Oxidation potentials of N-modified derivatives of the analgesic flupirtine linked to potassium KV7 channel opening activity but not hepatocyte toxicity

Article abstract of DOI:10.1002/cmdc.201402442

Openers of neuronal voltage-gated potassium channels (K_V) are of interest as therapeutic agents for treating pain (flupirtine) and epilepsy (retigabine). In an effort to better understand the mechanisms of action and toxicity of flupirtine, we synthesized nine novel analogues with varying redox behavior. Flupirtine can be oxidatively metabolized into azaquinone di-imines; thus, the oxidation potentials of flupirtine and its analogues were measured by cyclic voltammetry. K_V7.2/3 (KCNQ2/3) opening activity was determined by an established assay with HEK293 cells overexpressing these channels. A link was found between the oxidation potentials of the compounds and their EC₅₀ values for potassium channel opening activity. On the other hand, no correlation was observed between oxidation potentials and cytotoxicity in cultures of transgenic mouse hepatocytes (TAMH). These results support the idea that oxidative metabolites of flupirtine contribute to the mechanism of action, similar to what was recently proposed for acetaminophen (paracetamol), but not to hepatotoxicity.

Full text of DOI:10.1002/cmdc.201402442

CHEMMEDCHEM
FULL PAPERS
DOI: 10.1002/cmdc.201402442
Oxidation Potentials of N-Modified Derivatives of the
Analgesic Flupirtine Linked to Potassium KV7 Channel
Opening Activity But Not Hepatocyte Toxicity
Christian J. Lemmerhirt, Mirko Rombach, Anja Bodtke, Patrick J. Bednarski, and
Andreas Link*[a]
In memory of Professor Sidney D. Nelson
Openers of neuronal voltage-gated potassium channels (KV)
are of interest as therapeutic agents for treating pain (flupir-
tine) and epilepsy (retigabine). In an effort to better under-
stand the mechanisms of action and toxicity of flupirtine, we
synthesized nine novel analogues with varying redox behavior.
Flupirtine can be oxidatively metabolized into azaquinone di-
imines; thus, the oxidation potentials of flupirtine and its ana-
logues were measured by cyclic voltammetry. KV7.2/3 (KCNQ2/
3) opening activity was determined by an established assay
with HEK293 cells overexpressing these channels. A link was
found between the oxidation potentials of the compounds
and their EC50 values for potassium channel opening activity.
On the other hand, no correlation was observed between oxi-
dation potentials and cytotoxicity in cultures of transgenic
mouse hepatocytes (TAMH). These results support the idea
that oxidative metabolites of flupirtine contribute to the mech-
anism of action, similar to what was recently proposed for
acetaminophen (paracetamol), but not to hepatotoxicity.
Introduction
Flupirtine (1, Figure 1) is approved for use in a number of Euro-
pean countries as well as in Brazil and India as a centrally
acting analgesic, offering an alternative to non-opioid, nonster-
oidal anti-inflammatory drugs in patients that have contraindi-
cations to weak opioids or NSAIDs.[1] The mechanism of action
of flupirtine has been attributed to the ability of the drug to
open voltage-gated K+ channels (KV) in the central nervous
system, which indirectly leads to antagonism of N-methyl-d-as-
partate (NMDA) receptors.[2]
Depending on the primary factors that lead to channel
opening and closing, ion channels are divided into ligand- or
voltage-gated subfamilies, with or without a physiological
ligand binding site.[3] Even in the latter case, modulation by
small molecules is conceivable, and both subfamilies are drug-
gable. KV7 channels (also referred to as KCNQ after the corre-
sponding genes) belong to the voltage-gated K+ channel su-
perfamily. The five members (KV7.1–KV7.5) are six-transmem-
brane channels based on four subunits and an appended
transmembrane helix voltage-sensor domain.[4] The four subu-
nits create a central ion-conducting pore with a constricted se-
lectivity filter at the top, surrounded by the voltage-sensing
domain (VSD). The VSD responds to membrane depolarization
Figure 1. Flupirtine (1), retigabine (2), and their putative oxidation products
(1a,b and 2a,b, respectively) in comparison with acetaminophen (3) and its
known metabolite N-acetyl-p-quinone imine (NAPQI, 3a).
[a] C. J. Lemmerhirt, M. Rombach, Dr. A. Bodtke, Prof. Dr. P. J. Bednarski,
Prof. Dr. A. Link
and initiates opening of the gate at the base of the pore
through concerted movements of all subunits. In this way, the
ion channel converts between the closed and open states.
KV7.2 channels are expressed at high levels in the brain,
often as functional heterotetramers with KV7.3.[5,6] These het-
erotetrameric channels are also referred to as KCNQ2/3. A
Pharmaceutical and Medicinal Chemistry
Institute of Pharmacy, Ernst Moritz Arndt University
Friedrich-Ludwig-Jahn-Str. 17, 17487 Greifswald (Germany)
E-mail: link@uni-greifswald.de
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201402442.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 13
&
1
&
ÞÞ
These are not the final page numbers!

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
number of positive and negative KV7.2/3 channel modulators
have been reported, and the discovery of compounds that
affect these channels has attracted attention in the develop-
ment of novel analgesics.[7–9] KV7 openers, including flupirtine
and retigabine (2, an antiepileptic drug), are also potential
tools for understanding ion channel function, especially be-
cause considerable clinical data are available. However, it can
be difficult to draw firm conclusions from these data because
of poorly characterized polypharmacology.
