Computational design and discovery of minimally structured hERG blockers

Article abstract of DOI:10.1021/jm201194q

Molecular knowledge of hERG blocking liability can offer the possibility of optimizing lead compounds in a way that eliminates potentially lethal side effects. In this study, we computationally designed, synthesized, and tested a small series of minimally structured molecules. Some of these compounds were remarkably potent against hERG (6, IC₅₀ = 2.4 nM), allowing us to identify the minimal structural requirements for hERG blocking liability.

Full text of DOI:10.1021/jm201194q

Brief Article
pubs.acs.org/jmc
Computational Design and Discovery of “Minimally Structured”
hERG Blockers
Andrea Cavalli,*,†,‡ Rosa Buonfiglio,† Cristina Ianni,† Matteo Masetti,† Luisa Ceccarini,† Rachel Caves,§
Michael W. Y. Chang,§ John S. Mitcheson,§ Marinella Roberti,† and Maurizio Recanatini†
†Department of Pharmaceutical Sciences, University of Bologna, Via Belmeloro 6, I-40126 Bologna, Italy
‡Department of Drug Discovery and Development, Istituto Italiano di Tecnologia, Via Morego 30, I-16163 Genova, Italy
§Department of Cell Physiology and Pharmacology, University of Leicester, University Road, Leicester LE1 9HN, United Kingdom
S
* Supporting Information
ABSTRACT: Molecular knowledge of hERG blocking liability can offer the possibility of optimizing lead compounds in a way
that eliminates potentially lethal side effects. In this study, we computationally designed, synthesized, and tested a small series of
“minimally structured” molecules. Some of these compounds were remarkably potent against hERG (6, IC50 = 2.4 nM), allowing
us to identify the minimal structural requirements for hERG blocking liability.
understanding of hERG−drug binding and a potential
identification of pharmacophoric functions responsible for the
interactions between small organic molecules and hERG. All
this information has then been applied to the rational design of
new chemical entities in several drug discovery programs.19,20
One of the accepted pharmacophoric models for hERG
binders, mentioned above, shows (i) two aromatic rings (C0
and C1) almost coplanar and at a distance of ∼4.5−6.5 Å, (ii)
one basic nitrogen atom (N) ∼5−9 and ∼5.5−7 Å far away
from the rings C0 and C1, and (iii) a further aromatic moiety
(C2), which is required for increasing potency against hERG,
∼4.5−7.5 Å from the nitrogen atom (see Figure 1). This model
INTRODUCTION
■
Binding of different drugs into the pore domain of human
ether-a-go-go-related gene (hERG) channels (Kv11.1) can be
the concurrent cause of a severe tachyarrhythmia (torsades de
pointes), which is recognized as a very dangerous side effect of
drugs, eventually leading to sudden cardiac death.1 Several
approved drugs that induce this effect, referred to as drug-
induced long-QT syndrome, LQTS (e.g., astemizole, cisapride,
terfenadine, haloperidol, sertindole, etc.), have been withdrawn
from the market or restricted in their use because of their
effects on QT interval prolongation.2 In retrospect, it has been
shown that all these drugs are able to interact with and block
hERG channels and that hERG binding is probably one of the
key molecular events for drug-induced LQTS.3−7 Therefore,
the hERG binding liability of new chemical entities has greatly
challenged several drug discovery programs.
hERG blocking liability is usually assessed quite early in a
drug discovery program, and several relatively fast experimental
approaches are now available, spanning from planar electrode
patch clamp techniques and radiolabeled drug binding assays to
high-throughput fluorescent dye assays using Chinese hamster
ovary (CHO) cells stably transfected with hERG ion channels.8
In this scenario, in silico predictions of hERG binding have also
played an important role in drug discovery. Over the years, a
plethora of computational studies have appeared in the
literature.9,10 The first pharmacophore models developed
about 10 years ago by Ekins et al.,11 Cavalli et al.,12 Aronov
et al.,13 and Pearlstein et al.14 were characterized by a central
positive ionizable feature linked to aromatic or hydrophobic
groups. Conversely, other models have been based on
uncharged ligands,15 accounting for additional features
responsible for hERG block. In addition, more recent in silico
studies took advantage of remarkably large data sets, defining
the pool of molecules and the associated physicochemical
features responsible for binding to hERG.16−18 All these
investigations, carried out using ligand- or structure-based
approaches, have greatly contributed to achieving a superior
Figure 1. Pharmacophore model of hERG K+ channel blockers. C0,
C1, and C2 are the centers of mass of aromatic moieties, while N is a
protonable nitrogen atom. Potent hERG blockers fit with this
pharmacophore model if they adopt an extended conformation.
