Inhibition of hERG potassium channel by the antiarrhythmic agent mexiletine and its metabolite m-hydroxymexiletine

Article abstract of DOI:10.1002/prp2.160

Mexiletine is a sodium channel blocker, primarily used in the treatment of ventricular arrhythmias. Moreover, recent studies have demonstrated its therapeutic value to treat myotonic syndromes and to relieve neuropathic pain. The present study aims at investigating the direct blockade of hERG potassium channel by mexiletine and its metabolite m-hydroxymexiletine (MHM). Our data show that mexiletine inhibits hERG in a time- and voltage-dependent manner, with an IC₅₀ of 3.7 ± 0.7 lmol/L. Analysis of the initial onset of current inhibition during a depolarizing test pulse indicates mexiletine binds preferentially to the open state of the hERG channel. Looking for a possible mexiletine alternative, we show that m-hydroxymexiletine (MHM), a minor mexiletine metabolite recently reported to be as active as the parent compound in an arrhythmia animal model, is a weaker hERG channel blocker, compared to mexiletine (IC₅₀ = 22.4 ± 1.2 lmol/L). The hERG aromatic residues located in the S6 helix (Tyr652 and Phe656) are crucial in the binding of mexiletine and the different affinities of mexiletine and MHM with hERG channel are interpreted by modeling their corresponding binding interactions through ab initio calculations. The simulations demonstrate that the introduction of a hydroxyl group on the meta-position of the aromatic portion of mexiletine weakens the interaction of the drug xylyloxy moiety with Tyr652. These results provide further insights into the molecular basis of drug/hERG interactions and, in agreement with previously reported results on clofilium and ibutilide analogs, support the possibility of reducing hERG potency and related toxicity by modifying the aromatic pattern of substitution of clinically relevant compounds.

Full text of DOI:10.1002/prp2.160

ORIGINAL ARTICLE
Inhibition of hERG potassium channel by the antiarrhythmic
agent mexiletine and its metabolite m-hydroxymexiletine
Roberta Gualdani1, Francesco Tadini-Buoninsegni1, Mariagrazia Roselli2, Ivana Defrenza2,
Marialessandra Contino2, Nicola Antonio Colabufo2 & Giovanni Lentini2
1
`
Dipartimento di Chimica “Ugo Schiff”, Universita di Firenze, via della Lastruccia 3, Sesto Fiorentino, FI 50019, Italy
2
`
Dipartimento di Farmacia-Scienze del Farmaco, Universita degli Studi di Bari “A. Moro”, via Orabona 4, Bari 70125, Italy
Abstract
Keywords
Ab initio calculations, antiarrhythmic drug,
hERG channel, metabolite switch, mexiletine.
Mexiletine is a sodium channel blocker, primarily used in the treatment of ven-
tricular arrhythmias. Moreover, recent studies have demonstrated its therapeutic
value to treat myotonic syndromes and to relieve neuropathic pain. The present
study aims at investigating the direct blockade of hERG potassium channel by
mexiletine and its metabolite m-hydroxymexiletine (MHM). Our data show
that mexiletine inhibits hERG in a time- and voltage-dependent manner, with
an IC50 of 3.7 ꢀ 0.7 lmol/L. Analysis of the initial onset of current inhibition
during a depolarizing test pulse indicates mexiletine binds preferentially to the
open state of the hERG channel. Looking for a possible mexiletine alternative,
we show that m-hydroxymexiletine (MHM), a minor mexiletine metabolite
recently reported to be as active as the parent compound in an arrhythmia ani-
mal model, is a weaker hERG channel blocker, compared to mexiletine
(IC50 = 22.4 ꢀ 1.2 lmol/L). The hERG aromatic residues located in the S6
helix (Tyr652 and Phe656) are crucial in the binding of mexiletine and the dif-
ferent affinities of mexiletine and MHM with hERG channel are interpreted by
modeling their corresponding binding interactions through ab initio calcula-
tions. The simulations demonstrate that the introduction of a hydroxyl group
on the meta-position of the aromatic portion of mexiletine weakens the interac-
tion of the drug xylyloxy moiety with Tyr652. These results provide further
insights into the molecular basis of drug/hERG interactions and, in agreement
with previously reported results on clofilium and ibutilide analogs, support the
possibility of reducing hERG potency and related toxicity by modifying the aro-
matic pattern of substitution of clinically relevant compounds.
Correspondence
Roberta Gualdani, Dipartimento di Chimica
ꢀ
“Ugo Schiff”, Universita di Firenze, via della
Lastruccia 3, Sesto Fiorentino, FI 50019, Italy.
Tel: +39-(0)55-4573094; Fax: +39-(0)55-
4573142; E-mail: rgualdani@unifi.it
Funding Information
Research was supported by the Ministero
dell’Istruzione, dell’Universita‘ e della Ricerca
(PON project 2007–2013, 01_00937) and the
Ente Cassa di Risparmio di Firenze.
Received: 27 May 2015; Accepted: 4 June
2015
Pharma Res Per, 3(5), 2015, e00160,
doi: 10.1002/prp2.160
doi: 10.1002/prp2.160
Abbreviations
CHO, Chinese Hamster Ovary; HEK, human embryonic kidney 293; hERG, human
Ether-ꢀa-go-go-related gene; MHM, m-hydroxymexiletine; V1/2, voltage required for
half-maximal activation.
rapidly activating IKr and a slowly activating IKs compo-
nent (Sanguinetti and Jurkiewicz 1990).
Introduction
Potassium channels have a central role in the repolariza-
tion phase of the action potential and the control of the
cellular resting membrane potential in the heart (Tamargo
et al. 2004). The delayed rectifier K+ current IK can be
separated (on the basis of biophysical properties, pharma-
cological modulation, and molecular biology) into a
The “human Ether-ꢀa-go-go-Related Gene” (hERG)
encodes the a-subunit of the IKr potassium channel,
whereas the KCNQ1 (potassium voltage-gated channel,
KQT-like subfamily, member 1) with KCNE1 (potassium
voltage-gated channel, Isk-related family, member 1) regu-
latory subunits encodes the a-subunit of the IKs potassium
ª 2015 The Authors. Pharmacology Research & Perspectives published by John Wiley & Sons Ltd,
British Pharmacological Society and American Society for Pharmacology and Experimental Therapeutics.
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License,
which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and
no modifications or adaptations are made.
2015 | Vol. 3 | Iss. 5 | e00160
Page 1