Results
Chemistry
Synthesis of carbamate methylated flupirtine/retigabine
Regarding the nitrogen-substituted pyridine ring of flupirtine
(1) it can be assumed that the nitrogen atom incorporated
into the carbamate group is weakly acidic, whereas the other
two aromatic amines have weak basic properties. A prerequi-
site for the desired regioselective monoalkylation was there-
fore to exclusively improve the nucleophilicity of the appropri-
ate nitrogen atom. Deprotonation of the carbamate group
with subsequent addition of methyl iodide was applied
(Scheme 1) by using established procedures[19] for the alkyla-
Although rare, the most severe adverse effect of flupirtine is
related to hepatocellular toxicity.[10] The incidence of flupirtine-
related hepatotoxicity was estimated to be about eight in
100000 patients based on a re-evaluation of 226 unselected,
spontaneously reported cases in the German Health Authority
(BfArM) database.[11] Nonetheless, the Drug Commission of the
German Medical Association issued a drug safety warning in
2013 regarding the use of flupirtine in the management of
chronic pain.[12] Thus, an understanding of the molecular phar-
macology of the various channel subfamily members should
be complemented by toxicology data.
Flupirtine’s toxicity may be related to its metabolism, which
has been shown to undergo both oxidative and conjugative
reactions both in vitro[13] and in vivo.[14,15] Although the cyto-
chrome P450 system does not appear to be involved in oxida-
tion, peroxidases easily oxidize flupirtine to azaquinone diimine
metabolites (1a,b).[13] These are conjugated with glutathione
without the need for catalysis and after further metabolism
they appear in the urine as mercapturic acids. Despite the
striking molecular similarity between flupirtine and retigabine,
hepatotoxicity has never been reported to be a problem with
the latter. This goes to show that minute alterations in elec-
tronic features and/or reactivity might be important in terms
of drug safety and a deeper understanding of underlying
mechanisms. NAPQI (3a), the reactive quinone imine metabo-
lite formed during oxidation of acetaminophen (3), has been
implicated in both the activity[16] and toxicity[17,18] of this widely
used analgesic. In analogy to 3a, the azaquinone diimine me-
tabolites of flupirtine 1a,b may also be involved in both the
mechanism of action as well as drug hepatotoxicity.
Scheme 1. Selective alkylation of the carbamate nitrogen atom. Reagents
and conditions: a) NaH, CH3I, THF, 08C, 30 min, then CH3I, 08C!RT, over-
night.
tion of carbamates. Alkyl bromides with longer alkyl chains led
to better results than the corresponding iodides. However, in
the case of methylation, methyl iodide was preferred over gas-
eous methyl bromide due to safety and handling concerns.
Chromatographic analysis of the reaction products indicated
that only minor changes in lipophilicity resulted from methyla-
tion while the photochemical stability was altered.
Herein we report the synthesis of a series of novel flupirtine
analogues and their use in studying the activity at the pro-
posed molecular target as well their potential to cause toxicity
in liver cell culture models. These analogues closely resemble
flupirtine, but possess substituents designed to interfere with
azaquinone diimine formation by inductive electronic effects,
replacement of removable hydrogen atoms, and steric bulk.
Additionally, one derivative lacks an amino group necessary for
the formation of ortho-azaquinone diimine. The altered redox
properties associated with modified reactivity makes them
ideal probes for these studies.
Synthesis of bicyclic flupirtine derivatives
Another possible way to modify the properties of the carba-
mate nitrogen atom is to form a bicyclic derivative. A bridge
between nitrogen atoms on C2 and C3 of the flupirtine pyri-
dine ring, forming a five-membered ring, would leave the
backbone of the molecule unaltered, but would slightly
change the angle of the CꢀN bonds on C2 and C3. Attempts
to force ring closure by insertion of formic acid to obtain a me-
thine bridge or cyanogen bromide to form a guanidine group
were not successful. However, dry heating of 1 in an open
vessel, following similar solvent-free procedures[20,21] led to 6 in
good yield (Scheme 2).
Compound 6 is a known process impurity formed during
large-scale synthesis of flupirtine and was recently identified
and synthesized by Zhang et al.[22] Our straightforward prepara-
tion of 6 by dry heating, however, is advantageous over the
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 13
&
2
&
ÞÞ
These are not the final page numbers!

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
Scheme 2. Dry heating of flupirtine (1) followed by re-acylation. Reagents
and conditions: a) D (2008C), 30 min; b) ethyl pyridine-2-yl carbonate, DMF,
MeCN, 758C, overnight.
method described by Zhang et al., who used 2,3-diamino-6-(4-
fluorobenzylamino)pyridine, a compound extremely sensitive
to oxidation, as an intermediate in a multistep sequence.
As a consequence of ring closure by dry heating, the ethyl
carbamate group of 1 was lost. To improve the molecular simi-
larity of the bicyclic derivative 6 and flupirtine, we envisioned
reintroduction of such an ethyl carbamate substituent to 6.
This was achieved by use of an acylating carbonate with high
electrophilicity brought about by low mesomeric stabilization
due to incorporation of an electron-deficient aromatic ring.
