The Action of a Novel Fluoroquinolone Antibiotic Agent Antofloxacin Hydrochloride on Human-Ether-à-go-go-Related Gene Potassium Channel

The Action of a Novel Fluoroquinolone Antibiotic AgentAntofloxacin Hydrochloride on Human-Ether-�-go-go-Related Gene

Potassium Channel

Jia Guo1*, Sheng-Na Han3,4*, Jin-Xu Liu1, Xiang-Mei Zhang1, Zhen-Sheng Hu1, Jie Shi1, Li-Rong Zhang4, Zhong-Zhong Zhao1

and Zhao Zhang1,2

1College of Life Science, Jiangsu Key Laboratory for Molecular and Medical Biotechnology, 2Jiangsu Key Laboratory for SupermolecularMedicinal Materials & Applications, Nanjing Normal University, Nanjing, China, 3Department of Pharmacology and 4Department of

Physiology, School of Medicine, Zhengzhou University, Zhengzhou, China

(Received 30 August 2009; Accepted 10 December 2009)

Abstract: The administration of certain fluoroquinolone antibacterials has recently been linked to QT interval prolongation,raising the clinical concerns over the cardiotoxicity of these agents. In this study, the effects of a novel fluoroquinolone, anto-floxacin hydrochloride (AX) on human-ether-�-go-go-related gene (HERG) encoding potassium channels and the biophysicalmechanisms of drug action were performed with whole-cell patch-clamp technique in transiently transfected HEK293 cells.The administration of AX caused voltage- and time-dependent inhibition of HERG K+ current (IHERG ⁄ MiRP1) in a concentra-tion-dependent manner but did not markedly modify the properties of channel kinetics, including activation, inactivation,deactivation and recovery from inactivation as well. In comparison with sparfloxacin (SPX), levofloxacin lactate (LVFX), thepotency of AX to inhibit HERG tail currents was the least one, with an IC50 value of 460.37 lM. By contrast, SPX was themost potent compound, displaying an IC50 value of 2.69 lM whereas LVFX showed modest potency, with an IC50 value of43.86 lM, respectively. Taken together, our data suggest that AX only causes a minor reduction of IHERG ⁄ MiRP1 at the esti-mated free plasma level.

The human-ether-�-go-go-related gene (HERG) encodes thepore-forming a-unit of the K+ channel that resembles therapid component of the delayed rectifier current IKr in car-diac myocytes and non-cardiac cells [1]. It is well establishedthat the inhibition of cardiac IKr is associated with drug-induced QT prolongation torsades de pointes arrhythmias(TdP) and sudden cardiac death. Indeed, it is now knownthat structurally diverse drugs associated with LQT syn-dromes (LQTS), including class III anti-arrhythmics, antimi-crobials, prokinetic drugs, antihistamines, antidepressantsand antipsychotic agents as well, share a common ability ofblockage on HERG K+ channels, resulting in a delay of car-diac repolarisation [2]. The fluoroquinolone class of antibac-terials is widely prescribed for the treatment of infectionsdue to their broad-spectrum antibiotic activity. However, theadministrations of certain fluoroquinolone antibacterialshave recently been linked to QT interval prolongation, rais-ing the clinical concerns over the cardiotoxicity of theseagents. For instance, sparfloxacin (SPX) and grepafloxacin(GPX) have been shown to prolong QT interval on electro-

cardiogram (ECG) at clinical doses [3–5]. In addition, theycan induce TdP, a life-threatening ventricular arrhythmia[6,7], resulting in their withdrawn from the market in mostcountries. Antofloxacin hydrochloride (AX) is a newly devel-oped fluoroquinolone antibiotic in China (fig. 1). Comparedwith other available fluoroquinolones in clinical use, AX waswell tolerated and demonstrated rapid absorption, high-serum concentration, broad tissue distribution, and a longelimination half-life in both animal and human beings froma pharmacokinetic point of view [8,9]. This study aimed toexamine the effects of AX on HERG currents heterologouslyexpressed in HEK293 cells and characteristics of the bio-physical mechanisms of drug action. In addition, SPX andLVFX were used as reference drugs to compare their inhibi-tory potency.