hERG Channel Inhibition by Mexiletine and MHM
R. Gualdani et al.
channel (Barhanin et al. 1996; Sanguinetti et al. 1996b).
Polymorphisms in both hERG and KCNQ1 genes can
impair channel function and thereby lead to long QT syn-
drome (LQTS), which consists of an abnormal prolonga-
tion of the time between the Q wave and the T wave of
the heart’s electrical cycle (Sanguinetti et al. 1996a; Han-
cox et al. 2008). hERG channel may also mediate the
“acquired” (drug-induced) form of the LQTS syndrome
(Sanguinetti and Tristani-Firouzi 2006). Indeed, it is now
known that structurally different drugs (e.g., Class III anti-
arrhythmics, antibiotics, antihistamines, antidepressants,
and antipsychotic agents) are able to block the hERG
channel, causing a concomitant risk of sudden death, as a
side effect; for this reason hERG inhibition is an important
antitarget during the drug discovery process. Recent evi-
dence suggests that the pharmacological sensitivity of the
hERG channel is associated with (1) the larger inner cav-
ity, compared to other voltage-gated K+ channels, that
allows many molecules to enter and block the channel; (2)
the presence of particular aromatic amino acid residues in
the pore and in the S6 helix which are critical for high-
affinity binding of drugs that contain aromatic rings (San-
guinetti and Mitcheson 2005).
block (Brochu et al. 2006; Akinci et al. 2011; Eijkelkamp
et al. 2012).
A study on guinea pig ventricular myocytes reported
no effect of mexiletine on IKr at 30 lmol/L (Wang et al.
1996), whereas a significant block at this concentration
was observed in a later study (Mitcheson and Hancox
1997). Furthermore, a more recent article reported, on
the basis of a predictive model for hERG/drug interaction,
an estimation of IC50 ≫ 10 lmol/L for mexiletine (Roche
et al. 2002), suggesting a very low or even null effect of
the drug on the hERG channel. Taking these controversial
data into account, we decided to investigate in detail the
interaction of mexiletine with hERG channel, heterolo-
gously expressed in HEK cells.
Recently (Catalano et al. 2012), it has been shown that
a minor metabolite of mexiletine, meta-hydroxymexiletine
(MHM), has the same antiarrhythmic activity on guinea
pig isolated left atria and a more favorable toxicological
profile, when compared to mexiletine. This is the prere-
quisite property for the so-called “metabolite switch,”
namely the selection of an active metabolite as the substi-
tute for the parent compound, as long as the former has
more favorable properties compared with the latter. This
work was also intended to perform a comparative study
of the direct blockade of hERG channel by mexiletine and
its metabolite MHM.
Aside from the well-known QT-prolonging effects of
class III antiarrhythmic agents (Witchel and Hancox
2000), class I antiarrhythmic agents are also known to
prolong repolarization and QT interval, carrying a risk of
LQTS and associated pro-arrhythmia (Vaughan Williams
1984). Indeed, IKr blockade contributes to the actions of
the class Ia antiarrhythmics quinidine, procainamide, and
disopyramide (Paul et al. 2002; Ridley et al. 2003; El Har-
chi et al. 2012) and class Ib antiarrhythmic agents aprin-
dine (Horie and Yoshida 1999) and phenytoin
(Danielsson et al. 2003). IKr blockade has also been
reported for class Ic drugs such as propafenone (Windi-
sch et al. 2011) and flecainide (Paul et al. 2002).
Materials and Methods
Chemistry
m-hydroxymexiletine was prepared as previously reported
(Catalano et al. 2012). Mexiletine was synthesized modi-
fying a procedure previously applied to the preparation of
mexiletine analogs (Franchini et al. 2003). Briefly, 2,6-xy-
lyloxyacetone was obtained by Williamson etherification
of 2,6-dimethylphenol with chloroacetone under either
conventional conditions or microwave-assisted heating.
The so-obtained ketone underwent either conventional
(Bruno et al. 2006) or microwave assisted (Cavalluzzi
et al. 2013) reductive amination to give the target com-
pound (see Supporting Information for details).
Mexiletine is a class Ib antiarrhythmic drug which is
used predominantly in the treatment of ventricular
arrhythmias. Moreover, recent studies have demonstrated
its therapeutic value to treat many disorders associated
with voltage-gated sodium channel dysfunction, for
example, neuropathic pain (Challapalli et al. 2005), ery-
thromelalgia (Cregg et al. 2014), myotonic syndromes
(Conte Camerino et al. 2007; Logigian et al. 2010; Stat-
land et al. 2012), and Timothy syndrome (Gao et al.
2013). Unfortunately, mexiletine has been withdrawn
from the market in many countries (Matthews and
Hanna 2014) and attempts are made to develop a “mexi-
letine alternative.”
Maintenance of mammalian cell lines and
cell transfection
Patch-clamp studies were carried out in human embry-
onic kidney 293 (HEK) cells stably expressing hERG
channel and Chinese Hamster Ovary (CHO) cells tran-
siently expressing KCNQ1/KCNE1 channels or hERG-
F656A and Y652A mutants. For heterologous protein
expression, cells were plated in 6-well cell culture dishes
with 2-mL growth medium, 24 h before transfection.
While effective in treatment and prophylaxis of ventricu-
lar tachycardias, there is some controversial evidence that
mexiletine, especially at high doses, can cause adverse
effects such as hypotension and atrioventricular heart
2015 | Vol. 3 | Iss. 5 | e00160
Page 2
ª 2015 The Authors. Pharmacology Research & Perspectives published by John Wiley & Sons Ltd,
British Pharmacological Society and American Society for Pharmacology and Experimental Therapeutics.

R. Gualdani et al.
hERG Channel Inhibition by Mexiletine and MHM
Cells were transiently transfected using X-tremeGENE 9
transfection reagent (Roche, Roche Diagnostics, Man-
nheim, Germany), according to the protocol supplied by
the manufacturer. EGFP fluorescence was used as marker
of successful transfection. Electrophysiology studies were
performed 48–72 h after transfection.
tent calculated at the RHF/3-21G*level. All conformers
falling within a window of 5 kcal/mol above global mini-
mum were retained and submitted to RHF/3-21G* geom-
etry optimization. After removal of redundant conformers
(i.e., each conformer differing from a more stable one by
less than 5° in their corresponding dihedral values), the
single point energy content for all the remaining con-
formers were calculated at the RHF/6-31G** level. The
optimized structures were confirmed as real minima by
IR frequency calculation. The most stable conformer of
either MHMm or MEXm was then arbitrarily coupled
with the most stable conformer of either TYRm or
PHEm, and the so-obtained putative complexes under-
went geometry optimization by density function theory
(DFT) implemented in Spartan’14 with B3LYP functional
(Becke 1988) and several basis sets [6-31G*, 6-31G**, 6-
31+G*, 6-311G*, 6-311+G**, 6-311++G**, and 6-
311++G(2df,2p)] (Davidson and Feller 1986) in the gas
phase. The same geometry optimization procedures were
applied to the separated guest and host species. Where
solution calculations are concerned, water contribution to
the total energy was computed by means of the contin-
uum solvation model SM8 implemented for 6-31G*, 6-
31G**, and 6-31+G* basis sets, the only ones proposed as
reliable in Spartan’14 (Marenich et al. 2007). The highest
occupied and lowest unoccupied molecular orbital
(HOMO and LUMO, respectively) energies (EHOMO and
ELUMO, respectively) of the complexes were used to calcu-
late the corresponding electrophilicity indexes (x) (May-
nard et al. 1998; Parr et al. 1999) in agreement with the
following equations (Parthasarathi et al. 2004):
Electrophysiological recordings
Electrophysiological recordings were performed using the
whole-cell mode of the patch-clamp technique. The extra-
cellular recording solution used for patch-clamp recordings
had the following composition: 140 mmol/L NaCl,
5 mmol/L KCl, 1 mmol/L MgCl2, 2 mmol/L CaCl2,
10 mmol/L Glucose, 10 mmol/L HEPES, pH 7.4 with
NaOH. The following intracellular solution was used:
130 mmol/L KCl, 1 mmol/L MgCl2, 10 mmol/L HEPES,
10 mmol/L EGTA, 5 MgATP pH 7.2 with KOH. Solutions
were applied to the cell via a gravity-fed perfusion system
(VC-6 Six Channel Valve Controller, Warner Instruments,
Hamden, CT, USA). Patch-clamp electrode was pulled from
Sutter capillary glass (Novato, CA) on a Flaming/Brown
type puller (Sutter P-87), and fire polished to 3–4 MΩ
resistance, using a microforge (Narishige). Patch-clamp
recordings of cell cultures were carried out at room temper-
ature 48 h after transfection. For recordings, a Multiclamp
200B amplifier (Molecular Devices, Inc., Sunnyvale, CA)
and Digidata 1440 data acquisition board (Molecular
Devices, Inc.) with pCLAMP 10 software (Molecular
Devices, Inc.) were used. Series resistance was compensated
by ~60–80%. The data analysis was performed using Origin
8.0 (OriginLab Corporation, Northampton, MA). Further
information for data and statistical analysis is available in
the online supplementary information.
2
2
ðI þ AÞ
ðEHOMO þ ELUMO
Þ
x ¼
¼
;
8ðI ꢁ AÞ 8ðEHOMO ꢁ ELUMO
Þ
where I and A denote the ionization potential (IP) and
the electron affinity (EA), respectively. The interaction
energy involved in complex formation (DE) was simply
obtained by the energy of the complex (Ecomplex) sub-
tracted by the sum of energies of isolated constituents
(i.e., MEXm, MHMm, TYRm, and PHEm):
Quantum mechanical calculations
The models of 3-methoxy-2,4-dimethylphenol (MHM
moiety, MHMm), 2,6-dimethylphenyl methyl ether (mex-
iletine moiety, MEXm), p-cresol (Tyr moiety, TYRm),
and toluene (Phe moiety, PHEm) were generated from
atomic fragments incorporated into Spartan’14 (Wave-
function Inc., Irvine, CA) inner fragment library and
assuming the suggested default starting geometries. The
generated geometries were optimized by the molecular
mechanics MMFF routine offered by the software (Hal-
gren 1996) and then submitted to a systematic conforma-
tional distribution analysis using the default step sizes. All
conformers in a window of 10 Kcal/mol above the global
minimum conformer were retained. When two confor-
mers differed by dihedral values lower than 10˚, the less
stable conformer was left out. Conformers were then clas-
sified according to their ab initio gas phase energy con-
X
DE ¼ Ecomplex
ꢁ
Emoieties;
each energy being calculated at the above reported levels
(see Table S2 in the Supporting Material for details).
Results
Concentration dependence of hERG channel
block by mexiletine and MHM
To investigate the effect of mexiletine and MHM
(Fig. 1A) on hERG potassium channels, patch-clamp
ª 2015 The Authors. Pharmacology Research & Perspectives published by John Wiley & Sons Ltd,
British Pharmacological Society and American Society for Pharmacology and Experimental Therapeutics.
2015 | Vol. 3 | Iss. 5 | e00160
Page 3

