardial infarction. Lancet (1996) 348:7–12.[CrossRef][Web of Science][Medline]
  2. Karam R, Marcello S, Brooks R.R, Corey A.E, Moore A. Azimilide dihydrochloride, a novel antiarrhythmic agent. Am J Cardiol (1998) 81:40D–46D.[CrossRef][Web of Science][Medline]
  3. Singh B.N. Current antiarrhythmic drugs: an overview of mechanisms of action and potential clinical utility. J Cardiovasc Electrophysiol (1999) 10:283–301.[Web of Science][Medline]
  4. Black S.C, Butterfield J.L, Lucchesi B.R. Protection against programmed electrical stimulation-induced ventricular tachycardia and sudden cardiac death by NE-10064, a class III antiarrhythmic drug. J Cardiovasc Pharmacol (1993) 22:810–818.[Web of Science][Medline]
  5. Drexler A.P, Micklas J.M, Brooks R.R. Suppression of inducible ventricular arrhythmias by intravenous azimilide in dogs with previous myocardial infarction. J Cardiovasc Pharmacol (1996) 28:848–855.[CrossRef][Web of Science][Medline]
  6. Brooks R.R, Carpenter J.F, Miller K.E, Maynard A.E. Efficacy of the class III antiarrhythmic agent azimilide in rodent models of ventricular arrhythmia. Proc Soc Exp Biol Med (1996) 212:84–93.[CrossRef][Medline]
  7. Restivo M, Hegazy M, Caref E.B, et al. Effects of azimilide dihydrochloride on circus movement atrial flutter in the canine sterile pericarditis model. J Cardiovasc Electrophysiol (1996) 7:612–624.[Web of Science][Medline]
  8. Nattel S, Lieu L, St-Georges D. Effects of the novel antiarrhythmic agent azimilide on experimental atrial fibrillation and atrial electrophysiologic properties. Cardiovasc Res (1998) 37:627–635.[Abstract/Free Full Text]
  9. Camm A.J, Karam R, Pratt C.M. The azimilide post-infarct survival evaluation (ALIVE) trial. Am J Cardiol (1998) 81:35D–39D.[CrossRef][Web of Science][Medline]
  10. Pritchett E, Page R, Connolly S, et al. Azimilide treatment of atrial fibrillation. Circulation (1998) 98:I–633. (Abstract).
  11. Bosch R.F, Milek I.V, Popovic K, et al. Ambasilide prolongs the action potential and blocks multiple potassium currents in human atrium. J Cardiovasc Pharmacol (1999) 33:762–771.[CrossRef][Web of Science][Medline]
  12. Zhang Z.-H, Follmer C.H, Jonnalagedda S.M, Sarma F.C, Singh B.N. Effect of ambasilide, a new class III agent, on plateau currents in isolated guinea pig ventricular myocytes: block of delayed outward potassium current. J Pharmacol Exp Ther (1992) 263:40–48.[Abstract/Free Full Text]
  13. Wang J, Feng J, Nattel S. Class III antiarrhythmic drug action in experimental atrial fibrillation: differences in reverse use-dependence and effectiveness between d-sotalol and the new antiarrhythmic drug ambasilide. Circulation (1994) 90:2032–2040.[Abstract/Free Full Text]
  14. Zaza A, Malfatto G, Schwartz P.J. Effects on atrial repolarization of the interaction between K+ channel blockers and muscarinic receptor stimulation. J Pharmacol Exp Ther (1995) 273:1095–1104.[Abstract/Free Full Text]
  15. Gjini V, Korth M, Schreieck J, et al. Differential class III antiarrhythmic effects of ambasilide and dofetilide at different extracellular potassium and pacing frequencies. J Cardiovasc Pharmacol (1996) 28:314–320.[CrossRef][Web of Science][Medline]
  16. Schreieck J, Wang Y, Gjini V, et al. Differential effect of β-adrenergic stimulation on the frequency-dependent electrophysiological actions of the new class III antiarrhythmics dofetilide, ambasilide and chromanol 293B. J Cardiovasc Electrophysiol (1997) 8:1420–1430.[Web of Science][Medline]
  17. Takanaka C, Sarma J.S.M, Singh B.N. Electrophysiological effects of ambasilide (LU 47110), a novel class III antiarrhythmic agent, on the properties of isolated rabbit and canine cardiac muscle. J Cardiovasc Pharmacol (1992) 19:290–298.[Web of Science][Medline]
  18. Weyerbrock S, Schreieck J, Karch M, et al. Rate-independent effects of the new class III antiarrhythmic drug ambasilide on transmembrane action potentials in human ventricular endomyocardium. J Cardiovasc Pharmacol (1997) 30:571–575.[CrossRef][Web of Science][Medline]
  19. Trudeau M.C, Warmke J.W, Ganetsky B, Robertson G.A. HERG, a human inward rectifier in the voltage-gated potassium channel family. Science (1995) 269:92–95.[Abstract/Free Full Text]
  20. Curran M.E, Splawski I, Timothy K.W, et al. A molecular basis for cardiac arrhythmia: HERG mutations cause long QT syndrome. Cell (1995) 80:795–803.[CrossRef][Web of Science][Medline]
  21. Walker B.D, Valenzuela S.M, Singleton C.B, et al. Inhibition of HERG channels stably expressed in a mammalian cell line by the antianginal agent perhexiline maleate. Br J Pharmacol (1999) 127:243–251.[CrossRef][Web of Science][Medline]
  22. Sanguinetti M.C, Jiang C, Curran M.E, Keating M.T. A mechanistic link between an inherited and an acquired cardiac arrhythmia: HERG encodes the IKr potassium channel. Cell (1995) 81:299–307.[CrossRef][Web of Science][Medline]
  23. Snyders D.J, Chaudhary A. High affinity open channel block by dofetilide of HERG expressed in a human cell line. J Pharmacol Exp Ther (1996) 49:949–955.