Methods

Molecular biology. HEK293 cells were maintained in Dulbecco’smodified Eagle’s medium (DMEM) supplemented with 10% foetalbovine serum, 2 mM L-glutamine and 1% penicillin and strepto-mycin at 37�C in an atmosphere of 95% air 5% CO2 incubator.Cells were transiently transfected with HERG cDNA, green fluo-rescent protein (GFP) cDNA and MiRP1 cDNA (hKCNE2)(kindly provided by Dr Nipavan Chiamvimonvat, University ofCalifornia, Davis, CA, USA) performed with the calcium phos-phate precipitation method (Invitrogen, Carlsbad, CA, USA).

Author for correspondence: Zhao Zhang, Jiangsu Key Laboratoryfor Molecular and Medical Biotechnology, College of Life Science,Nanjing Normal University, 1 Wenyuan Road, Nanjing, 210046,China (fax + 86 25 858913530, e-mail [email protected]).

*These authors contributed equally.

� 2010 The Authors Doi: 10.1111/j.1742-7843.2010.00550.xJournal compilation � 2010 Nordic Pharmacological Society. Basic & Clinical Pharmacology & Toxicology, 107, 643–649

GFP-positive cells were identified performed with the epifluores-cence system and studied within 24–48 hr after transfection.

Whole-cell HERG K+ currents recording. The cells were superfusedwith bath solution containing (in mM): NaCl, 140; KCl, 5.4; MgCl2,1; HEPES, 10; Glucose, 10; CaCl2, 2 (titrated to a pH of 7.4 withNaOH). Drugs were also added to this to make-up test solutions atthe final concentrations mentioned in the results section. Patch-pipettes (Sutter Instrument, borosilicate glass, Novato, CA, USA)were pulled to resistances of 2–3 MX (Narishige, Tokyo, Japan,PB-7) and fine polished to 3–4 MX (Narishige, MF-9). The internaldialysis solution contained (in mM): KCl, 140; MgCl2, 1; HEPES,10; K-ATP, 4; EGTA, 5 (titrated to a pH of 7.2 with KOH). The‘pipette-to-bath’ liquid junction potential was measured for thisfilling solution and was )2.7 mV. As this value was small, no correc-tions of membrane potential were made. HERG K+ currents wererecorded performed with the whole-cell configuration of thepatch-clamp technique (Hamill et al., 1981) at room temperature(22 € 1�C). Uncompensated capacitance currents in response tosmall hyperpolarising voltage steps were recorded for offline integra-tion as a means of measuring cell capacitance. Voltage commandsand data acquisition were controlled by pClamp (v10.0; MolecularDevices, Union City, CA, USA) software, which controlled Axon200B patch-clamp amplifier through a digitiser, DigiData 1440A(Molecular Devices). Series resistance compensation (65–80%) wasachieved in all experiments. Current records were filtered at 2 kHzand sampled at 10 kHz performed with a four-pole Bessel filter.

The voltage-clamp protocol to generate HERG K+ currents(refer as IHERG ⁄ MiRP1) was as follows: the holding potential was)80 mV and a family of voltage steps from )70 to +80 mV for 2sec., in 10-mV increments to evoke step currents, followed by are-polarisation step to )40 mV for 2 sec., to induce tail currents(stimulation frequency of 0.1 Hz). HERG tail currents were leakcorrected.

Half-maximal voltage values for IHERG ⁄ MiRP1 activation wereobtained by fitting IHERG ⁄ MiRP1 tail–voltage (I–V) relations with aBoltzmann distribution equation:

I=Imax ¼ 1=ð1þ expððV1=2 � VmÞ=SÞ ð1Þ

where I is the IHERG ⁄ MiRP1 tail amplitude following test potentialVm, Imax is the maximal IHERG ⁄ MiRP1 tail observed during the proto-col, V1 ⁄ 2 is the potential at which IHERG ⁄ MiRP1 was half-maximallyactivated and S is the maximal slope factor describing IHERG ⁄ MiRP1

activation when Vm = V1 ⁄ 2. Data from each individual experimentwere fitted by this equation to derive the steady activation curve andV1 ⁄ 2 in ‘control’ and with drugs.

Parameters describing voltage-dependent inactivation ofIHERG ⁄ MiRP1 were derived from fits to voltage–dependent availabil-ity plots with the equation same as the activation. But someparameters change: V1 ⁄ 2 the potential at which IHERG ⁄ MiRP1 washalf-maximally inactivated, and S is the maximal slope factordescribing IHERG ⁄ MiRP1 inactivation when Vm = V1 ⁄ 2.

The effects of the fluoroquinolones on peak tail current ampli-tudes were measured in order to determine the extent of block ofIHERG ⁄ MiRP1.