hERG Channel Inhibition by Mexiletine and MHM
R. Gualdani et al.
(A)
(B)
(C)
(D)
Figure 1. (A) Structures of mexiletine and its hydroxylated metabolite m-hydroxymexiletine (MHM). (B) Representative hERG currents obtained
before (control) and after addition of 50 lmol/L mexiletine, recorded using the pulse protocol shown. The arrow indicates zero current level. (C)
Time course of mexiletine-induced hERG tail current inhibition from the same cell shown in (B). (D) Concentration–response relationship
(mean ꢀ SEM) for block of hERG tail current by mexiletine and MHM. A Hill equation fit yielded an IC50 of 3.7 ꢀ 0.7 and 22.4 ꢀ 1.2 lmol/L and
a Hill slope of 0.93 ꢀ 0.09 and 1.3 ꢀ 0.08, for mexiletine and MHM, respectively (n ≥ 5 per data point).
experiments were performed on HEK cells heterologously
expressing the hERG channel. To elicit the activating out-
ward current, a depolarizing step from a holding potential
of ꢁ80 mV to ꢁ10 mV (3 sec in duration) was applied.
Tail current was evoked by repolarizing to ꢁ50 mV for
3 sec (Fig. 1B). Successive command pulses were applied
at 10-sec intervals. In the presence of 50 lmol/L mexile-
tine, the tail current was reduced almost completely and
the IhERG block induced by mexiletine was reversible
(Fig. 1C). Fitting the concentration–response relationship
for block of hERG tail currents with a Hill equation
yielded an IC50 of 3.7 ꢀ 0.7 and 22.4 ꢀ 1.2 lmol/L and
a Hill slope of 0.93 ꢀ 0.09 and 1.3 ꢀ 0.08, for mexiletine
and MHM, respectively (Fig. 1D).
interfere with the activity of the pumps as they did not
influence the P-gp and MRP1-mediated efflux of the
probe (CalceinAM), in cell line overexpressing MDCK-
MDR1 and MDCK-MRP1 proteins. Moreover, the
consumption of ATP cell level has been also measured
to confirm the absence of P-gp and MRP1-mediated
efflux (data not shown; for details see Supporting Infor-
mation).
Voltage- and time-dependent block of hERG
channel by mexiletine and MHM
The voltage dependence of mexiletine- and MHM-
induced inhibition was measured on hERG K+ current
amplitudes, applying the protocol shown in Figure 2A.
Figure 2B shows the current–voltage relationships (I–V
curve) for currents measured at the end of the test pulse.
The hERG current has an activation threshold of
ꢁ40 mV, increases progressively with potentials up to
0 mV, but then declines at more positive voltages, as a
consequence of the inward rectification. Tail currents
were elicited by repolarization to ꢁ60 mV. Individual tail
current amplitudes, recorded before (control) or after
10 min incubation with 10 lmol/L mexiletine and MHM,
were normalized to the maximal control amplitude
and fitted with a Boltzmann function (Fig. 2C). The
voltage required for half-maximal activation (V1/2) was
slightly shifted from ꢁ17.4 ꢀ 0.6 mV (control), to
ꢁ19.6 ꢀ 1.3 mV (MHM) and ꢁ21.2 ꢀ 0.9 mV (mexile-
tine). This shift to more negative potentials might be
In order to evaluate the binding selectivity of mexile-
tine and MHM, we tested their modulatory effect on two
different cardiac targets: KCNQ1/KCNE1 channel and
Na+,K+-ATPase. In both cases, we observed little or no
effect (Fig. S2, Supporting Information) on the current
signals generated by the two transport proteins. Finally,
we studied the interaction of mexiletine and MHM with
two ATPase proteins highly expressed in biological mem-
branes, such as P-glycoprotein (P-gp) and MultiDrug
Resistance-associated protein 1 (MRP1), which affect
pharmacokinetics, efficacy, safety, or tissue levels of drugs.
Each compound has been tested in a cell line overexpress-
ing the ATPase proteins P-gp and MRP1 (MDCK-MDR1
and MDCK-MRP1 cells) in order to investigate the inter-
action of mexiletine and MHM with the selected proteins.
Both compounds, at 100 lmol/L concentration, did not
2015 | Vol. 3 | Iss. 5 | e00160
Page 4
ª 2015 The Authors. Pharmacology Research & Perspectives published by John Wiley & Sons Ltd,
British Pharmacological Society and American Society for Pharmacology and Experimental Therapeutics.

R. Gualdani et al.
hERG Channel Inhibition by Mexiletine and MHM
explained by mexiletine and MHM binding preferentially
to the channel in an open state, thus limiting a hERG
conductance increase at greater depolarized test potentials
(Smith et al. 1996).
inhibitory effect of mexiletine or MHM caused a slight
increase in the fast component of the deactivation current
(Fig. 2E), whereas the slow component (sslow) was not
affected by the presence of both compounds (Fig. S1,
Supporting Information).
Drug block of ion channels may affect protein kinetics.
For this reason we investigated a possible effect of mexile-
tine and MHM on activation or deactivation kinetics of
hERG channel. To evaluate the time course of hERG
channel activation, ionic currents were recorded using the
envelop tail protocol shown in Figure 2E (Liu et al. 1996;
Vandenberg et al. 2012). The time constants, determined
by fitting with a single exponential function the latter part
of the activation time course, were plotted against the
membrane voltage (Fig. 2E). We observed that in the
presence of 10 lmol/L mexiletine or MHM there was a
significant decrease in the activation time constant. More-
over, deactivation time constants were obtained by fitting
with a double exponential function the currents recorded
using the protocol shown in Figure 2D. In this case, the
Effect of mexiletine and MHM on hERG
channel inactivation
Steady-state inactivation currents were measured follo-
wing the protocol shown in Figure 3A: channels were
inactivated at +60 mV, before short test pulses from
ꢁ100 mV to +20 mV were applied, to recover the chan-
nels from inactivation. Depolarization to +60 mV after
these test pulses evoked a large outward inactivating cur-
rent. Current amplitudes measured after the return to
+60 mV were normalized, plotted against the voltage and
fitted by a Boltzmann function. Our data show that mex-
iletine and MHM do not cause any significant change in
(A)
(B)
(C)
(D)
(E)
Figure 2. (A) Representative hERG current traces elicited by the voltage protocol shown in the lower panel, testing steady-state activation
currents. The arrow indicates zero current level. (B, C) Normalized (with respect to the control currents) I–V relationships for current measured at
the end of the test pulse (B) and at the tail current peaks (C), before (control) and after addition (time of incubation ≥ 10 min) of 10 lmol/L
mexiletine and MHM. Error bars for some points are masked by symbols (n ≥ 5 cells per data-point). I–V relationships for tail current amplitudes
were fitted with a Boltzmann function. V1/2 shifted from ꢁ17.4 ꢀ 0.6 mV (control), to ꢁ19.6 ꢀ 1.3 mV (MHM; P < 0.05 paired Student’s t-test;
n = 6 cells) and ꢁ21.2 ꢀ 0.9 mV (mexiletine; P < 0.001, paired t-test; n = 6 cells). *P < 0.05, two-way ANOVA followed by a Bonferroni
posttest; ***P < 0.001 two-way ANOVA followed by a Bonferroni posttest. (D) Representative hERG current traces elicited by the voltage protocol
shown in the lower panel, testing deactivation currents. The arrow indicates zero current level. (E) Activation time constants were derived from
currents acquired using the envelope of tails voltage protocol shown on the right-hand side of the figure, fitted with a single exponential
function. Deactivation time course was fitted by a double exponential function and derived from currents acquired using the protocol shown in
(D). *P < 0.05, two-way ANOVA followed by a Bonferroni posttest; n ≥ 5 cells.
ª 2015 The Authors. Pharmacology Research & Perspectives published by John Wiley & Sons Ltd,
British Pharmacological Society and American Society for Pharmacology and Experimental Therapeutics.
2015 | Vol. 3 | Iss. 5 | e00160
Page 5