can account for the hERG blocking activity of several quite
flexible drugs (e.g., astemizole, terfenadine, cisapride, etc.),
which should adopt a kind of “extended” conformation to
properly fit with the pharmacophore.12 Such an extended
Received: September 9, 2011
Published: March 28, 2012
© 2012 American Chemical Society
4010
dx.doi.org/10.1021/jm201194q | J. Med. Chem. 2012, 55, 4010−4014

Journal of Medicinal Chemistry
Brief Article
conformation was subsequently confirmed by structure-based
studies, where docking simulations have shown that flexible
molecules binding into the pore domain of hERG can adopt a
conformation very similar to that reported in the model of
Figure 1.21−23 Furthermore, this model was also used to align a
series of 31 hERG blockers to carry out a comparative
molecular field analysis (CoMFA)24 and to develop a 3D
QSAR model for predicting hERG binding potencies of novel
drug candidates.12
In the present study, we designed a small set of “minimally
structured” molecules to fulfill the pharmacophoric hypothesis
of Figure 1. Eight new derivatives were synthesized as reported
in Schemes 1 and 2. Six of these molecules (1−6, Table 1)
Figure 1 and further highlights the key role of the positively
charged nitrogen for high hERG blocking potency. On this
basis, we propose that three phenyl rings and one basic
nitrogen, suitably spaced, represent the stereotype of a
minimally structured potent hERG blocker.
RESULTS AND DISCUSSION
■
hERG Inhibitor Design. A small set of “minimally
structured” hERG blockers was designed on the basis of the
pharmacophoric model of Figure 1. Compounds 1 and 2 are
diphenylpropanamines substituted by an ethyl- or a propyl-
phenyl chain, respectively. In 3 and 4, the methine carbon atom
of the benzydryl moieties was replaced by a tertiary nitrogen
atom. In addition, in 5 and 6, a 4-4′di-fluoro substitution was
introduced in the diarylamino moieties to modify the electronic
features of the C0 and C1 aromatic rings, without significantly
altering volume and lipophilicity of the molecules. Compounds
7 and 8 were the amide derivatives of 3 and 6 and were
designed and synthesized to investigate the role of the positive
charge on hERG block.
a
Scheme 1
The rational design of 1−6 had also been strongly supported
by known mutagenesis25,26 and docking23 studies showing that
Tyr652 and Phe656 play a pivotal role in the hERG drug
binding by promoting cation−π, π−π, and hydrophobic
interactions with the basic nitrogen and aromatic rings of drugs.
CoMFA Model. An extended CoMFA model for hERG K+
channel block based on that previously developed in 200212
was constructed (see Supporting Information for computa-
tional details). The training set was composed of 75 (vs 31 of
the 2002 model) hERG blockers selected from the literature as
representative compounds of different therapeutic classes
(antipsychotics, antihistamines, antibacterials, etc.) able to
block hERG. These molecules showed a homogeneous
distribution of drug activities within the pIC50 range 3.8−9.0.
The CoMFA model was then used to support the above-
reported design strategy and to help the definition of the
proper length of the chain connecting N to C2. Finally, the
model was utilized to predict the activity of 1−8 (Table 1). All
the statistical parameters and the CoMFA contour maps are
reported in Supporting Information.
a
Reagents and conditions: NaBH3CN, MeOH dry, room temperature
for 7 days.