The following equations were used for numerical analysis andgraphical fits to data: The extent of IHERG ⁄ MiRP1 inhibition by variedconcentrations of different drugs was determined by the equation:

Fractionof block¼½1�ðIHERG�drug=IHERG�controlÞ=IHERG�min��100%

ð2Þ

where ‘Fractional block’ refers to the degree of inhibition forIHERG ⁄ MiRP1 by a given concentration of drugs; IHERG-control,IHERG-drug and IHERG-min represent the current amplitudes in theabsence and presence of a working concentration of drugs and thecurrent in the presence of lowest concentration of drug, respectively.

Concentration–response data were fitted by the logistic equationas follows through Origin for Windows 6.0 software:

y ¼ A2þ ðA1�A2Þ=ð1þ ðx=x0Þ^PÞ ð3Þ

where A1 and A2 mean the maximal and minimal inhibition, xrefers to the working concentration of drug, x0 is the IC50 at whichdrugs produce a half-maximal inhibition of the IHERG ⁄ MiRP1. And pis the slope factor for the fit.

Drugs. AX (fig. 1, Anhui Global Pharmaceutical Co., Ltd., Bengbu,China), yellow powder, was dissolved in extracellular solution toobtain a 10 or 100 mM stock solution, which was stored at )20�Cand then diluted to extracellular solution when necessary to the finalworking concentrations (range from 10)3 to 104 lM). SPX andLVFX were used as reference drugs of HERG assay. All chemicalswere obtained from Sigma–Aldrich Co. (St. Louis, MO, USA), withthe exception of KOH and NaOH (Nanjing Chemical and ReagentsCo Ltd., Nanjing, China).

Statistical analysis. Data are presented as mean € S.E.M. The statis-tical differences of paired or unpaired data were evaluated by Stu-dent’s paired t-test or one-way repeated measures ANOVA followed byq test for mean values comparison when appropriate. A p value of<0.05 was considered significant.

Results

Inhibitory effects of AX on IHERG ⁄ MiRP1.

Figure 2A shows the example traces of IHERG ⁄ MiRP1

recorded from HEK293 cells expressed with HERG andMiRP1 genes in the absence (left) and the presence (right) of100 lM AX. Under the control condition, upon a stepwisedepolarisation for 2 sec., from holding potential of )80 mVin 10-mV increments (recording protocol shown in inset),time-dependent outward currents were elicited, which acti-vated at potentials greater than )40 mV, reached peakat +10 mV, and decreased thereafter at more positive poten-tials, giving the current–voltage (I–V) relationship with typi-cal bell-shaped appearance. Following families ofdepolarisation, with applying of a constant repolarising stepto )40 mV for 2 sec., a large amplitude of outward tail cur-rents were induced, exhibiting typical profiles of HERG K+

currents. After recording current traces under control condi-tion, the cell was exposed to AX while repetitive pulse wasapplied once every 30 sec. The inhibitory effect of AX wasonset approximately 3 min., after drug exposure andachieved steady-state response after 5 min. There was alittle change in amplitude of IHERG ⁄ MiRP1 after washout,indicating an irreversible blockade. A total of eight differentconcentrations of AX from 10)3 to 104 lM were tested on

Fig. 1. Chemical structure of antofloxacin hydrochloride (left) andlevofloxacin lactate (right). A small modification was made by sub-stitution of hydrogen with an amino moiety at position 8 in antoflox-acin.

644 JIA GUO ET AL.

� 2010 The AuthorsJournal compilation � 2010 Nordic Pharmacological Society. Basic & Clinical Pharmacology & Toxicology, 107, 643–649

IHERG ⁄ MiRP1 (performed with between four and nine cells foreach concentration). The effect of AX on voltage-depen-dence of steady-state activation of IHERG ⁄ MiRP1 was analysedbefore and after drug exposure performed with Boltzmannequation mentioned in section ‘Methods’. When normalizedtail currents were plotted as a function of the proceeding testpotential, the value of half-maximal activation voltage wasslightly left shifted from )2.4 € 0.2 mV to )7.6 € 0.1 mV(p < 0.05, fig. 2B) but the slope factor was not changed (8.8versus 8.2).