Thus, ethyl pyridin-2-yl carbonate was prepared as described[23]
and used without further purification. The main product of the
reaction of this electrophile with the cyclic urea 6 was the de-
sired carbamate 7, but a small quantity of a second com-
pound, 8, bearing two carbamate groups, was isolated. The
formation of compound 8 was not expected, as Zhang et al.
only found the regioisomer with N1 and N3 acylated among
the process impurities.[22]
Scheme 3. Synthesis of flupirtine derivatives, introducing methyl groups at
intermediary reaction steps (a, d, and g). Reagents and conditions: a) CH3NH2,
MeOH, reflux, overnight; b) NaBH4, MeOH, 08C!RT, overnight; c) conc.
HNO3, conc. H2SO4, reflux (~100–1058C), overnight; d) (for 15): aq. NH3, 2-
propanol, 358C, 3 d; (for 16): (CH3)2NH, Et3N, MeCN, 08C!RT, 30 min; e) (for
17): 15, 12, 2-propanol, Et3N, reflux, overnight; (for 18, 19): 4-fluorobenzyl-
amine (11 or 12), 16, Et3N, MeCN, RT, overnight, then reflux, ꢁ24 h; f) (for
20): 17, Pd/C, H2, EtOH, overnight, then Et3N, ethyl chloroformate, 4 h; (for
21, 22): nitropyridine (18 or 19), Pd/C, H2, EtOH, overnight, then Et3N, ethyl
chloroformate, 8 h; g) 20, NaH, THF, 08C, 30 min, then CH3I, 08C!RT, over-
night.
tine,[24–26] which allows the regioselective introduction of
methyl groups in intermediary reaction steps (Scheme 3).
Whereas 4-fluorobenzylamine (11) was commercially avail-
able, the N-methylated secondary amine 12 had to be pre-
pared from 4-fluorobenzaldehyde (9) by reductive amination in
a two-step synthesis.[27,28] The second component was the
chlorinated pyridine ring 13, which was nitrated by using hot
nitrosulfuric acid to obtain 14. In the next step, selective nucle-
ophilic aromatic substitution was achieved; in this reaction,
ammonia was far less reactive than dimethylamine, which was
an advantage in matters of selectivity, and thus 15 was far
more easily purified than 16.
Synthesis of N-methylated derivatives following an
established route for the synthesis of flupirtine
Direct alkylations of flupirtine in our hands always yielded N-
methylated flupirtine derivatives with substitution at the carba-
mate nitrogen only. To obtain derivatives with a divergent N-al-
kylation pattern, multistep syntheses were required. In retro-
spect, there were no conformities between any side products
of the synthesis of 4 detected by thin-layer chromatography
(TLC) and any monoalkylated compounds described later on.
This indicates that the alkylation of N2 and N6 cannot be ac-
complished by direct alkylation with standard procedures. We
focused on a well-described route for the synthesis of flupir-
The subsequent transformation of compounds 15 and 16 in
a second nucleophilic substitution with the 4-fluorobenzyla-
mines 11 and 12 at elevated temperatures afforded 17–19.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 13
&
3
&
ÞÞ
These are not the final page numbers!

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
Again, the derivatives with a dimethylamino group proved to
be more difficult to purify, as the resulting compounds (18 and
19) were not crystalline like 17, but brown sticky oils. Whereas
18 has been reported previously in a patent[29] along with simi-
lar derivatives with various N2 substituents,[30–32] 19 has not
been reported until now.
Analytical investigations
Cyclic voltammetry studies
Electrochemistry can be used for mimicking some types of bio-
logical reactions, such as oxidative drug metabolism, including
formation of reactive metabolites. Specifically, cyclic voltam-
metric determination of anodic peak potential is suited to indi-
cate chemical properties in the context of oxidation; for exam-
ple, the electrochemical properties of the widely used analge-
sic diclofenac were studied by cyclic voltammetry by Madsen
et al.[33,34]
Subsequent hydrogenation of the nitro groups of 17, 18,
and 19 was carried out without workup of the resulting amino
intermediates because of their notorious sensitivity toward
oxygen. Thus, ethyl chloroformate and triethylamine were
added directly to the reaction mixture in the closed vessel and
allowed to react for several hours.
Flupirtine (1) undergoes facile autoxidation even when dis-
solved in aqueous solution at room temperature, as can be vi-
sualized directly by a change from a colorless solution to one
with a blue–green hue within hours. Peroxidases are known to
accelerate this reaction considerably, leading to reactive aza-
quinone diimines that can react further to form dimers and
trimers of flupirtine.[13] Thus, we investigated the oxidation po-
tentials of the new analogues by cyclic voltammetry and com-
pared them with that of flupirtine (Figure 2).
Lastly, crystallization was enforced by the addition of a mix-
ture of water and 2-propanol, as described for the maleate salt
of flupirtine.[26] In the case of 20, this could also be achieved,
but even the free base was directly obtained in crystalline
form. On the other hand, the crystalline salts of 21 and 22
could not be obtained in an analogous fashion, neither with
nor without maleic acid. Thus, the crude products were sub-
jected to liquid–liquid extraction by using dilute hydrochloric
acid and dichloromethane, where the salts 21·HCl and 22·HCl
fractioned into the organic phase. These salts could, in part, be
obtained in solid form by lyophilization. Recrystallization was
achieved by dissolution in ethyl acetate followed by the addi-
tion of n-hexane.
Compound 20 was alkylated with methyl iodide in an analo-
gous manner to the preparation of 4 and 5 to obtain the di-
methylated derivative 23, which required an even more de-
manding workup und crystallized only slowly.