  24. Busch A.E, Eigenberger B, Jurkiewicz N.K, et al. Blockade of HERG channels by the class III antiarrhythmic azimilide: mode of action. Br J Pharmacol (1998) 123:23–30.[CrossRef][Web of Science][Medline]
  25. Hancox J.C, Levi A.J, Witchel H.J. Time course and voltage dependence of expressed HERG current compared with the native ‘rapid’ delayed rectifier K current during the cardiac action potential. Pflugers Arch (1998) 436:843–853.[CrossRef][Web of Science][Medline]
  26. Zhou Z, Gong Q, Ye B, et al. Properties of HERG channels stably transfected in HEK293 cells studied at physiological temperature. Biophys J (1998) 74:230–241.[Web of Science][Medline]
  27. Singh B.N. Antiarrhythmic drugs: a reorientation in light of recent developments in the control of disorders of rhythm. Am J Cardiol (1998) 81:3D–13D.[CrossRef][Web of Science][Medline]
  28. Torp-Pedersen C, Moller M, Bloch-Thomsen P.E, et al. Dofetilide in patients with congestive heart failure and left ventricular dysfunction. New Engl J Med (1999) 341:857–865.[Abstract/Free Full Text]
  29. Kodama I, Kamiya K, Toyama J. Cellular electrophysiology of amiodarone. Cardiovasc Res (1997) 35:13–29.[Free Full Text]
  30. Hohnloser S, Klingenheben T, Singh B.N. Amiodarone associated proarrhythmic effects: a review with special reference to torsade de pointes tachycardia. Ann Intern Med (1994) 121:529–535.[Abstract/Free Full Text]
  31. Stramba-Badiale M, Pessano P, Kirchengast M, Schwartz P.J. Effects of the potassium channel blocking agent ambasilide on ventricular arrhythmias induced by acute myocardial ischemia and sympathetic activation. Am Heart J (1995) 129:549–556.[CrossRef][Web of Science][Medline]
  32. Conder M.L, Smith M.A, Atwal K.S, McCullough J.R. Effects of NE-10064 on K+ currents in cardiac cells. Biophys J (1994) 66:A326. Abstract.
  33. Fermini B, Jurkiewicz N.K, Jow B, et al. Use-dependent effects of the class III antiarrhythmic agent NE-10064 (azimilide) on cardiac repolarization: block of delayed rectifier potassium and L-type calcium currents. J Cardiovasc Pharmacol (1995) 26:259–271.[Web of Science][Medline]
  34. Gintant G. Azimilide causes reverse rate-dependent block while reducing both components of delayed rectifier current in canine ventricular myocytes. J Cardiocvasc Pharmacol (1998) 31:945–953.[CrossRef]
  35. Dumaine R, Roy M.-L, Brown A.M. Blockade of HERG and Kv1.5 by ketoconazole. J Pharmacol Exp Ther (1998) 286:727–735.[Abstract/Free Full Text]
  36. Baskin E.P, Lynch J.J. Differential atrial versus ventricular activities of class III potassium channel blockers. J Pharmacol Exp Ther (1998) 285:135–142.[Abstract/Free Full Text]
  37. Davies M.P, Freeman L.C, Kass R.S. Dual actions of the novel class III antiarrhythmic drug NE-10064 on delayed potassium channel currents in guinea pig ventricular and sinoatrial node cells. J Pharmacol Exp Ther (1996) 276:1149–1154.[Abstract/Free Full Text]
  38. Jiang M, Dun W, Fan J.-S, Tseng G.-N. Use-dependent ‘agonist’ effect of azimilide on the HERG channel. J Pharmacol Exp Ther (1999) 291:1324–1336.[Abstract/Free Full Text]
  39. Carmeliet E. Use-dependent block and use-dependent unblock of the delayed rectifier K+ current by almokalant in rabbit ventricular myocytes. Circ Res (1993) 73:857–868.[Abstract/Free Full Text]
  40. Yao J.-A, Fan J.-S, Brooks R, Tseng G.-N. Outer mouth mutations affect azimilide's agonist action on hERG. Circulation (1997) 96:I–425. Abstract.