Effects of AX on voltage- and time-dependence of HERG K+

channel.In order to clarify the action of AX on biophysical propertiesof HERG K+ channel, the voltage- and time-dependentblock of IHERG ⁄ MiRP1 by AX was studied in greater detailperformed with a concentration of 1 lM. Reduction ofIHERG ⁄ MiRP1 in the presence of 1 lM AX was compared at

the different potentials. As illustrated in fig. 2C and D, theinhibitory effect was most prominent at the potential rangefrom +20 to +60 mV for step currents and from +20 to+80 mV for tail currents, respectively. The extent of blockwas significantly enhanced as the membrane became moredepolarised and reached a plateau level at +20 mV, indicatinga voltage-dependent blockade. An ‘envelope of tails’ protocolwas used to investigate further the time-dependence of AXblock (inset in fig. 3). Typical traces of elicited currents in theabsence and the presence of AX are shown in fig. 3A whencells were depolarised to +20 mV for incrementing period of50–2100 msec. The extent of block at each pulse durationwas calculated and then pooled data plotted in fig. 3C. A sig-nificant block was achieved with a 50 msec., depolarisingstep after application of AX. The extent of block was furtherincreased with increasing pulse duration and maximal blockwas achieved at 1100 msec., demonstrating the inhibitoryeffect of AX on IHERG ⁄ MiRP was time-dependent.

A

B C D

Fig. 2. Effect of AX on IHERG ⁄ MiRP1 elicited by depolarising voltage pulses. (A) Representative IHERG ⁄ MiRP1 traces elicited by 2-sec., depolaris-ing voltage pulses between )70 to +80 mV in 10-mV increments (upper panel) from a holding potential of )80 mV and then by a 2-sec., repola-rising pulse of )40 mV (inset in panel A) to generate tail currents in the absence of AX. (B) Steady-state activation curves in the absence of(open square, n = 7) and in the presence of 1 lM AX (filled squares, n = 7). Normalized tail currents were displayed as a function of the pre-ceding test pulse voltages and fitted with a Boltzmann function. Symbols with error bars represent mean € S.E.M. (C,D) Summary of currentdensity–voltage relationship for step and tail of IHERG ⁄ MiRP1 from control values compared with AX application (n = 7; *,**p < 0.05 andp < 0.01 AX versus control).

A B

Fig. 3. Time-dependence of IHERG ⁄ MiRP inhibition. (A) Envelope tail currents before (left) and after (right) 1 lM AX exposure were recordedby a depolarising pulse to + 20 mV from holding potential of )80 mV with different duration followed by a repolarising pulse to )40 mV for5 sec. (shown inset in A s a protocol). (B) Mean fractional block of IHERG ⁄ MiRP (n = 5) produced by 1 lM AX during different duration ofdepolarising step to + 20 mV.

FLUOROQUINOLONE ANTIBACTERIALS AND HUMAN-ETHER-�-GO-GO-RELATED GENE K+ CURRENTS 645


Effect of AX on deactivation and recovery from inactivation ofHERG K+ channel.The effects of AX on deactivation kinetics of HERG chan-nel were examined at the voltage range of +40 to )100 mVperformed with a protocol shown in the inset of fig. 4B.The time course of the decaying tail currents was fitted witha double exponential function to obtain the fast and slowtime constants. As illustrated in fig. 4A, both fast and slowdeactivation in the presence of AX were slightly lower butthey were not significantly different from that in theabsence of AX at most of the test potentials. However, slowcomponent of time constant for channel deactivation atpotential from )130 to )120 mV were markedly slowed byAX (p < 0.05).

The rate of recovery from inactivation was also analysedby applying a mono-exponential fitting of the ascendingphase of tail currents from fig. 4C in the absence and pres-ence of AX. Fig. 4D shows that the time course of recoveryfrom inactivation for HERG K+ channel was not signifi-cantly altered by AX treatment.

Effect of AX on inactivation of HERG K+ channel.Onset rate of HERG K+ channel inactivation was furtherinvestigated performed with a three-pulse protocol (inset offig. 5A). Cells were depolarised to +20 mV for 500 mses.,from the holding potential of )80 mV (to allow the current