a
Scheme 2
Chemistry. 1−6 were synthesized in a parallel fashion,
taking advantage of a reductive amination between the
appropriate amines 9−11 and the commercial aldehydes 12,
13 with NaBH3CN as reducing agent (Scheme 1; see
Supporting Information for experimental details). 7 and 8
were obtained by a nucleophilic substitution between the
amines 10, 11 and the corresponding commercially available
acyl chlorides 14, 15 (Scheme 2; see Supporting Information
for experimental details). The 3,3-diphenylpropan-1-amine 9
was commercially available. The N1,N1-diphenylethane-1,2-
diamine 10 was prepared as described in the literature.27 The
N1,N1-bis(4-fluorophenyl)ethane-1,2-diamine 11 was obtained
from hydrazine cleavage of phthalimidic derivative 17, which
was synthesized from N-(2-chloroethyl)-4-fluoro-N-(4-
fluorophenyl)aniline 1628 through a nucleophilic substitution
reaction with potassium phthalimide (Scheme S1 and
Supporting Information for experimental details).
a
Reagents and conditions: diisopropylethylamine, MeOH dry, 0° C
for 4 h.
displayed the three aromatic moieties (C0, C1, C2) and the
basic nitrogen atom (N), spatially arranged to fit the
pharmacophoric scheme, whereas 7 and 8 carried a
methanamide instead of the methylenamine group to evaluate
the role of a positive charge on hERG block potency. The
activity of the new compounds was predicted using the 3D
QSAR equation derived from a CoMFA model, which
represented an extension of that reported in 2002.12 Finally,
all molecules were tested to assess the hERG block activity
using patch clamp experiments carried out in human embryonic
kidney cells. The new compounds were active against hERG,
and some of them turned out to be nanomolar inhibitors of this
channel. This study confirms the pharmacophoric hypothesis of
Biology. hERG K+ current inhibition by 1−8 was
determined from whole-cell voltage clamp recordings made in
HEK cells stably expressing hERG channels (see Supporting
Information for experimental details). The molecules all
inhibited hERG currents in a concentration dependent manner.
Inhibition was open state dependent, and repetitive pulsing to 0
4011
dx.doi.org/10.1021/jm201194q | J. Med. Chem. 2012, 55, 4010−4014

Journal of Medicinal Chemistry
Brief Article
Table 1. Experimental and Predicted hERG K+ Channel Blocking Activity of 1−8
mV was required until steady-state inhibition was achieved (see
Figure 2A for representative current traces). The mean
concentration−response relationships for 1−6 and 7, 8 are
shown in parts B and C of Figure 2, respectively, and the IC50
values are reported in Table 1. In addition, Figure 2C shows the
mean concentration−response relationships of 3 and 6 (in
dashed lines) to compare hERG blocking activity with
corresponding amide derivatives 7 and 8, bearing methanamide
instead of methylenamine. As predicted by the CoMFA model,
the biological activities of the molecules were in the nanomolar
to micromolar range.
Structure−Activity Relationships. Comparison of the
biological activity of 1, 2 and 5, 6 showed the importance of the
length of the N−C2 linker for potency, which is in agreement
with in silico predictions. However, the same correlation could
not be observed for 3 and 4. Experimental data for 1, 3, and 5
also revealed that a positive contribution to the activity could
be achieved by substituting the carbon between C0 and C1
with a tertiary nitrogen. Moreover, adding fluorine atoms at
para positions of the C0 and C1 rings had a positive effect on
the activity, probably as a consequence of their electronic effect
on aromatic rings. In particular, the electron-withdrawal
behavior of fluorines decreased the electron density of the
phenyl groups, thus likely strengthening possible π−π
interactions with Tyr652 and Phe656. All these features (i.e.,
three methylenes between N and C2, and fluorine atoms on
both C0 and C1) were present on 6, which turned out to be the
most potent hERG blocker of this series (IC50 = 2.4 nM) and
one of the most potent hERG blockers ever reported in the
literature.
To investigate the role of the positive charge on the hERG
blocking activity, we designed and synthesized 7 and 8, which
represented the amide derivatives of 3 and 6. While both
compounds preserved a certain activity against hERG, the drop
in potency of 2 orders of magnitude (7860 nM vs 62 nM for 7
vs 3; 192 nM vs 2.4 nM for 8 vs 6) clearly showed that,
although not strictly necessary, the positive charge plays a
major role in determining the overall potency of a compound
toward hERG. In parallel, these data show that a possible way
to lower the hERG blocking liability of a newly designed
compound could be “switching off” a possible positive charge,
located on the new molecule in the way reported in Figure 1.