hERG Channel Inhibition by Mexiletine and MHM
R. Gualdani et al.
(A)
(B)
(C)
(D)
Figure 3. (A) Representative hERG current traces elicited by the voltage protocol shown in the lower panel, testing steady-state inactivation
currents. The arrow indicates zero current level. (B) Representative onset of inactivation currents, elicited by a test pulse from +60 mV, to
potentials ranging from ꢁ20 to +60 mV, in 10 mV-increments. The arrow indicates zero current level. (C) Normalized steady-state I–V inactivation
curves before (control) and after application of 10 lmol/L mexiletine and MHM. Data are expressed as mean ꢀ SEM; error bars for some points
are masked by symbols (n ≥ 5 per data point). Solid lines represent fits with Boltzmann function. The V1/2 obtained are ꢁ90.6 ꢀ 4.7 mV
(control), ꢁ90.4 ꢀ 2.1 mV (mexiletine), ꢁ89.1 ꢀ 3.5 mV (MHM). (D) Inactivation time constants were derived from currents acquired using the
protocol shown in (B), fitted with a single exponential function and plotted against the membrane potential. Data are expressed as mean ꢀ SEM
(n ≥ 5 per data point).
steady-state inactivation, as shown by the V1/2 values:
ꢁ90.6 ꢀ 4.7 mV in the absence of drugs (control),
ꢁ89.1 ꢀ 3.5 mV for mexiletine and ꢁ90.4 ꢀ 4.2 mV for
MHM (Fig. 3C). This indicates that both drugs have no
effect on the inactivation gating of the hERG channel.
Finally, the effect of mexiletine and MHM on the inacti-
vation time course was investigated using the protocol
amplitude during long depolarizing steps. Current was
activated by a 100 ms depolarization step from ꢁ80 mV
(to keep the channels in the closed state) to +80 mV,
then the cell was held at 0 mV for 1 sec (the protocol is
shown in Fig. 4A, B). This protocol is a useful strategy
to study in details the possibility of hERG close-state
block by drugs (Kiehn et al. 1996; Thomas et al. 2001).
After performing the control measurements, 10 lmol/L
mexiletine was applied for 10 min without pulsing. After
the pulse-free incubation period, the pulse protocol was
repeated. Figure 4A shows that the initial current during
the step to +80 mV and the current activation peak,
recorded in the absence (control, trace “a”) or presence
of mexiletine (trace “b”), overlap perfectly, suggesting
minimal closed-state block by mexiletine. Moreover, in
trace “b” we observe a time-dependent increase in
blockage, which can be fitted by a single exponential
function with a time constant of 1.18 ꢀ 0.11 sec. This is
shown more clearly in Figure 4B, reporting the “mexile-
tine-sensitive” current, obtained by subtracting trace “b”
from trace “a.” The trace in Figure 4B shows an
shown in Figure 3B. From
ꢁ80 mV, a 200 ms test pulse to +60 mV was applied to
inactivate the channel. Moreover, short pulse to
a holding potential of
a
ꢁ100 mV and voltage steps from ꢁ20 mV to +60 mV
(300 ms, 10 mV-increments) were applied to elicit a large
outward inactivating current. Inactivation currents were
fitted by a single exponential function to extrapolate time
constant values. Figure 3D shows that the inactivation
time course was not affected by the presence of mexiletine
and MHM.
State dependence of hERG channel block by
mexiletine
To investigate the state dependence of hERG inhibition,
we evaluated the effect of mexiletine on hERG current
exponential development with
a time constant of
0.88 ꢀ 0.08 sec, supporting the hypothesis that the
2015 | Vol. 3 | Iss. 5 | e00160
Page 6
ª 2015 The Authors. Pharmacology Research & Perspectives published by John Wiley & Sons Ltd,
British Pharmacological Society and American Society for Pharmacology and Experimental Therapeutics.

R. Gualdani et al.
hERG Channel Inhibition by Mexiletine and MHM
(A)
(B)
(C)
Figure 4. (A) Representative hERG current traces recorded before (control) or after addition of 10 lmol/L mexiletine (10 min incubation), using
the protocol shown in the lower panel. To verify the reproducibility of the experiments, each measurement was repeated n ≥ 5 times. (B)
“Mexiletine sensitive” current obtained by subtracting traces “a” and “b” shown in (A). (C) Concentration–response relationship (mean ꢀ SEM)
for block by mexiletine of hERG tail current, induced by repolarization from +80 mV to ꢁ50 mV. A Hill equation fit yielded an IC50 of
47.2 ꢀ 5.3 lmol/L and a Hill slope of 1.13 ꢀ 0.20 (n ≥ 5 per data point).
closed state of hERG channel is not targeted by
mexiletine.
42.6 ꢀ 4.6 lmol/L, and 436 ꢀ 130 lmol/L for WT,
Y652A and F656A, respectively. Thus, both mutant
channels Tyr652 and Phe656 exhibit decreased sensitivity
to block by mexiletine and MHM, particularly in the case
of Phe656. Moreover, our results suggest that Y652A
mutation affects MHM binding less than mexiletine bind-
ing. In fact the greatest fold shift in IC50 of the mutant
Y652A is observed in the presence of mexiletine (~11),
compared to MHM (~2) (Fig. 5E).
Finally we evaluated the possibility of block induced by
mexiletine when the hERG channel is in the inactivated
state. To this end, we observed that the value of maxile-
tine IC50, calculated at +80 mV (at which all channels are
inactivated), is 47.2 ꢀ 5.3 lmol/L (Fig. 4C), compared to
ꢁ10 mV (at which most channels are open) is almost 15-
fold higher than the corresponding value at ꢁ10 mV
(Fig. 1D), suggesting mexiletine preferential binding to
the open state of hERG channel, rather than the inacti-
vated state.
Ab initio calculations
It has been suggested that both Tyr652 and Phe656 may
give p–p stacking interactions with the same aromatic
portion of hERG blockers (Gemma et al. 2012) while
Tyr652 should preferentially be involved in p-cation
interactions with the protonated basic nitrogen of the
ligand (Pearlstein et al. 2003). Since mexiletine and
MHM share the same basic aliphatic moieties, we focused
our attention on the aromatic portions of the ligands, the
ones that should elicit the differences observed when
comparing their corresponding pharmacological profiles.
In particular, (1) we assumed the moieties drawn in bold
in Figure 6 (panels A and B) as representative of the
whole MHM and mexiletine molecules, (2) modelled the
corresponding truncated compounds – 3-methoxy-2,4-
dimethylphenol (MHM moiety, MHMm) and 2,6-dim-
ethylphenyl methyl ether (mexiletine moiety, MEXm),
respectively –, and (3) optimized them at the RHF/3-
21G* level (Fig. 6, panel C). The same reductionist pro-
cess (1–3) was used to represent the aromatic rings of the
two amino acids supposed to be involved in direct inter-
actions with the aromatic portions of the guest molecules.
Thus, p-cresol and toluene were assumed as models of
Tyr652 (Tyr moiety, TYRm) and Phe656 (Phe moiety,
PHEm), respectively (Fig. 6, panels A–C).
Role of S6 inner helix aromatic residues on
mexiletine binding
Most hERG channel blockers bind within the channel
inner cavity, interacting with one or two aromatic resi-
dues of the S6 domain (Tyr652 and Phe656). To evaluate
if mexiletine binds to sites within the inner cavity of
hERG channel, we recorded ionic currents through CHO
cells expressing Y652A or F656A hERG mutants, in the
presence of mexiletine. Due to the known low expression
and altered kinetic properties of F656A hERG clone, we
measured F656A hERG inward tail current using a high
[K+] external solution (94 mmol/L, Fig. 5A) (Mitcheson
et al. 2000a; Du et al. 2014); on the contrary, the Y652A
hERG outward tail current (Fig. 5B) was acquired in the
“normal” [K+] solution (5 mmol/L).
Figure 5C, D show the inhibition curves of mexiletine
and MHM for wild-type and mutant hERG channels. The
extrapolated IC50 values calculated in the presence of
mexiletine are 3.7 ꢀ 0.7 lmol/L, 41.6 ꢀ 6.6 lmol/L, and
111.5 ꢀ 1.8 lmol/L for WT, Y652A, and F656A, respec-
tively. Moreover, the extrapolated IC50 values calculated
in the presence of MHM are 22.4 ꢀ 1.2 lmol/L,
ª 2015 The Authors. Pharmacology Research & Perspectives published by John Wiley & Sons Ltd,
British Pharmacological Society and American Society for Pharmacology and Experimental Therapeutics.
2015 | Vol. 3 | Iss. 5 | e00160
Page 7