to fully activate and inactivate) and then a briefly (2 msec.)stepped to )80 mV (to relieve inactivation) followed by asubsequent step to test potentials ranging from )60 to+40 mV (the third pulse in the protocol) to induce there-onset of channel inactivation. The example tracesrecorded in the absence of AX are shown in fig. 5A. Out-ward tail currents in the absence and the presence of 1 lMAX were fit with a single exponential function to yield timeconstant at each test potential. As shown in fig. 5B, the timecourses of inactivation were slightly accelerated by AX treat-ment at voltages more positive than )60 mV. But, the differ-ences were not significant compared with control. Thevoltage-dependence of steady-state inactivation was assessedby a protocol illustrated in inset of fig. 5C. After a 50-msec.,pre-pulse to +20 mV, various test pulses from )140 to +50mV for 30 msec., followed by a return step to 20 mV for130 msec., were applied to the membrane. Subsequently,peak current amplitudes elicited by the second step to+20 mV were plotted as function of the voltage of 30 msec.,test pulses and the curve was fitted with a Boltzmann equa-tion. Half-maximal inactivation voltage and slope factor ofsteady-state inactivation were not significantly differentbetween control and AX application ()23.7 € 1.7 versus)26.8 € 0.7 mV with slop factor of 11.7 versus 14.7), sug-gesting no modification in voltage dependence of steady-stateinactivation of HERG channels.

A B

C D

Fig. 4. Effects of AX on deactivation and recovery from inactivation of IHERG ⁄ MiRP. (A) The time constant for fast (square) and slow (circle)components of IHERG ⁄ MiRP deactivation in the absence (open square ⁄ circle) and presence of AX (filled square ⁄ circle) obtained by a doubleexponential fit of the decaying tail current amplitude in B was plotted as a function of test potential. (B) Representative traces under the controlcondition were induced by a pulse protocol (inset) consisting of a depolarising step to +20 mV from holding potential of )80 mV, followed by abrief repolarisation of )80 mV and a 1.6 sec., proceeding test pulse to potentials between +40 and )140 mV. (C) Representative traces ofIHERG ⁄ MiRP were elicited by a pulse protocol shown in inset, to measure recovery from inactivation. Each current was obtained by depolarisa-tion to +50 mV for 200 msec., to reach a steady-state level before repolarisation to potentials varying from )100 to )20 mV. (D) Time constantof recovery from inactivation obtained by a single exponential fitting to the initial increase in tail-current amplitude was plotted against testpotential. There was no significant difference between control and AX application.

646 JIA GUO ET AL.


Comparison with LVFX and SPX.Similar to AX, at the same recording condition, step and tailcurrents of IHERG ⁄ MiRP1 were attenuated by either LVFX orSPX exposure in a concentration-dependent manneralthough with distinct potencies. The most significant inhibi-tion by 1 lM LVFX was achieved at the potential rangingfrom 0 to +40 mV for the step current and +10 to +80 mVfor the tail current; whereas, greater inhibition by 1 lM SPXwas observed at more depolarisation levels ranging from 0 to+60 mV for step current and 0 to +80 mV for tail current,respectively (data not shown). Therefore, regarding tail cur-rent, the results through logistic fitting for all tested drugsshowed that AX was the weakest compound, with an IC50

value of 460.4 lM. In contrast, SPX was the most powerfulone, displaying an IC50 value of 2.7 lM whereas LVFXshowed modest potency, with an IC50 value of 43.9 lM,respectively (fig. 6).

Discussion

In this study, the effects of novel fluoroquinolone antibioticagent AX on HERG K+ currents and its biophysical proper-

ties were described. In addition, the potential cardiotoxicitywas evaluated in comparison with that of LVFX and SPX.

Most of the drugs that induce QT prolongation have beenreported to possess the same ability to block the rapid com-ponent of the delayed rectifier K+ current (IKr) mediated byHERG K+ channel [10]. On this point, the data from ourprevious work and this study indicated that AX was noexception. Using the conventional microelectrode recordingfrom isolated papillary muscles of guinea pig, it has beendemonstrated that, with a perfusion of AX at concentrationsfrom 0.1 to 100 lM, APD90 were prolonged in a concentra-tion-dependent manner, whereas the RP, APA, Vmax andAPD50 were obviously not affected at any of the concentra-tions used. The prolongation of APD90 induced by AX wasabolished by pre-treatment with E4031 or by E4031 plusChromanol 293B, but not Chromanol 293B alone, suggest-ing that the prolongation of APD caused by AX was mainlyassociated with blockade of the rapid component of the car-diac delayed rectifier K+ current (IKr) rather than other ioncurrents, such as INa, slow component of delayed rectifierK+ current, IKs [11]. In accordance with our speculation,it was demonstrated in this study by the whole-cell