Synthesis of deaminoflupirtine derivatives
Starting from 2-chloro-5-nitropyridine (24), the flupirtine ana-
logues 27 and 28 could be obtained in just three steps as
their hydrochloride salts (Scheme 4). The synthetic route was
analogous to that of 21 and 22, yielding the solid intermedi-
ates 25 and 26. Nucleophilic aromatic substitution again was
driven by the electron-withdrawing effect of the nitro group
and the electron deficiency of the pyridine ring.
Figure 2. Cyclic voltammograms of selected compounds in 100 mm Tris
buffer (pH 7.4); shown is the first cycle, swept through ꢀ1.0 to 1.0 V, display-
ing only ꢀ0.5 to 1.0 V.
Figure 2 shows representative cyclic voltammograms of flu-
pirtine and three of the analogues, which are characterized by
more or less well-defined anodic peaks between 0.25 and
0.75 V (Table 1). In contrast, the cathodic peaks were very weak
or nonexistent, evidence of the irreversible nature of the redox
process, and consistent with the high reactivity of the azaqui-
none diimine products, which deposit as a polymer film on the
surface of the electrode. In fact, after each cyclic voltammetric
run, the glassy carbon electrode had to be polished, or else
the next cyclic voltammogram showed poorly formed peaks of
decreased intensity.
Interestingly, the most intense anodic and cathodic peaks
were found with the N,N-dimethyl-substituted analogue 21
(Figure 2); substitution of the primary aromatic amine likely
prevents nucleophilic attack of the azaquinone diimine struc-
ture, lowering the propensity to form polymers. Nevertheless,
substitution of either the primary or secondary aromatic
amines of flupirtine with methyl groups (21 and 20, respective-
ly) had little effect on the position of the anodic peak poten-
tials (Ep.a) relative to flupirtine. On the other hand, substitution
Scheme 4. Synthesis of deaminoflupirtine derivatives. Reagents and condi-
tions: a) Et3N, MeCN, RT, overnight, then reflux, 24 h; b) Pd/C, H2, EtOH, over-
night, then Et3N, ethyl chloroformate, 8 h.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 13
&
4
&
ÞÞ
These are not the final page numbers!

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
whereas the inactive derivatives have considerably
greater Ep.a values (Table 1).
Table 1. Comparisons of anodic peak potentials (Ep.a) of flupirtine (1) and derivatives,
EC50 values toward KV7.2/3 channels in HEK293 cells, and LD50 values in TAMH cells.
In comparing the active and inactive compounds
based on their anodic peak potentials by the Mann–
Whitney U-test, a highly significant difference was
found between the Ep.a of the two groups of com-
pounds (Figure 4a). This indicates that more easily
oxidized flupirtine derivatives are more potent than
those less easily oxidized. In fact, a slight trend
toward greater potency of the active compounds can
be observed with decreasing Ep.a values (Table 1).
Compound
Ep.a [mV][a]
EC50 [mm][b]
LD50 [mm][c]
20
294
299
320
350
385
1.9ꢂ0.1
1.9ꢂ0.2[d]
2.6ꢂ0.4
3.6ꢂ0.9[e]
3.5ꢂ0.7
94ꢂ34
174ꢂ27
119ꢂ49
87ꢂ60[e]
35ꢂ17
retigabine (2)
21·HCl
flupirtine (1)
22·HCl
5
7
591
612
642
652
653
667
>100
205ꢂ53
202ꢂ57
204ꢂ47
131ꢂ67
69ꢂ33
>100
>100
>100
>100
>100
27·HCl
4
28·HCl
23
Cytotoxicity in TGF-a transgenic mouse hepatocytes
125ꢂ61
TGF-a transgenic mouse hepatocytes (TAMH cells)
were used to estimate the hepatotoxicity of the com-
pounds in cell culture. This cell line overexpresses
TGF-a, but retains many of the characteristics of hep-
atocytes while replicating for months in culture.[37]
Acute cytotoxicity was estimated by exposing conflu-
ent cell cultures to test compounds for 24 h and
measuring cell viability by MTT assay. Compound
concentrations that decreased cell viability by 50%
relative to untreated controls were defined as LD50.
Figure 5 shows dose–response curves for three representative
compounds in the TAMH cell line.
acetaminophen (3)
N-methylacetaminophen (29)
602
773
ND[f]
ND[f]
>1000
>1000
[a] Determined with 1.0 mm compound in 100 mm Tris buffer (pH 7.4); values are the
mean ꢂSD of 4–5 independent determinations. [b] Determined after 30 min expo-
sure; values are the mean ꢂSD of 4–5 independent determinations. [c] Determined
after 24 h exposure. [d] Ref. [36]. [e] Flupirtine maleate salt was used. [f] Not deter-
mined.
of the carbamate nitrogen (4, 5, 7, 23) as well as lack of the
amino group on C2 (27·HCl and 28·HCl) caused a very notable
shift in the anodic potential of about +0.3 V.
Table 1 summarizes the LD50 values for flupirtine and the var-
ious analogues, of which some show relatively high and
others, like flupirtine, low Ep.a values. It is clear that there is no
correlation between Ep.a and LD50 values. Moreover, Figure 4b
shows that there is no significant difference between the LD50
values for the active and inactive compounds; i.e., KV7.2/3
channel opening activity and hepatotoxicity are not linked.