hERG Channel Inhibition by Mexiletine and MHM
R. Gualdani et al.
(C)
(A)
(D)
(B)
(E)
Figure 5. (A) Representative IhERG traces of WT hERG and F656A hERG in the absence (black line) and presence (gray line) of 50 lmol/L
mexiletine. The tail current was evoked by repolarization from +20 to ꢁ120 mV (protocol shown in the lower panel) and was recorded using
“high” [K+] solution (94 mmol/L). The arrows indicate zero current level. (B) Representative IhERG traces of WT hERG and Y652A hERG mutant in
the absence (black line) and presence (gray line) of 50 lmol/L mexiletine. The tail current was evoked by repolarization from 0 to ꢁ50 mV (same
protocol shown in the lower panel of Fig. 1B) and was recorded under the “normal” external solution ([K+] = 5 mmol/L). The arrows indicate
zero current level. (C, D) Concentration–response relationships for block of hERG channel currents by mexiletine (C) and MHM (D). Each
value represents mean ꢀ SEM of n ≥ 5 cells. (E) Summary of fold changes in IC50 values for Y652A and F656A hERG mutants compared with
WT-hERG.
To evaluate the possibility that the aromatic rings of
the ligands would display different tendencies to form
sandwich-like stacked complexes with both Tyr652 and
Phe656 (Gemma et al. 2012), the geometries of the com-
plexes MHMm/TYRm, MEXm/TYRm, MHMm/PHEm,
and MEXm/PHEm were optimized at several levels of
DFT calculation (see Materials and Methods). Since the
binding site of hERG blockers is supposed to be located
into a restricted space lined by four residues each of
Phe656 and Tyr652, calculations were performed in vac-
uum (i.e., e = 0) to simulate a highly lipophilic environ-
ment. Generally, complexes were T shaped denoting the
contribution of both hydrogen bonding and dispersion
interactions (Paton and Goodman 2009). In particular,
TYRm behaved as a hydrogen bond donor while the phe-
nolic hydroxy and methoxy groups of MHMm and
MEXm, respectively, performed as hydrogen bond accep-
tors (Fig. 6, panels D and E). MHMm and MEXm inter-
act rotated of 120° with respect to TYRm. This
apparently dramatic difference may be accepted as possi-
ble, given the reduced size of both mexiletine and MHM
compared with the relatively large inner cavity of the
channel. Furthermore, the symmetry of the latter and the
presence of four equally accessible Tyr residues let
hypothesize the possible flipping of ligand/Tyr interac-
tions within the pore. However, given the limitations of
the model, the above considerations should be considered
as mere speculations and should be assumed with caution
when predicting the actual binding orientation of the
whole binding partners. For each optimized complex, the
corresponding electrophilicity index (x) value was
assumed as a measure of the residual tendency to “soak”
electrons from a further aromatic ring as an electron
donor. Regardless of the level of calculation used, the
2015 | Vol. 3 | Iss. 5 | e00160
Page 8
ª 2015 The Authors. Pharmacology Research & Perspectives published by John Wiley & Sons Ltd,
British Pharmacological Society and American Society for Pharmacology and Experimental Therapeutics.

R. Gualdani et al.
hERG Channel Inhibition by Mexiletine and MHM
(A)
(B)
(C)
(D)
(E)
Figure 6. Structures of the model compounds used to predict the interaction energies between mexiletine and MHM with their putative binding
aromatic residues – Tyr652 and Phe656. (A) guest and host molecules presenting in bold the moieties studied by ab initio calculations; (B)
corresponding truncated molecules assumed as representative of the whole guest compounds and host aromatic residues; (C) tube representation
of the most stable conformer [B3LYP/6-311++G(2df,2p)//B3LYP/6-311++G(2df,2p)] of each model compound undergoing complex formation; (D)
complex formed by MHMm with TYRm, optimized at the above level of calculation (hydrogen bond shown as a green broken line); (E) complex
formed by MEXm with TYRm, optimized at the above level of calculation (hydrogen bond shown as a green broken line).
complexes formed by MHM showed slightly higher x val-
ues (see Table S1 in the Supporting Information); how-
ever, no significant difference was observed when the
corresponding mean values were compared. When the
energies of formation of the complexes were considered,
the complexes involving TYRm resulted significantly more
stable than the corresponding complexes formed with
PHEm (mean energies of formation of about 6–7 and 1–3
Kcal/mol, respectively; see Table S2 in the Supporting
Information), MHMm/TYRm being significantly less sta-
ble than MEXm/TYRm (mean difference = 1.2 ꢀ
0.3 Kcal/mol, P = 0.01). When the interaction with
PHEm is concerned, the complex formed with MHM
resulted more stable than the corresponding complex
formed with MEXm of about the same difference previ-
ously observed between the two complexes formed with
TYRm. However, compensation should be ruled out con-
sidering the significantly higher levering interactions with
Tyr652. Thus, the interaction of the aromatic moieties of
mexiletine and MHM with Tyr652 might explain the dif-
ference observed in their relative potency of hERG block.
Since the same interaction was proposed as a pivotal
contribution to the blocking activity of both compounds
on cardiac sodium channels (Desaphy et al. 2012) where
MHM was at least as potent as mexiletine, we hypothe-
sized that this apparently contradictory result might be
explained assuming that the binding site for the two
blocking agents is located in a relatively less lipophilic
environment in the sodium channel in comparison with
the one present in the hERG vestibule. Running our cal-
culations in the presence of water, we were gratified by
the results indicating comparable stability for MHMm/
TYRm (E ꢀ SE = ꢁ6.2 ꢀ 1.8 Kcal/mol) in respect of
MEXm/TYRm (E ꢀ SE = ꢁ6.7 ꢀ 1.8 Kcal/mol) (Table
S3 in the Supporting Information). Once again, the com-
plexes involving TYRm resulted significantly more stable
than the corresponding complexes formed with PHEm
(mean energies of formation of about 6–7 and 1–2 Kcal/
mol, respectively; see Table S4 in the Supporting Informa-
tion), thus supporting the suggestion that, when both
hydrogen bonding and dispersion interactions contribute
to binding, the Tyr aromatic ring should display a higher
levering interaction in comparison with the one involving
Phe residues. MHMm/PHEm and MEX/PHEm also
ª 2015 The Authors. Pharmacology Research & Perspectives published by John Wiley & Sons Ltd,
British Pharmacological Society and American Society for Pharmacology and Experimental Therapeutics.
2015 | Vol. 3 | Iss. 5 | e00160
Page 9