A B

C D

Fig. 5. Effects of AX on the inactivation of IHERG ⁄ MiRP. (A) Representative traces recorded by a three-pulse protocol (see inset). To assess onsetof inactivation, cells were depolarised to +20 mV for 500 msec., from holding potential of )80 mV and then briefly repolarised to )80 mV fol-lowed by a subsequent test potentials ranging from )60 to +40 mV for 500 msec. (B) The time constants were determined by fitting a singleexponential function to the inactivating current (in A) and then were plotted as a function of test potential. The time constant of onset inactiva-tion is not significantly altered by application of AX. (C) Steady-state inactivation was measured with two-pulse protocol (see inset), in which apulse step to +20 mV followed by a successive 30-msec. test to potentials ranging from )140 to +50 mV in 10-mV increments and a second testpulse to +20 mV. (D) The steady-state inactivation curves were illustrated under control (open square) condition and AX application (filledsquare). Outward current amplitudes induced by second step to +20 mV were measured and normalized values were plotted as a function ofthe preceding test potentials. By fitting the curves with a Boltzmann equation, mean values of V1 ⁄ 2 inactivation were )23.7 mV (slope 11.7) and)26.8 mV (slope 14.7) for control and AX application, respectively. No significant shift of the inactivation curve was observed (n = 8).



patch-clamp recording in HEK293 cell heterologouslyexpressed HERG. HERG K+ current (IHERG ⁄ MiRP1) wasinhibited by AX in a concentration-dependent manner.However, comparative data on LVFX or SPX, which havebeen reported to cause LQTS, obtained under identicalexperimental conditions suggest that relatively higher con-centrations of AX are required to block HERG K+ channelsexpressed in HEK293 cells. In the aspect of tail currents, theIC50 value for AX in this study was 460.37 lM while thatfor LVFX and SPX were 43.9 and 2.7 lM, respectively.Thus, relative rank order for IHERG ⁄ MiRP1 blocking abilitiesis SPX > LVFX > AX. This agrees with our previous resultsfrom APD prolonging extent in isolated papillary musclesfrom guinea pig, namely, SPX > LVFX > AX [11]. However,the values of IC50 for HERG inhibition induced by SPX andLVFX obtained in this study are less than other investiga-tions [12–14]. To our knowledge, the differences might bederived from the use of different cell line, different experi-mental conditions or co-transfection of MiRP1.

It is well known that drugs with different structures exhibitdifferent profiles on the gating properties of HERG K+

channel. Our data suggest that AX, like many other com-pounds [15,16] exhibits voltage- and time-dependence ofchannel block. A left shift in the half-maximal activationvoltage (�5 mV) was observed and pronounced inhibitionappeared at positive potentials in consistence with a require-ment for the channel to be opened for AX to block. Mean-while, it appears that AX did not act strongly to stabilise theinactivated state of the channel on the basis of the small shiftin the HERG inactivation curve with AX (fig. 5D) and the

time course of inactivation for HERG K+ channel was notsignificantly accelerated by the drug treatment. Accordingly,the observed time-dependence of IHERG ⁄ MiRP inhibition byAX in this study is consistent with either a mixed state-dependence of blockade (with components of both closedand open channel blockade) or with the presence of a veryrapidly developing component of activation-dependent inhi-bition immediately on depolarisation.

The minor difference in the chemical structure of AXcompared with LVFX, (the substitution of hydrogen with anamino moiety at position 8 in AX, fig. 1) does not changethe in vitro antibacterial activity; it instead prolongs the elim-ination half-life. AX is primarily metabolized by CYP3Ainto pharmacologically inactive demethylated, hydroxylatedand N-oxide derivatives, such as other known fluroquilonon-es [17,18]. Its average serum elimination half-life was 20.3–20.6 hr, but this interval might be prolonged in patients withrenal insufficiency [19]. Cmax of AX after single oral dose of300, 400 and 500 mg administration were 2.91, 3.53,4.32 mg ⁄ ml [8], equivalent to 7.05, 8.55, 10.47 lM, respec-tively, yielding the ratio of 65.3, 53.85, 43.97 between IC50

observed for HERG blockade and maximal effective plasmaconcentration under normal conditions. With regard to therelationship among the HERG K+ current blockade, QTprolongation and TdP arrhythmia, Redfern et al. [20] haveprovisionally recommended a thirty times safety margin. Ourpresent results revealed that AX would appear to have agreater safety margin than LVFX or SPX (460.37 versus 43.9for AX and LVFX; 460.37 versus 2.7 for AX and SPX,respectively). These findings suggest that serological levels ofAX might be out of the range expected to be associated withTdP arrhythmia, which is consistent with the data from pre-clinic trials reported by Xiao et al. [9].