  41. Wang S, Morales M.J, Liu S, Strauss H.C, Rasmussen R.L. Modulation of HERG affinity for E-4031 by [K+]o and C-type inactivation. FEBS Lett (1997) 417:43–47.[CrossRef][Web of Science][Medline]
  42. Yang T, Snyders D.J, Roden D.M. Ibutilide, a methanesulfonanilide antiarrhythmic, is a potent blocker of the rapidly activating delayed rectifier K+ current (IKr) in AT-1 cells. Circulation (1995) 91:1799–1806.[Abstract/Free Full Text]
  43. Furukawa T, Tsujimura Y, Kitamura K, Tanaka H, Habuchi Y. Time- and voltage-dependent block of the delayed K+ current by quinidine in rabbit sinoatrial and atrioventricular nodes. J Pharmacol Exp Ther (1989) 251:756–763.[Abstract/Free Full Text]
  44. Follmer C.H, Cullinan C.A, Colatsky T.J. Differential block of cardiac delayed rectifier current by class Ic antiarrhythmic drugs: evidence for open channel block and unblock. Cardiovasc Res (1992) 26:1121–1130.[Abstract/Free Full Text]
  45. Snyders D.J, Knoth K.M, Roberds S.L, Tamkun M.M. Time-, voltage- and state-dependent block by quinidine of a cloned human potassium channel. Mol Pharmacol (1991) 41:322–330.[Web of Science]
  46. Spector P.S, Curran M.E, Zou A, Keating M.T, Sanguinetti M.C. Fast inactivation causes rectification of the IKr channel. J Gen Physiol (1996) 107:611–619.[Abstract/Free Full Text]
  47. Miller K.E, Carpenter J.F, Brooks R.R. Inhibition of isoproterenol-induced tachycardia by azimilide in the isolated perfused guinea pig heart. Cardiovasc Drugs Ther (1998) 12:83–91.[CrossRef][Web of Science][Medline]
  48. Yao J.-A, Tseng G.-N. Azimilide (NE-10064) can prolong or shorten the action potential duration in canine ventricular myocytes: dependence on blockade of K, Ca and Na channels. J Cardiovasc Electrophysiol (1997) 8:184–198.[Web of Science][Medline]
  49. Busch A.E, Malloy K, Groh W.J, et al. The novel class III antiarrhythmics NE-10064 and NE-10133 inhibit IsK channels expressed in Xenopus oocytes and IKs in guinea pig cardiac myocytes. Biochem Biophys Res Commun (1994) 202:265–270.[CrossRef][Web of Science][Medline]
  50. Busch A.E, Herzer T, Takumi T, Krippeit-Drews P, Waldegger S, et al. Blockade of human IsK channels expressed in Xenopus oocytes by the novel class III antiarrhythmic NE-10064. Eur J Pharmacol (1994) 264:33–37.[CrossRef][Web of Science][Medline]
  51. Feng J, Wang Z, Li G.-R, Nattel S. Effects of class III antiarrhythmic drugs on transient outward and ultra-rapid delayed rectifier currents in human atrial myocytes. J Pharmacol Exp Ther (1997) 281:384–392.[Abstract/Free Full Text]
  52. Koidl B, Flaschberger P, Schaffer P, et al. Effects of the class III antiarrhythmic drug ambasilide on outward currents in human atrial myocytes. Naunyn Schmeideberg's Arch Pharmacol (1996) 353:226–232.[Web of Science][Medline]

Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 Circ. Res.Home page
H. Sale, J. Wang, T. J. O'Hara, D. J. Tester, P. Phartiyal, J.-Q. He, Y. Rudy, M. J. Ackerman, and G. A. Robertson
Physiological Properties of hERG 1a/1b Heteromeric Currents and a hERG 1b-Specific Mutation Associated With Long-QT Syndrome
Circ. Res., September 26, 2008; 103(7): e81 - e95.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow E-letters: Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when E-letters are posted
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Walker, B. D
Right arrow Articles by Campbell, T. J
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Walker, B. D
Right arrow Articles by Campbell, T. J
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?