hERG Channel Inhibition by Mexiletine and MHM
R. Gualdani et al.
presented comparable stability in water (E ꢀ SE =
ꢁ0.6 ꢀ 0.7 Kcal/mol and E ꢀ SE = ꢁ1.6 ꢀ 1.1 Kcal/
mol, respectively).
channel (Sanguinetti and Mitcheson 2005; Perry et al.
2006). Mexiletine has two pharmacophoric points: the
aromatic portion that could interact with Tyr652, and a
charged group, which could form p-cationic interaction
with Phe656. Thus, considering also the analogy between
Nav1.5- and hERG-binding sites, we hypothesized that
both strictly related compounds Mex and MHM should
preferentially reside in the bottom corner of the pore, the
one lined by Tyr652 and Phe656. To evaluate this hypoth-
esis, we recorded ionic currents through Y652A or F656A
hERG mutants. Our data show that mutation of Tyr652
and Phe656 residues dramatically reduces the blocking
effect by mexiletine (Fig. 5); in particular the Phe656 resi-
due seems to play a prominent role in the binding of
mexiletine. Moreover, we observed that Y652A mutation
affects MHM binding less than mexiletine binding
(Fig. 5C and D), in accordance to ab initio calculations.
According to the modulated receptor theory for chan-
nel block (Hille 1977) and in agreement with previous
results on both sodium (Desaphy et al. 2012) and hERG
(Mitcheson et al. 2000a,b) channels, it may be assumed
that block by mexiletine and MHM occurs from the cyto-
plasmic side of the membrane. Thus, the lower potency
Discussion
This is the first report of direct blockade of heterolo-
gously expressed hERG potassium channel by the antiar-
rhythmic agent mexiletine and one of its metabolite m-
hydroxymexiletine (MHM).
We found that mexiletine is a moderate-potency inhib-
itor of hERG channel, with an IC50 of 3.7 ꢀ 0.7 lmol/L,
a value close to its therapeutic concentration (4–11 lmol/
L) (Campbell et al. 1978). Interestingly, MHM, a minor
metabolite of mexiletine, performed as a weaker hERG
blocker, compared to mexiletine, with an IC50 of
22.4 ꢀ 1.2 lmol/L.
Target specificity is critical to limit toxicity. To assess
the specific relationship between the risk of mexiletine-
induced cardiotoxicity and other potential targets for car-
diac adverse effects, we investigated the effect of mexile-
tine and MHM on the following proteins: (1) the Na+,
K+-ATPase, an ion pump of cardiac cells, which is
blocked by the action of cardiac glycosides like digoxin
used to treat heart disease; (2) the KCNQ1/KCNE1 chan-
nel, which is responsible for the slow component of the
action potential repolarization phase (IKs); (3) two AT-
Pase proteins highly expressed in biological membranes
such as P-glycoprotein (P-gp) and MultiDrug Resistance-
associated protein 1 (MRP1), which affect pharmacoki-
netics, efficacy, safety, and tissue distribution of drugs.
The absence of interaction with these antitargets (see Sup-
porting Information), together with the previously
reported absence of CNS side effects and null cytotoxicity
(Catalano et al. 2012), reinforce the role of MHM as a
possible metabolite switch candidate.
found for the less lipophilic analog
– MHM (log
D = 0.02) – in comparison with the one found for mex-
iletine (log D = 0.53) (Desaphy et al. 2012) might be
explained as a reflex of reduced access of the former to
its binding site when compared with an easier access of
the latter. However, lipophilicity might also play a critical
role as a determinant of the intimate interaction between
the compounds and their binding site. Our ab initio
investigations on possible subtle differences in MHM and
mexiletine relative binding modes suggest that the two
compounds should display similar tendencies to attract
electrons from aromatic ring residues as indicated by the
mean x values found for the complexes formed by their
corresponding truncated aromatic moieties with the ones
corresponding to Tyr652 (0.90 ꢀ 0.13 and 0.88 ꢀ 0.10,
respectively) and Phe656 (0.94 ꢀ 0.13 and 0.89 ꢀ 0.12,
respectively). However, in high lipophilic environments
MHM should give relatively less stable complexes with
Tyr residues with respect to those formed by mexiletine: a
significant mean difference of 1.2 ꢀ 0.3 Kcal/mol
(P = 0.01) was found between the interaction energy of
the truncated aromatic moiety of MHM with the moiety
representative of Tyr (TYRm), and that of the truncated
moiety representative of mexiletine with the same TYRm.
Indeed, mexiletine was about six times more potent as a
hERG blocker than MHM, in good quantitative agree-
ment with the prediction. In more hydrophilic environ-
ments MHM interactions are predicted to be at least as
strong as the ones of mexiletine. This seems to be the case
for MHM and mexiletine when their corresponding
As most open-channel blockers, mexiletine and MHM
accelerate significantly the rate of activation and cause a
small deceleration of the deactivation time-course
(Fig. 2E). This kinetic effect is due to the reopening of
channels caused by charged drug unbinding (Wang et al.
1999; Mitcheson et al. 2000b; Tie et al. 2000). Given these
observations, we conclude that mexiletine blocks the
hERG channel preferentially in an open state, with fea-
tures similar to those of other open-channel blockers, for
example, miconazole, mesoridazine, and ketanserin (Su
et al. 2004; Kikuchi et al. 2005; Tang et al. 2008).
The molecular mechanism of many hERG inhibitors is
an open-state-dependent trapping model of drugs into
the inner cavity of the channel (Vandenberg et al. 2012).
In this model, aromatic residues on the S6 helix (Tyr652
and Phe656) and polar residues at the base of the pore
helix (Thr623 and Ser624) line the inner cavity of the
2015 | Vol. 3 | Iss. 5 | e00160
Page 10
ª 2015 The Authors. Pharmacology Research & Perspectives published by John Wiley & Sons Ltd,
British Pharmacological Society and American Society for Pharmacology and Experimental Therapeutics.

R. Gualdani et al.
hERG Channel Inhibition by Mexiletine and MHM
sodium channel blocking activities are considered: MHM
was only twice as potent as mexiletine on cardiac sodium
channels (Catalano et al. 2012).
the mechanism of reverse use dependence of select APD/
QT-prolonging agents (Guo et al. 2010). With regard to
our data and in accordance to the literature, we can specu-
late that the low documented cardiotoxicity of mexiletine
could be due to the fact that mexiletine inhibition of INaL
may counteract any simultaneous prolongation of the QT
interval induced by the block of IKr/hERG channel. Note-
worthy, clinical relevant combined effects on INa and IKr
have been also reported for the antianginal agent ranol-
azine, which has been shown to be of potential interest
(Hancox and Doggrell 2010; Du et al. 2014) for the treat-
ment of atrial fibrillation; in the case of ranolazine the
inhibition of IKr/hERG appears even as a “positive” side
effect.
The possibility to reduce hERG potency by modifying
the pattern of aromatic substitution of hERG blocking
compounds has been previously explored, and results
from elegant mutagenesis and molecular modeling studies
on both clofilium and ibutilide analogs suggested that
phenyl ring para-substituents strongly affect Tyr binding,
with polar and electronegative substituents reinforcing the
binding interaction (Perry et al. 2006). Perry et al. con-
cluded that modifying the para-substituent could be a
useful strategy for reducing hERG potency. Our results
are in general agreement with the latter statement. How-
ever, when the relatively less explored meta-position was
substituted with a polar group such as the hydroxyl one
in MHM, a reduction in hERG potency was observed,
thus opposing to what previously observed with para-sub-
stituents in both clofilium and ibutilide series (Perry et al.
2006). Puzzling as it may appear, this observations pro-
vide further insights into the molecular basis of drug/
hERG interactions.
From the point of view of synthetic chemist, the bene-
ficial effect offered by the introduction of a meta-phenol
function as a structural feature to reduce hERG affinity
might be of general applicability. If validated through
the study of other well-known hERG blockers oppor-
tunely hydroxylated on their phenyl rings, this chemical
modification might provide the key to rescue useful
drugs previously withdrawn from clinical use because of
their hERG-related toxicity, and orientate the design of
safer new drugs. Studies in this respect have been under-
taken.
Of course, our model is tainted with a number of limi-
tations (current inhibition measurements as indirect evi-
dence of binding energies; truncated moieties assumed as
representative of the whole partner entities with conse-
quent overlooking of protonated amine group contribu-
tions and conformation issues; Manichean assumptions
about dielectric constants; relative limited calculation lev-
els used for energy predictions in water solution; allosteric
perturbations as consequences of mutations). However,
some of the above were attenuated by the fact that two
strictly related compounds were considered where the
only structural difference was an OH group in the meta-
position of the xylyloxy moiety.
Acknowledgements
We thank M. D’Amico and A. Arcangeli (University of
Florence) for the kind gift of HEK-hERG trasfected cells;
J. Barhanin (Universitꢁe de Nice-Sophia Antipolis, Nice,
France) for PCI-hKCNQ1 and pIRES-CD8-hKCNE1 plas-
mids; D. Golenbock (University of Massachusetts Medical
School, Boston, USA) for providing pcDNA3-EGFP (Add-
gene plasmid 13031); X. Koenig for providing F656A and
Y652A hERG plasmids.
Finally, in the light of our results, it is important to
observe that, even though our data show that mexiletine
inhibits hERG at physiologically relevant concentration,
mexiletine has not been documented to significantly affect
cardiac repolarization (normal QT-interval in the ECG)
and is generally a drug well tolerated in man at low doses.
Conversely, mexiletine has been shown to cause a signifi-
cant reduction in action potential duration (APD) mainly
through its inhibition of the sodium current (Matsuo
et al. 1985; Roden et al. 1987); this mechanism has been
considered as an advantage for the treatment of patients
with LQT3 (Wang et al. 1997) and LQT8 (Gao et al.
2013) syndromes, suggesting the possibility of gene-spe-
cific therapy for the two distinct forms of the congenital
LQTS. More specifically, mexiletine has been shown to
preferentially block at therapeutic concentrations the late
sodium current (INaL) compared to the peak sodium cur-
rent (Wang et al. 1997) and this selectivity contributes to
Author Contributions
R. G., G. L. designed the research study; R. G., G. L., F.
T.-B., M. R., I. D., M. C. performed the experiments; R.
G., G. L., F. T.-B., N. A. C. analyzed the data; R. G., G.
L., F. T.-B. wrote or contribute to the writing of the
article.
Disclosure
None declared.
References
Akinci E, Yu€zbaßsioglu Y, Coßskun F (2011). Hemodialysis as
an alternative treatment of mexiletine intoxication. Am J
Emerg Med 29: 1235–1236.
ª 2015 The Authors. Pharmacology Research & Perspectives published by John Wiley & Sons Ltd,
British Pharmacological Society and American Society for Pharmacology and Experimental Therapeutics.
2015 | Vol. 3 | Iss. 5 | e00160
Page 11