Our present results and data from previous studies wereobtained with heterologously expression system and isolatedcardiac preparation in accordance with ICH S7B guidelines(http://www.ich.org). However, we should be aware of thelimitations of our studies to differentiate pro-arrhythmicdrugs from non-arrhythmic drugs. For instance, with HERGassays, there was no direct evidence to indicate the effects ofAX on other ion channels natively expressed in cardiomyo-cytes, including INa, ICa-L [21], and IKs [13] (although bothIKr and IKs blockers have been used in previous studies withisolated cardiac preparation). In addition, the data obtainedfrom preparation of papillary muscle not only show less sen-sitivity to detect drug-induced changes in the APD than theone from rabbit isolated Purkinje fibres [22], but also no con-cern with regard to transmural dispersion of repolarisation[23]. Therefore, in order to have a better estimation of possi-ble cardiotoxicity of AX in clinical use, integration of datafrom HERG assays with measurement of dispersion of repo-larisation and incidence of EADs by the use of Purkinjefibres and ⁄ or left ventricular wedge preparations should beconsidered in future investigations.

In conclusion, therapeutic concentration of AX would notexpect cardiotoxity because of the low potency of inhibitoryIHERG ⁄ MiRP. Although administration of AX only causes a

Fig. 6. Comparison of concentration-dependent blockade ofIHERG ⁄ MiRP1. Normalised tail current of IHERG ⁄ MiRP1 was plotted asthe function of compounds at the tested concentrations. Mean datawere fitted with the equation 2 in section ‘Methods’, giving a half-maximal inhibitory concentration (IC50) of 460.37 lM with slopefactor of 0.58 for tail currents at +20 mV. The potency ranking ofthe compounds as IHERG ⁄ MiRP1 inhibitors was SPX > LVFX > AX.(n = 7–10 for AX, n = 5–9 for LVFX and n = 4–9 for SPX at eachtested concentration, respectively; *,**p < 0.05 and p < 0.01 for AXversus LVFX; #,##p < 0.05 and p < 0.01 for AX versus SPX).

648 JIA GUO ET AL.


minor reduction of HERG K+ currents expressed inHEK293 cells at the estimated free plasma level, specialattention should be paid to those patients with other riskfactors, such as hereditary LQT syndromes, hyperkaliaemia,or elevated plasma levels of compounds induced by alter-ation in their pharmacokinetics, etc.

AcknowledgementsThis project was supported by National Hi-Tech Research

and Development Program (863 projects, No.2002AA2Z3100) and partly supported by National NaturalScience Foundation (No. 30570662, No. 30871228) fromMinistry of Science and Technology of P.R. China and Qual-ified Personnel Foundation of Nanjing Normal University.We thank Dr Nipavan Chiamvimonvat (University of Cali-fornia, Davis, USA) for providing us the plasmids of HERGand MiRP1.

Conflicts of InterestAuthors state no conflicts of interest.

References

1 Sanguinetti MC, Tristani-Firouzi M. hERG potassium channelsand cardiac arrhythmia. Nature 2006;440:463–9.

2 Drici MD, Barhanin J. Cardiac K+ channels and drug-acquiredlong QT syndrome. Therapie 2000;55:185–93.

3 Morganroth J, Hunt T, Dorr MB, Magner D, Talbot GH. Thecardiac pharmacodynamics of therapeutic doses of sparfloxacin.Clin Ther 1999;21:1171–81.

4 Morganroth J, Hunt T, Dorr MB, Magner D, Talbot GH. Theeffect of terfenadine on the cardiac pharmacodynamics of spar-floxacin. Clin Ther 1999;21:1514–24.

5 Lipsky BA, Baker CA. Fluoroquinolone toxicity profiles: a reviewfocusing on newer agents. Clin Infect Dis 1999;28:352–64.

6 Takahara A, Sugiyama A, Satoh Y, Iwasaki H, Nakamura Y,Hashimoto K. Cardiovascular profile of the canine torsades depointes arrhythmia model assessed by echocardiographic and hae-modynamic methods. Basic Clin Pharmacol Toxicol 2007;101:35–40.

7 Thomsen MB, Volders PG, Beekman JD, Matz J, Vos MA. Beat-to-Beat variability of repolarization determines proarrhythmicoutcome in dogs susceptible to drug-induced torsades de pointes.J Am Coll Cardiol 2006;48:1268–76.