Notably, our CoMFA model nicely predicted the quite low
pIC50 of 7, while it underestimated the potency of 8 (see Table
1). This could be due to the underestimation of the role of the
fluorine atoms and the linker (as occurred for 6) and to the
overestimation of the role of the positive charge, which was
present on the most of the most potent derivatives of the
training set. In addition, we observed that the activities of 1 and
4 were overestimated whereas 2, 3, 5, 6, and 8 were
underestimated by the 3D QSAR, probably as a consequence
of their “minimal” chemical structures. In fact, these new
molecules lacked additional functional groups protruding into
the CoMFA steric and electrostatic regions (see Supporting
Information), despite a perfect fit with the pharmacophore of
Figure 1.
4012
dx.doi.org/10.1021/jm201194q | J. Med. Chem. 2012, 55, 4010−4014

Journal of Medicinal Chemistry
Brief Article
features that can positively or negatively affect drug penetration
into the cell and interaction with hERG.
In conclusion, by synthesizing and testing a small set of
hERG blockers, designed on the basis of the pharmacophoric
hypothesis of Figure 1, we demonstrate that three phenyl rings,
suitably spaced, and one protonable nitrogen atom can
represent the minimal structural requirements for high potency
for hERG block.
ASSOCIATED CONTENT
* Supporting Information
■
S
Computational details of the CoMFA model generation;
experimental chemical procedures and compound character-
ization; details of patch clamp experiments. This material is
available free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
*Phone/Fax: +39 051 2099735/4. E-mail: andrea.cavalli@
unibo.it.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The work has received funding from the European
Community’s Seventh Framework Programme (FP7/2007−
2013) under Grant No. 241679 (ARITMO Project; L.C.,
M.Re.), Istituto Italiano di Tecnologia, Genova, Italy (Seed
Project HERGSYM; R.B., C.I., M.M., M.Ro., M.Re.), and Alma
Mater Studiorum, University of Bologna, Bologna, Italy.
We greatly thank Prof. Craig January (University of
Wisconsin, U.S.) for the kind gift of HEK cells stably
transfected with hERG.
Figure 2. (A) Representative traces of hERG currents before (control)
and after steady state inhibition by indicated concentrations of 5. (B)
Mean ( SEM.) concentration response relationships for inhibition of
hERG currents by compounds 1-6 (n = 5 - 9 cells). Hill function fits of
the data are shown by solid lines. (C) Mean SEM concentration−
response relationships for inhibition of hERG currents by 7 and 8 with
3 and 6 (charged analogues) for comparison.
ABBREVIATIONS USED
CHO, Chinese hamster ovary; LQTS, long-QT syndrome;
pIC50, negative logarithm of IC50; pKa, negative logarithm of Ka
■
In Figure S1, the 3D structure of a low energy conformation
of 2 was superimposed onto the pharmacophore model (Figure
S1A) and the CoMFA contour maps (Figure S1B) are shown.
It is readily apparent that 2 shows the minimal structural
requirements for hERG block, as this molecule carries three
phenyl rings and one basic nitrogen atom suitably spaced to fit
C0, C1, C2, and N. Despite the CoMFA-based prediction
pointing to this compound as a low micromolar blocker of
hERG, 2 actually turned out to be a nanomolar inhibitor of this
channel. Therefore, we can speculate that the pharmacophoric
functions C0, C1, C2, and N are required for favorable
interactions with hERG, while accessory groups, such as the
fluorine atoms of 6, can be important for increasing the hERG
blocking potency. We also comment that the present CoMFA
model was greatly influenced by the chemical nature of the
training set, and therefore, it was only partially successful in
predicting the hERG inhibition potencies of structurally
different compounds, such as those here synthesized and
tested. In addition, other physicochemical features, such as pKa,
membrane permeability, etc., which were not taken into
account by the CoMFA model, can play a role in drugs applied
to the extracellular side of the membrane binding to hERG. In
fact, binding sites for hERG blockers have been mapped within
the inner cavity of the channel, which drugs can only access
from the intracellular side of the membrane.29 Therefore,
lipophilicity and the presence of formal charges are relevant
REFERENCES
■
(1) Sanguinetti, M. C.; Tristani-Firouzi, M. hERG potassium
channels and cardiac arrhythmia. Nature 2006, 440, 463−469.