Biology
In vitro tests on KV7.2/3 channels
The nine synthesized flupirtine derivatives were tested in
a FLIPRꢁ-based assay for their ability to increase KV7.2/3 potas-
sium channel activity relative to flupirtine. Special HEK293 cells
co-expressing the KCNQ2 and KCNQ3 genes were used in this
assay, which was carried out by using a FluxOR potassium ion
channel assay (Life Technologies).[35] In this assay, a specific
thallium-sensitive dye is loaded into the HEK293 cells which is
then cleaved by esterase activity to the active Tl-complexing
agent, which forms fluorescent complexes with thallium ions
that enter the cells selectively via the opened KV7.2/3 potassi-
um channel. The intensity of cellular fluorescence is an indica-
tion of the efficacy of ion channel opening.
Discussion
Earlier structure–activity relationship (SAR) studies with flupir-
tine were concerned mainly with the effect of the substitution
pattern in the phenyl ring, as well as substitutions at the car-
Figure 3 shows the dose–response curves for the active
compounds 1, 20, 21, and 22 as KV7.2/3 channel openers.
Table 1 reports the average EC50 values. All flupirtine deriva-
tives with a substituted carbamate nitrogen atom were ineffec-
tive at channel opening up to 100 mm. The two deaminoflupir-
tine derivatives 27·HCl and 28·HCl were not completely inac-
tive at the measured concentrations, but estimates of their
EC50 values were not possible, as they were >100 mm, the
highest concentration used (Figure 3). In contrast, derivatives
that are methyl substituted at only N2 and/or N6 (20, 21·HCl,
and 22·HCl) had EC50 values in the same low micromolar range
as flupirtine. Interestingly, these are the same three flupirtine
derivatives that show similar Ep.a values to that of flupirtine,
Figure 3. Dose–response curves of 20, 21·HCl, 22·HCl, and 27·HCl compared
with that of flupirtine (1) in the KV7.2/3 channel-opening assay. Relative fluo-
rescence intensity was measured by setting stimulus buffer alone at 0%,
and 100 mm flupirtine maleate at 100%; data are the average ꢂSD of four
independent determinations.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 13
&
5
&
ÞÞ
These are not the final page numbers!

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
Figure 4. Statistical evaluation of the differences between active and inactive
compounds with respect to a) Ep.a and b) LD50. Significance: **p<0.01 by
two-tailed Mann–Whitney U-test (GraphPad Prism 6.0).
Figure 6. Structures of compounds active as KV7.2/3 channel openers (left)
compared with their inactive analogues (right). Shaded areas highlight struc-
tural differences from the lead compound flupirtine (1).
SAR of the 11 compounds investigated in this work with re-
spect to their ability to open KV7.2/3 channels.
The ease of oxidation, based on the cyclic voltammetric
anodic peak potentials of the N-methylated analogues relative
to flupirtine (1), can be rationalized based on their structures.
For example, if the carbamate nitrogen atom is methylated
(i.e., 4 and 23), oxidation to an azaquinone diimine would
place a permanent positive charge on that carbamate nitrogen
directly adjacent to a partial positively charged carbonyl
carbon, which is energetically unfavorable. On the other hand,
methylation at N2 and N6 (i.e., 20–22) can theoretically facilitate
oxidation due to the electropositive nature of the two methyl
groups on that nitrogen. Furthermore, the primary amino
group of flupirtine would be expected to increase the electron
density in the pyridine ring owing to resonance effects, which
should also facilitate oxidation and explain the higher anodic
peak potentials of the derivatives without this specific group
(i.e., 27 and 28).
Figure 5. Dose-dependent cytotoxicity of flupirtine (1) and representative
derivatives in TAMH cells after 24 h exposure, determined by the MTT
method. Data are the average ꢂSD of four independent determinations.
bamate oxygen atom, on analgesic activity.[38] Until now, no re-
ports have been published concerning the effects of N-alkyla-
tion either on activity or toxicity. In this study, all nitrogen
atoms attached to the pyridine ring of flupirtine were selec-
tively methylated, although only the carbamate nitrogen could
be alkylated directly from flupirtine itself. All nine N-modified
flupirtine derivatives were previously unreported, thus offering
a unique opportunity to conduct SAR studies with respect to
oxidation potential, KV7.2/3 channel-opening activity, and hep-
atotoxicity of this widely used drug. Figure 6 summarizes the
SAR clearly show that substitution of the proton on the car-
bamate nitrogen atom with methyl group causes a total loss
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 13
&
6
&
ÞÞ
These are not the final page numbers!

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
of KV7.2/3 channel-opening activity while dramatically increas-
ing the oxidation potential by +300 mV. This observation is
reminiscent of earlier studies with the analgesic acetamino-
phen, in which N-methylation increases the oxidation potential
by +171 mV, resulting in a complete loss of analgesic activi-
ty.[39] This is consistent with recent evidence indicating that re-
active quinone imine metabolite of acetaminophen, NAPQI
(3a), is the active agent responsible for analgesic activity.[16]
An alternative explanation for the poor KV7.2/3 channel-
opening activity of the analogues with an N-methylated carba-
mate could be that the loss of the carbamate hydrogen results
in the loss of a critical hydrogen bond within the binding site
of the potassium channel. However, 27 and 28, both of which
belong to the group with higher oxidation potentials (Table 1),
retain this hydrogen, but are still poor channel openers. Thus,
oxidation potential appears more closely correlated with good
channel-opening activity. The primary aromatic amino group
of flupirtine also appears important for good channel-opening
activity, as analogues 27 and 28, which lack it, are notably less
active. In contrast, methylation at either N2 or N6 resulted in
analogues (i.e., 21 and 22) with good channel-opening activity,
but methylation also had little effect on their oxidation poten-
tials (Table 1).