8 Xiao Y, Lu Y, Kang Z, Hou F, Wang S, Li T et al. Tolerabilityand pharmacokinetics of antofloxacin hydrochloride after multi-ple oral dose administration in healthy Chinese male volunteers.Int J Clin Pharmacol Ther 2008;46:172–9.

9 Xiao Y, Lu Y, Kang Z, Zhang M, Liu Y, Zhang M et al. Pharma-cokinetics of antofloxacin hydrochloride, a new fluoroquinolone

antibiotic, after single oral dose administration in Chinese healthymale volunteers. Biopharm Drug Dispos 2008;29:167–72.

10 Yap YG, Camm J. Risk of torsades de pointes with non-cardiacdrugs. Doctors need to be aware that many drugs can cause qtprolongation. BMJ 2000;320:1158–9.

11 Shang SY, Han SN, Jing Y, Cheng SM, Qiao P, Zhang Z. Prolon-gation of cardiac action potential duration induced by antofloxa-cin in isolated guinea pig papillary muscles. Bas Clin Med2008;28:752–755. (Abstract in English).

12 Bischoff U, Schmidt C, Netzer R, Pongs O. Effects of fluoroqui-nolones on HERG currents. Eur J Pharmacol 2000;406:341–3.

13 Kang J, Wang L, Chen XL, Triggle DJ, Rampe D. Interactions ofa series of fluoroquinolone antibacterial drugs with the humancardiac K+ channel HERG. Mol Pharmacol 2001;59:122–6.

14 Zunkler BJ, Claassen S, Wos-Maganga M, Rustenbeck I, Holzg-rabe U. Effects of fluoroquinolones on HERG channels and onpancreatic beta-cell ATP-sensitive K+ channels. Toxicology2006;228:239–48.

15 Alexandrou AJ, Duncan RS, Sullivan A, Hancox JC, LeishmanDJ, Witchel HJ et al. Mechanism of hERG K+ channel blockadeby the fluoroquinolone antibiotic moxifloxacin. Br J Pharmacol2006;147:905–16.

16 Duncan RS, McPate MJ, Ridley JM, Gao Z, James AF, LeishmanDJ et al. Inhibition of the HERG potassium channel by the tricy-clic antidepressant doxepin. Biochem Pharmacol 2007;74:425–37.

17 Gargallo-Viola D, Ferrer S, Tudela E, Robert M, Coll R, RoserR et al. Antibacterial activities and pharmacokinetics of E-4767and E-5065, two new 8-chlorofluoroquinolones with a 7-azetidinring substituent. Antimicrob Agents Chemother 2001;45:3113–21.

18 Gargallo-Viola D, Esteve M, Moros M, Coll R, Xicota MA, deAndres C et al. Comparative in vitro and in vivo activities of sixnew monofluoroquinolone and difluoroquinolone 3-carboxylicacids with a 7-azetidin ring substituent. Antimicrob Agents Che-mother 1990;34:2318–26.

19 Pang XY, Liu L, Zhang DM, Wang GJ, Xie L, Liu XD. Thepharmacokinetics of antofloxacin in renally impaired rats.J Pharm Pharmacol 2008;60:667–70.

20 Redfern WS, Carlsson L, Davis AS, Lynch WG, MacKenzie I,Palethorpe S et al. Relationships between preclinical cardiac elec-trophysiology, clinical QT interval prolongation and torsade depointes for a broad range of drugs: evidence for a provisionalsafety margin in drug development. Cardiovasc Res 2003;58:32–45.

21 Zebedin V, Koenig X, Radenkovic M, Pankevych H, Todt H, Fre-issmuth M et al. Effects of duramycin on cardiac voltage-gatedion channels. Naunyn Schmiedeberg’s Arch Pharmacol2008;377:87–100.

22 Lu HR, Vlaminckx E, Van de Water A, Rohrbacher J, HermansA, Gallacher DJ. In vitro experimental models for the risk assess-ment of antibiotic-induced QT prolongation. Eur J Pharmacol2007;577:222–32.

23 Wang D, Patel C, Cui C, Yan GX. Preclinical assessment of drug-induced proarrhythmias: role of the arterially perfused rabbit leftventricular wedge preparation. Pharmacol Ther 2008;119:141–51.



Documents

The Action of a Novel Fluoroquinolone Antibiotic Agent Antofloxacin Hydrochloride on Human-Ether-à-go-go-Related Gene Potassium Channel