hERG Channel Inhibition by Mexiletine and MHM
R. Gualdani et al.
Barhanin J, Lesage F, Guillemare E, Fink M, Lazdunski M,
Romey G (1996). K(V)LQT1 and lsK (minK) proteins
associate to form the I(Ks) cardiac potassium current. Nature
384: 78–80.
interactions and reduced potency against inactivation mutants.
J Mol Cell Cardiol 74: 220–230.
Eijkelkamp N, Linley JE, Baker MD, Minett MS, Cregg R,
Werdehausen R, et al. (2012). Neurological perspectives on
voltage-gated sodium channels. Brain 135: 2585–2612.
Becke AD (1988). Density-functional exchange-energy
approximation with correct asymptotic behavior. Phys Rev A
38: 3098–3100.
El Harchi A, Zhang YH, Hussein L, Dempsey CE, Hancox JC
(2012). Molecular determinants of hERG potassium channel
inhibition by disopyramide. J Mol Cell Cardiol 52: 185–195.
Brochu RM, Dick IE, Tarpley JW, McGowan E, Gunner D,
Herrington J, et al. (2006). Block of peripheral nerve sodium
channels selectively inhibits features of neuropathic pain in
rats. Mol Pharmacol 69: 823–832.
Franchini C, Carocci A, Catalano A, Cavalluzzi MM, Corbo F,
Lentini G, et al. (2003). Optically active mexiletine analogues
as stereoselective blockers of voltage-gated Na+ channels.
J Med Chem 46: 5238–5248.
Bruno C, Carocci A, Catalano A, Cavalluzzi MM, Corbo F,
Franchini C, et al. (2006). Facile, alternative route to
lubeluzole, its enantiomer, and the racemate. Chirality 18:
227–231.
Gao Y, Xue X, Hu D, Liu W, Yuan Y, Sun H, et al. (2013).
Inhibition of late sodium current by mexiletine: a novel
pharmotherapeutical approach in timothy syndrome. Circ
Arrhythm Electrophysiol 6: 614–622.
Campbell NP, Kelly JG, Adgey AA, Shanks RG (1978). The
clinical pharmacology of mexiletine. Br J Clin Pharmacol 6:
103–108.
Gemma S, Camodeca C, Brindisi M, Brogi S, Kukreja G,
Kunjir S, et al. (2012). Mimicking the intramolecular
hydrogen bond: synthesis, biological evaluation, and molecular
modeling of benzoxazines and quinazolines as potential
antimalarial agents. J Med Chem 55: 10387–10404.
Catalano A, Desaphy JF, Lentini G, Carocci A, Di Mola A,
Bruno C, et al. (2012). Synthesis and toxicopharmacological
evaluation of m-hydroxymexiletine, the first metabolite of
mexiletine more potent than the parent compound on voltage-
gated sodium channels. J Med Chem 55: 1418–1422.
Guo D, Lian J, Liu T, Cox R, Margulies KB, Kowey PR, et al.
(2010). Contribution of late sodium current (I(Na-L)) to rate
adaptation of ventricular repolarization and reverse use-
dependence of QT-prolonging agents. Heart Rhythm 8:
762–769.
Cavalluzzi MM, Viale M, Bruno C, Carocci A, Catalano A,
Carrieri A, et al. (2013). A convenient synthesis of lubeluzole
and its enantiomer: evaluation as chemosensitizing agents on
human ovarian adenocarcinoma and lung carcinoma cells.
Bioorg Med Chem Lett 23: 4820–4823.
Halgren TA (1996). Merck molecular force field. I. Basis, form,
scope, parameterization, and performance of MMFF94. J
Computational Chem 17: 490–519.
Challapalli V, Tremont-Lukats IW, McNicol ED, Lau J, Carr
DB 2005. Systemic administration of local anesthetic agents to
relieve neuropathic pain. Cochrane Database Syst Rev 4:
CD003345.
Hancox JC, Doggrell SA (2010). Perspective: does ranolazine
have potential for the treatment of atrial fibrillation? Expert
Opin Investig Drugs 19: 1465–1474.
Conte Camerino D, Tricarico D, Desaphy J-F (2007). Ion
Hancox JC, McPate MJ, El Harchi A, Zhang YH (2008). The
hERG potassium channel and hERG screening for drug-
induced torsades de pointes. Pharmacol Ther 119: 118–132.
channel pharmacology. Neurotherapeutics 4: 184–198.
Cregg R, Cox JJ, Bennett DL, Wood JN, Werdehausen R
(2014). Mexiletine as a treatment for primary erythromelalgia:
normalization of biophysical properties of mutant L858F
NaV1.7 sodium channels. Br J Pharmacol 171: 4455–4463.
Hille B (1977). Local anesthetics: hydrophilic and hydrophobic
pathways for the drug-receptor reaction. J Gen Physiol 69:
497–515.
Danielsson BR, Lansdell K, Patmore L, Tomson T (2003).
Phenytoin and phenobarbital inhibit human hERG potassium
channels. Epilepsy Res 55: 147–157.
Horie M, Yoshida H 1999. An antiarrhythmic agent with class
Ib action, aprinidine, inhibits human IKr but not IKs
heterologously expressed in COS7 cells. Circulation 100:I–280.
Davidson ER, Feller D (1986). Basis set selection for molecular
Kiehn J, Lacerda AE, Wible I, Brown AM (1996). Molecular
physiology and pharmacology of hERG. Single-channel
currents and block by dofetilide. Circulation 94: 2572–2579.
calculations. Chem Rev 86: 681–696.
Desaphy JF, Dipalma A, Costanza T, Carbonara R, Dinardo
MM, Catalano A, et al. (2012). Molecular insights into the
local anesthetic receptor within voltage-gated sodium channels
using hydroxylated analogs of mexiletine. Front Pharmacol 3:
17.
Kikuchi K, Nagatomo T, Abe H, Kawakami K, Duff HJ,
Makielski JC, et al. (2005). Blockade of hERG cardiac K+
current by antifungal drug miconazole. Br J Pharmacol 144:
840–848.
Du C, Zhang Y, El Harchi A, Dempsey CE, Hancox JC (2014).
Ranolazine inhibition of hERG potassium channels: drug-pore
Liu S, Rasmusson RL, Campbell DL, Wang S, Strauss HC
(1996). Activation and inactivation kinetics of an E-4031-
2015 | Vol. 3 | Iss. 5 | e00160
Page 12
ª 2015 The Authors. Pharmacology Research & Perspectives published by John Wiley & Sons Ltd,
British Pharmacological Society and American Society for Pharmacology and Experimental Therapeutics.