with acetaminophen (3), for which N-methylation led to an in-
crease in Ep.a of 171 mV and also protected against hepatotox-
icity. Nelson and co-workers rationalized this by the failure of
N-methylacetaminophen (29) to oxidatively form the hepato-
toxic NAPQI.[39] In the case of flupirtine, however, hepatotoxici-
ty would not appear to be directly linked to oxidative process-
es. The very different physicochemical properties of acetamino-
phen (hydrophilic, weak acid) relative to flupirtine (lipophilic,
weak base) could result in differences as to how these com-
pounds are taken up and compartmentalized by hepatocytes.
Ongoing studies on the biochemical changes in hepatocytes
treated with flupirtine and various analogues are aimed at
gaining a better understanding of the mechanism of toxicity.
Conclusions
While the redox behavior appears to be linked to the pharma-
cological activity of flupirtine, as is also the case with acetamin-
ophen, unlike acetaminophen, it does not appear to play a cen-
tral role in hepatocyte toxicity. Thus, analgesic activity and
hepatotoxicity would appear separable in this class of drugs.
This knowledge should be useful for the rational design of
new KV7.2/3 channel openers.
It is not yet known if these effects at the KV7.2/3 channel
translate directly to analgesic activity in animals. However,
a number of other variables will affect in vivo activity that do
not factor into an in vitro system, e.g., pharmacokinetics and
metabolism. Thus, mechanism of action studies with a simpler
cellular system would appear suitable for defining SAR.
Experimental Section
General
Reaction control by TLC was carried out with ALUGRAM SIL G/
UV254 silica gel plates from Macherey–Nagel (4ꢂ8 cm, silica layer:
20 mm, substances plotted in 1 cm bends). Melting Points were
measured with a Kofler hot-stage microscope (Boꢃtius PHMK 81/
3035, VEB Wꢄgetechnik Rapido) with 16-fold amplification. Com-
pound purities were determined after HPLC by DAD chromato-
grams (100% method, 254 nm). NMR spectra were recorded with
a Bruker Avance III instrument with Ultrashield 400 (258C, ppm
scale) internal standard, resonance frequency, and solvent as indi-
cated. HRMS data were obtained after HPLC with an LC-IT-TOF
mass spectrometer (Shimadzu). IR spectra were recorded on a Nico-
let IR200 FT-IR instrument (Thermo Electron Corporation) with dia-
mond ATR accessory. For cyclic voltammetry, a 797 CA Computrace
from Metrohm was used.
There is growing interest in the regulation of potassium
channels by the cellular redox environment.[40] Studies indicate
that these heteromeric KV7.2/3 channels are sensitive to oxida-
tion, for example, through hydrogen peroxide mediated oxida-
tion of cysteine residues, which affects current activity.[41] Once
bound to the channel in the oxidized form (i.e., as the azaqui-
none diimine 1a or 1b), flupirtine might participate directly in
such redox reactions, facilitating the opening of KV7.2/3 chan-
nels. This could explain why easily oxidized analogues are
better channel openers than those less easily oxidized. Under-
standing how the cellular redox environment influences the
KV7.2/3 channel-opening properties of flupirtine and deriva-
tives will be at the center of future research.
For evaluation of the hepatotoxic potential of flupirtine and
derivatives, an established in vitro model system based on cul-
tured TAMH cells was used. TAMH cells show many of the
same attributes as hepatocytes,[37] and have been used to
study mechanisms of hepatotoxicity in drugs such as acet-
aminophen,[42] flutamide,[43] and lapatinib[44] (not shown).
In the case of flupirtine, a compound known to cause liver
toxicity in a small group of patients, 24 h exposure led to a no-
table decrease in cell viability at concentrations below 100 mm.
In fact, all the analogues tested showed roughly the same cy-
totoxic potency regardless of their oxidation potential
(Table 1). In particular, 4 showed no significant decrease in cy-
totoxicity relative to flupirtine, even though methylation of the
carbamate nitrogen atom led to an increase in oxidation po-
tential by ~300 mV. This is in contrast to the case observed
Chemistry
Ethyl [2-amino-6-(4-fluorobenzylamino)pyridine-3-yl]carbamate
(flupirtine; 1): Five capsules of Trancopal Dolo (each containing
100 mg flupirtine maleate salt) were cut open, and the contents
were added to a separation funnel containing H2O (50 mL), EtOAc
(50 mL) and Et3N (1 mL). After shaking the mixture, the insoluble
components were filtered off, and the organic layer of the filtrate
was washed once with brine (50 mL), dried over Na2SO4 and con-
centrated in vacuo. Light-yellow solid, turning grey (which subse-
quently turned brown) upon exposure of the substance on the
TLC plate to UV light for 15 min (315 mg, 87%); mp: 114–1168C
(from EtOAc); 1H NMR (400 MHz, [D6]DMSO): d=1.20 (brs, 3H,
CH3), 4.03 (q, J=6.8 Hz, 2H, O-CH2), 4.35 (d, J=6.0 Hz, 2H, C(a’)H2),
5.19 (s, 2H, N2H2), 5.67 (t, J=8.0 Hz, 1H, C(5)H), 6.52 (t, J=6.0 Hz,
1H, N6H), 7.01 (m, 1H, C(4)H), 7.10 (m, 2H, C(3’)H and C(5’)H), 7.34
(m, 2H, C(2’)H and C(6’)H), 8.17 ppm (brs, 1H, N3H); 13C NMR
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 13
&
7
&
ÞÞ
These are not the final page numbers!