(2) Shah, R. R. Can pharmacogenetics help rescue drugs withdrawn
from the market? Pharmacogenomics 2006, 7, 889−908.
(3) Kang, J.; Wang, L.; Chen, X. L.; Triggle, D. J.; Rampe, D.
Interactions of a series of fluoroquinolone antibacterial drugs with the
human cardiac K+ channel HERG. Mol. Pharmacol. 2001, 59, 122−
126.
(4) Bischoff, U.; Schmidt, C.; Netzer, R.; Pongs, O. Effects of
fluoroquinolones on HERG currents. Eur. J. Pharmacol. 2000, 406,
341−343.
(5) Taglialatela, M.; Pannaccione, A.; Castaldo, P.; Giorgio, G.; Zhou,
Z.; January, C. T.; Genovese, A.; Marone, G.; Annunziato, L.
Molecular basis for the lack of HERG K+ channel block-related
cardiotoxicity by the H1 receptor blocker cetirizine compared with
other second-generation antihistamines. Mol. Pharmacol. 1998, 54,
113−121.
(6) Suessbrich, H.; Schonherr, R.; Heinemann, S. H.; Attali, B.; Lang,
F.; Busch, A. E. The inhibitory effect of the antipsychotic drug
haloperidol on HERG potassium channels expressed in Xenopus
oocytes. Br. J. Pharmacol. 1997, 120, 968−974.
(7) Suessbrich, H.; Waldegger, S.; Lang, F.; Busch, A. E. Blockade of
HERG channels expressed in Xenopus oocytes by the histamine
receptor antagonists terfenadine and astemizole. FEBS Lett. 1996, 385,
77−80.
4013
dx.doi.org/10.1021/jm201194q | J. Med. Chem. 2012, 55, 4010−4014

Journal of Medicinal Chemistry
Brief Article
(8) Stork, D.; Timin, E. N.; Berjukow, S.; Huber, C.; Hohaus, A.;
Auer, M.; Hering, S. State dependent dissociation of HERG channel
inhibitors. Br. J. Pharmacol. 2007, 151, 1368−1376.
(28) Wise, L. D.; Pattison, I. C.; Butler, D. E.; DeWald, H. A.; Lewis,
E. P.; Lobbestael, S. J.; Nordin, I. C.; Poschel, B. P.; Coughenour, L. L.
1-[3-(Diarylamino)propyl]piperidines and related compounds, poten-
tial antipsychotic agents with low cataleptogenic profiles. J. Med. Chem.
1985, 28, 606−612.
(29) Perry, M.; Stansfeld, P. J.; Leaney, J.; Wood, C.; de Groot, M. J.;
Leishman, D.; Sutcliffe, M. J.; Mitcheson, J. S. Drug binding
interactions in the inner cavity of HERG channels: molecular insights
from structure−activity relationships of clofilium and ibutilide analogs.
Mol. Pharmacol. 2006, 69, 509−519.
(9) Recanatini, M.; Cavalli, A.; Masetti, M. Modeling HERG and its
interactions with drugs: recent advances in light of current potassium
channel simulations. ChemMedChem 2008, 3, 523−535.
(10) Raschi, E.; Ceccarini, L.; De Ponti, F.; Recanatini, M. hERG-
related drug toxicity and models for predicting hERG liability and QT
prolongation. Expert Opin. Drug Metab. Toxicol. 2009, 5, 1005−1021.
(11) Ekins, S.; Crumb, W. J.; Sarazan, R. D.; Wikel, J. H.; Wrighton,
S. A. Three-dimensional quantitative structure−activity relationship for
inhibition of human ether-a-go-go-related gene potassium channel. J.
Pharmacol. Exp. Ther. 2002, 301, 427−434.
(12) Cavalli, A.; Poluzzi, E.; De Ponti, F.; Recanatini, M. Toward a
pharmacophore for drugs inducing the long QT syndrome: insights
from a CoMFA study of HERG K(+) channel blockers. J. Med. Chem.