R. Gualdani et al.
hERG Channel Inhibition by Mexiletine and MHM
sensitive current from single ferret atrial myocytes. Biophys J
inner cavity of HERG channels: molecular insights from
structure-activity relationships of clofilium and ibutilide
analogs. Mol Pharmacol 69: 509–519.
70: 2704–2715.
Logigian EL, Martens WB, Moxley RT IV, McDermott MP,
Dilek N, Wiegner AW, et al. (2010). Mexiletine is an effective
antimyotonia treatment in myotonic dystrophy type 1.
Neurology 74: 1441–1448.
Ridley JM, Milnes JT, Benest AV, Masters JD, Witchel HJ,
Hancox JC (2003). Characterization of recombinant hERG K+
channel blockade by the Class Ia antiarrhythmic drug
procainamide. Biochem Biophys Res Commun 306: 388–393.
Marenich AV, Olsen RM, Kelly CP, Cramer CJ, Truhlar DG
(2007). Self-consistent reaction field model for aqueous and
nonaqueous solutions based on accurate polarized partial
charges. J Chem Theory Comput 3: 2011–2033.
Roche O, Trube G, Zuegge J, Pflimlin P, Alanine A, Schneider
G (2002). A virtual screening method for prediction of the
hERG potassium channel liability of compound libraries.
ChemBioChem 3: 455–459.
Matsuo S, Kishida H, Munakata K, Atarashi H (1985). The
effects of mexiletine on action potential duration and its
restitution in guinea pig ventricular muscles. Jpn Heart J 26:
271–287.
Roden DM, Iansmith DH, Woosley RL (1987). Frequency-
dependent interactions of mexiletine and quinidine on
depolarization and repolarization in canine Purkinje fibers.
J Pharmacol Exp Ther 243: 1218–1224.
Matthews E, Hanna MG (2014). Repurposing of sodium
channel antagonists as potential new anti-myotonic drugs. Exp
Neurol 261: 812–815.
Sanguinetti MC, Jurkiewicz NK (1990). Two components of
cardiac delayed rectifier K+ current. Differential sensitivity to
block by class III antiarrhythmic agents. J Gen Physiol 96:
195–215.
Maynard AT, Huang M, Rice WG, Covell DG (1998).
Reactivity of the HIV-1 nucleocapsid protein p7 zinc finger
domains from the perspective of density-functional theory.
Proc Natl Acad Sci USA 95: 11578–11583.
Sanguinetti MC, Mitcheson JS (2005). Predicting drug-hERG
channel interactions that cause acquired long QT syndrome.
Trends Pharmacol Sci 26: 119–124.
Mitcheson JS, Hancox JC (1997). Modulation by mexiletine of
action potentials, L-type Ca current and delayed rectifier K+
current recorded from isolated rabbit atrioventricular nodal
myocytes. Pflu€gers Arch 434: 855–858.
Sanguinetti MC, Tristani-Firouzi M (2006). hERG potassium
channels and cardiac arrhythmia. Nature 440: 463–469.
Sanguinetti MC, Curran ME, Spector PS, Keating MS
(1996a). Spectrum of hERG K+-channel dysfunction in an
inherited cardiac arrhythmia. Proc Natl Acad Sci USA 93:
2208–2212.
Mitcheson JS, Chen J, Lin M, Culberson C, Sanguinetti MC
(2000a). A structural basis for drug-induced long QT
syndrome. Proc Natl Acad Sci USA 97: 12329–12333.
Mitcheson JS, Chen J, Sanguinetti MC (2000b). Trapping of a
methanesulfonanilide by closure of the hERG potassium
channel activation gate. J Gen Physiol 115: 229–240.
Sanguinetti MC, Curran ME, Zou A, Shen J, Spector PS,
Atkinson DL, et al. (1996b). Coassembly of K(V)LQT1 and
minK (IsK) proteins to form cardiac I(Ks) potassium channel.
Nature 384: 80–83.
Parr RG, Szentpꢁaly LV, Liu S (1999). Electrophilicity Index.
J Am Chem Soc 121: 1922–1924.
Smith PL, Baukrowitz T, Yellen G (1996). The inward
rectification mechanism of the hERG cardiac potassium
channel. Nature 379: 833–836.
Parthasarathi R, Subramanian V, Roy DR, Chattaraj PK
(2004). Electrophilicity index as a possible descriptor of
biological activity. Bioorg Med Chem 12: 5533–5543.
Statland JM, Bundy BN, Wang Y, Rayan DR, Trivedi JR,
Sansone VA, et al. (2012). Consortium for Clinical
Paton RS, Goodman JM (2009). Hydrogen bonding and pi-
stacking: how reliable are force fields? A critical evaluation of
force field descriptions of nonbonded interactions. J Chem Inf
Model 49: 944–955.
Investigation of Neurologic Channelopathies. Mexiletine for
symptoms and signs of myotonia in nondystrophic myotonia:
a randomized controlled trial. JAMA 308: 1357–1365.
Su Z, Martin R, Cox BF, Gintant G (2004). Mesoridazine: an
open-channel blocker of human ether-a-go-go-related gene K+
channel. J Mol Cell Cardiol 36: 151–160.
Paul AA, Witchel HJ, Hancox JC (2002). Inhibition of the
current of heterologously expressed hERG potassium channels
by flecainide and comparison with quinidine, propafenone and
lignocaine. Br J Pharmacol 136: 717–729.
Tamargo J, Caballero R, Gomez R, Valenzuela C, Delpon E
(2004). Pharmacology of cardiac potassium channels.
Cardiovasc Res 62: 9–33.
Pearlstein R, Vaz R, Rampe D (2003). Understanding the
structure-activity relationship of the human ether-a-go-go-
related gene cardiac K+ channel. A model for bad behavior.
J Med Chem 46: 2017–2022.
Tang Q, Li Z-Q, Li W, Guo J, Sun H-Y, Lau C-P, et al.
(2008). The 5-HT2 antagonist ketanserin is an open channel
blocker of human cardiac ether-ꢀa-go-go-related gene (hERG)
potassium channels. Br J Pharmacol 155: 365–373.
Perry M, Stansfeld PJ, Leaney J, Wood C, de Groot MJ,
Leishman D, et al. (2006). Drug binding interactions in the
ª 2015 The Authors. Pharmacology Research & Perspectives published by John Wiley & Sons Ltd,
British Pharmacological Society and American Society for Pharmacology and Experimental Therapeutics.
2015 | Vol. 3 | Iss. 5 | e00160
Page 13

hERG Channel Inhibition by Mexiletine and MHM
R. Gualdani et al.
Thomas D, Wendt-Nordahl G, Rockl K, Ficker E, Brown AM,
Kiehn J (2001). High-affinity blockade of human ether-a-go-
go-related gene human cardiac potassium channels by the
novel antiarrhythmic drug BRL-32872. J Pharmacol Exp Ther
297: 753–761.
propafenone derivatives in hERG channels. Br J Pharmacol
162: 1542–1552.
Witchel HJ, Hancox JC (2000). Familial and acquired long QT
syndrome and the cardiac rapid delayed rectifier potassium
current. Clin Exp Pharmacol Physiol 27: 753–766.
Tie H, Walker BD, Singleton CB, Valenzuela SM, Bursill JA,
Wyse KR, et al. (2000). Inhibition of hERG potassium
channels by the antimalarial agent halofantrine. Br J
Pharmacol 130: 1967–1975.
Supporting Information
Additional Supporting Information may be found in the
online version of this article:
Vandenberg JI, Perry MD, Perrin MJ, Mann SA, Ke Y, Hill AP
(2012). hERG K(+) channels: structure, function, and clinical
significance. Physiol Rev 92: 1393–1478.
Data S1. Materials and methods.
Figure S1. Slow time constants (sslow) of hERG deactiva-
tion elicited before and after application of mexiletine
and MHM.
Figure S2. Effect of mexiletine and MHM on Na+,K+-AT-
Pase and KCNQ1/KCNE1 channels.
Vaughan Williams EM (1984). A classification of
antiarrhythmic actions reassessed after a decade of new drugs.
J Clin Pharmacol 24: 129–147.
Wang DW, Kiyosue T, Sato T, Arita M (1996). Comparison of
the effects of class I anti-arrhythmic drugs, cibenzoline,
mexiletine and flecainide, on the delayed rectifier K+ current
of guinea-pig ventricular myocytes. J Mol Cell Cardiol 28:
893–903.
Table S1. Calculated electrophilicity index x (Kcal/mol)
for complexes formed by either MHMm or MEXm with
either TYRm or PHEm (in vacuum).
Table S2. Calculated interaction energies (Kcal/mol) for
complexes formed by either MHMm or MEXm with
either TYRm or PHEm (in vacuum).
Wang DW, Yazawa K, Makita N, George AL Jr, Bennett PB
(1997). Pharmacological targeting of long QT mutant sodium
channels. J Clin Invest 99: 1714–1720.
Table S3. Calculated interaction energies (Kcal/mol) for
complexes formed by either MHMm or MEXm with
TYRm (in water solution).
Wang HZ, Shi H, Liao SJ, Wang Z (1999). Inactivation gating
determines nicotine blockade of human hERG channels. Am J
Physiol 277: H1081–H1088.
Table S4. Calculated interaction energies (Kcal/mol) for
complexes formed by either MHMm or MEXm with
PHEm (in water solution).
Windisch A, Timin E, Schwarz T, Stork-Riedler D, Erker T,
Ecker G, et al. (2011). Trapping and dissociation of
2015 | Vol. 3 | Iss. 5 | e00160
Page 14
ª 2015 The Authors. Pharmacology Research & Perspectives published by John Wiley & Sons Ltd,
British Pharmacological Society and American Society for Pharmacology and Experimental Therapeutics.