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
(100 MHz, [D6]DMSO): d=14.5, 43.5, 59.1, 106.5, 114.7, 128.9, 135.6,
137.2, 153.0, 155.0, 155.6, 159.7, 162.1 ppm; IR: n˜ =3371, 3202,
118.0, 128.1, 128.9, 136.8, 144.6, 148.2, 155.9, 159.8, 162.2 ppm; IR:
n˜ =3428, 3381, 3340, 2981, 2856, 1686, 1507, 828 cmꢀ1; HRMS-ESI
m/z [M+H]+ calcd for C17H20N3O2F: 318.1612, found: 318.1603;
Anal. calcd for C17H20N3O2F: C 64.3, H 6.35, N 13.2, found: C 64.0, H
6.0, N 13.1.
2985, 2904, 2841, 1687, 1616, 1495, 1213 cmꢀ1
.
Ethyl [2-amino-4-(4-fluorobenzylamino)phenyl]carbamate (reti-
gabine; 2): One film-coated tablet of Trobalt (containing 300 mg
retigabine) was pulverized in a mortar and diluted with Et2O
(25 mL) and brine (25 mL). After shaking the mixture, the insoluble
components were filtered off, and the organic layer of the filtrate
was washed once with brine (25 mL), dried over Na2SO4 and con-
centrated in vacuo. Light-yellow solid, evolving a pale-pink color
when the substance on a TLC plate was exposed to UV light for
15 min (250 mg, 83%); mp: 136–1388C (from Et2O); 1H NMR
(400 MHz, [D6]DMSO): d=1.19 (brs, 3H, CH3), 4.02 (q, J=7.2 Hz,
2H, O-CH2), 4.16 (d, J=6.0 Hz, 2H, C(a’)H2), 4.52 (s, 2H, N2H2), 5.82
(dd, J=8.8 Hz, J=2.4 Hz, 1H, C(3)H), 5.89 (t, J=6.0 Hz, 1H, N4H),
5.92 (d, J=2.4 Hz, 1H, C(3)H), 6.79 (d, J=8.0 Hz, 1H, C(6)H), 7.12
(m, 2H, C(3’)H and C(5’)H), 7.35 (m, 2H, C(2’)H and C(5’)H)
8.14 ppm (s, 1H, N1H); 13C NMR (100 MHz, [D6]DMSO): d=14.7,
45.8, 59.8, 98.9, 101.6, 113.0, 114.8, 127.0, 128.8, 136.8, 143.4, 147.1,
155.1, 159.8, 162.2 ppm; IR: n˜ =3421, 3375, 3268, 3187, 3037, 2984,
5-(4-Fluorobenzylamino)-1,3-dihydro-2H-imidazo[4,5-b]pyridine-
2-one (6): Compound 1 (2 mmol, 608 mg) was heated at 2008C for
30 min in an open vessel. After cooling, the crude product was
washed with a small volume of MeOH and dried under reduced
pressure. Light-brown solid (538 mg, 88%); purity: 100%; mp:
242–2448C; 1H NMR (400 MHz, [D6]DMSO): d=4.37 (d, J=6.0 Hz,
2H, C(a’)H2), 6.08 (d, J=8.4 Hz, 1H, C(6)H), 6.61 (t, J=6.0 Hz, 1H,
N5H), 6.98 (d, J=8.0 Hz, 1H, C(7)H), 7.12 (m, 2H, C(3’)H and C(5’)H),
7.35 (m, 2H, C(2’)H and C(6’)H), 10.18 (s, 1H, N(1)H) and 10.79 (s,
1H, N(3)H); 13C NMR (100 MHz, [D6]DMSO): d=44.1. 99.1, 113.7,
114.7, 117.7, 128.9, 136.9, 142.8, 153.8, 154.1, 159.7, 162.1 ppm; IR:
n˜ =3433, 3132, 2997, 2829, 1682, 1613, 1219 cmꢀ1; HRMS-ESI m/z
[M+H]+ calcd for C13H11N4OF: 259.0990, found: 259.0985; Anal.
calcd for C13H11N4OF: C 60.5, H 4.3, N 21.7, found: C 60.1, H 4.1, N
21.5.
2918, 2850, 1697, 1507, 1214, 825 cmꢀ1
.
General procedure B
General procedure A
Compound 6 (1 mmol, 258 mg) and ethyl pyridine-2-yl carbonate
(2 mmol; 333 mg) were suspended in a mixture of DMF (2.5 mL)
and MeCN (2.5 mL) and stirred overnight in a capped round-
bottom flask at 758C. The reaction was concentrated under re-
duced pressure and acidified with HCl (1m) to pH~1, filtered, and
the residue was washed with H2O and dried in vacuo. The ob-
tained crude product was purified by column chromatography (n-
hexane/EtOAc gradient) with silica gel.