2002, 45, 3844−3853.
(13) Aronov, A. M.; Goldman, B. B. A model for identifying HERG
K+ channel blockers. Bioorg. Med. Chem. 2004, 12, 2307−2315.
(14) Pearlstein, R. A.; Vaz, R. J.; Kang, J.; Chen, X. L.;
Preobrazhenskaya, M.; Shchekotikhin, A. E.; Korolev, A. M.;
Lysenkova, L. N.; Miroshnikova, O. V.; Hendrix, J.; Rampe, D.
Characterization of HERG potassium channel inhibition using
CoMSiA 3D QSAR and homology modeling approaches. Bioorg.
Med. Chem. Lett. 2003, 13, 1829−1835.
(15) Aronov, A. M. Common pharmacophores for uncharged human
ether-a-go-go-related gene (hERG) blockers. J. Med. Chem. 2006, 49,
6917−6921.
(16) Thai, K. M.; Ecker, G. F. A binary QSAR model for classification
of hERG potassium channel blockers. Bioorg. Med. Chem. 2008, 16,
4107−4119.
(17) Nisius, B.; Goller, A. H.; Bajorath, J. Combining cluster analysis,
feature selection and multiple support vector machine models for the
identification of human ether-a-go-go related gene channel blocking
compounds. Chem. Biol. Drug Des. 2009, 73, 17−25.
(18) Doddareddy, M. R.; Klaasse, E. C.; Shagufta; Ijzerman, A. P.;
Bender, A. Prospective validation of a comprehensive in silico hERG
model and its applications to commercial compound and drug
databases. ChemMedChem 2010, 5, 716−729.
(19) Gavaghan, C.; Arnby, C.; Blomberg, N.; Strandlund, G.; Boyer,
S. Development, interpretation and temporal evaluation of a global
QSAR of hERG electrophysiology screening data. J. Comput.-Aided
Mol. Des. 2007, 21, 189−206.
(20) Pollard, C. E.; Abi Gerges, N.; Bridgland-Taylor, M. H.; Easter,
A.; Hammond, T. G.; Valentin, J. P. An introduction to QT interval
prolongation and non-clinical approaches to assessing and reducing
risk. Br. J. Pharmacol. 2010, 159, 12−21.
(21) Osterberg, F.; Aqvist, J. Exploring blocker binding to a
homology model of the open hERG K+ channel using docking and
molecular dynamics methods. FEBS Lett. 2005, 579, 2939−2944.
(22) Stansfeld, P. J.; Gedeck, P.; Gosling, M.; Cox, B.; Mitcheson, J.
S.; Sutcliffe, M. J. Drug block of the hERG potassium channel: insight
from modeling. Proteins 2007, 68, 568−580.
(23) Masetti, M.; Cavalli, A.; Recanatini, M. Modeling the hERG
potassium channel in a phospholipid bilayer: molecular dynamics and
drug docking studies. J. Comput. Chem. 2008, 29, 795−808.
(24) Cramer, R. D.; Patterson, D. E.; Bunce, J. D. Comparative
molecular field analysis (CoMFA). 1. Effect of shape on binding of
steroids to carrier proteins. J. Am. Chem. Soc. 1988, 110, 5959−5967.
(25) Mitcheson, J. S.; Chen, J.; Lin, M.; Culberson, C.; Sanguinetti,
M. C. A structural basis for drug-induced long QT syndrome. Proc.
Natl. Acad. Sci. U.S.A. 2000, 97, 12329−12333.
(26) Fernandez, D.; Ghanta, A.; Kauffman, G. W.; Sanguinetti, M. C.
Physicochemical features of the HERG channel drug binding site. J.
Biol. Chem. 2004, 279, 10120−10127.
(27) Tilley, J. W.; Levitan, P.; Welton, A. F.; Crowley, H. J.
Antagonists of slow-reacting substance of anaphylaxis. 1. Pyrido[2,1-
b]quinazolinecarboxylic acid derivatives. J. Med. Chem. 1983, 26,
1638−1642.
4014
dx.doi.org/10.1021/jm201194q | J. Med. Chem. 2012, 55, 4010−4014