Compound 1 or 2 (1 mmol) and NaH (1.5 mmol, 60 mg; 65% solu-
tion in n-hexane) were set under an argon atmosphere, stirred in
THF (12 mL), cooled to 08C in the dark. After 30 min CH3I (2 mmol,
124 mL) was added, then stirred overnight while the reaction was
allowed to reach room temperature. TLC of the reaction progress
showed almost no change in Rf values, but a change in color
under UV light occurred after 15 min. The reaction was deemed
complete when the spot was a consistent orange color without
any hue of the original substance on the lower rim. Afterward, the
solvent was removed under reduced pressure, the residue diluted
with CH2Cl2 (50 mL) and then washed with brine (50 mL). The or-
ganic layer was dried over Na2SO4, concentrated in vacuo, and if
necessary lyophilized to give a brown solid, which was washed
with small amounts of Et2O to give the desired product.
Ethyl 5-(4-fluorobenzylamino)-2-oxo-2,3-dihydro-1H-imidazo[4,5-
b]pyridine-1-carboxylate (7): Following general procedure B yield-
ed a light-beige solid (70 mg, 21%); purity: 100%; mp: 206–2088C
(from n-hexane/EtOAc); 1H NMR (400 MHz, [D6]DMSO): d=1.32 (t,
J=7.2 Hz, 3H, CH3), 4.34 (q, J=7.2 Hz, 2H, O-CH2), 4.40 (d, J=
6.0 Hz, 2H, C(a’)H2), 6.19 (d, J=8.8 Hz, 1H, C(6)H), 7.05 (t, J=
6.0 Hz, 1H, N5H), 7.13 (m, 2H, C(3’)H and C(5’)H), 7.35 (m, 2H,
C(2’)H and C(5’)H), 7.61 (d, J=8.4 Hz, 1H, C(7)H), 11.48 ppm (s, 1H,
N(3)H); 13C NMR (100 MHz, [D6]DMSO): d=14.0, 43.7, 62.6, 99.9,
110.6, 114.8, 123.3, 128.9, 136.4, 141.8, 149.6, 149.8, 155.6, 159.8,
162.2 ppm; IR: n˜ =3392, 3060, 2995, 2934, 2856, 2746, 1755, 1721,
1607, 1514 cmꢀ1; HRMS-ESI m/z [M+H]+ calcd for C16H15N4O3F:
331.1201, found: 331.1186; Anal. calcd for C16H15N4O3F: C 58.2, H
4.6, N 17.0, found: C 58.3, H 4.5, N 16.8.
Ethyl [2-amino-6-(4-fluorobenzylamino)pyridine-3-yl](methyl)car-
bamate (4): Following general procedure A starting from
1 (304 mg) gave a light-brown solid (127 mg, 40%); purity: 100%;
mp: 112–1138C (from CH2Cl2); 1H NMR (400 MHz, [D6]DMSO): d=
1.06 and 1.22 (brs and brs, 2H and 1H, O-CH2-CH3), 2.96 (s, 3H,
N3CH3) 3.97 (brs, 2H, O-CH2), 4.36 (m, 2H, C(a’)H2), 5.37 (s, 2H,
N2H2), 5.65 (d, J=8.4 Hz, 1H, C(5)H), 6.63 (t, J=6.0 Hz, 1H, N6H),
6.90 (d, J=8.4 Hz, 1H, C(4)H), 7.12 (t, J=9.0 Hz, 2H, C(3’)H and
C(5’)H), 7.37 ppm (m, 2H, C(2’)H and C(6’)H), 13C NMR (100 MHz,
[D6]DMSO): d=14.7, 36.5, 43.6, 60.5, 95.3, 111.4, 114.8, 129.2, 137.0,
137.1, 154.1, 155.7, 156.5, 159.8 and 162.2 ppm; IR: n˜ =3419, 3166,
1676, 1589, 1508, 1212 cmꢀ1; HRMS-ESI m/z [M+H]+ calcd for
C16H19N4O2F: 319.1565, found: 319.1550; Anal. calcd for
C16H19N4O2F: C 60.4, H 6.0, N 17.6, found: C 60.25; H 5.7, N 17.4.
Ethyl 5-[(ethoxycarbonyl)(4-fluorobenzyl)amino]-2-oxo-2,3-dihy-
dro-1H-imidazo[4,5-b]pyridine-1-carboxylate (8): Following gen-
eral procedure B yielded a light-brown solid (7 mg, 2%); purity
1
100%; mp: 157–1608C (from n-hexane/EtOAc); H NMR (400 MHz,
[D6]DMSO): d=1.15 (t, J=7.2 Hz, 3H, N5-COO-CH2-CH3), 1.35 (t, J=
7.2 Hz, 3H, N(1)-COO-CH2-CH3), 4.13 (q, J=7.2 Hz, 2H, N5-COO-CH2),
4.39 (q, J=7.2 Hz, 2H, N(1)-COO-CH2), 4.99 (s, 2H, C(a’)H2), 7.11 (m,
2H, C(3’)H and C(5’)H), 7.21 (d, J=8.4 Hz, 1H, C(6)H), 7.30 (m, 2H,
C(2’)H and C(6’)H), 7.88 (d, J=8.4 Hz, 1H, C(7)H), 11.98 ppm (s, 1H,
N(3)H); 13C NMR (100 MHz, [D6]DMSO): d=14.0, 14.3, 49.7, 61.7,
63.2, 113.4, 115.0, 118.3, 122.2, 129.3, 134.4, 142.1, 149.5, 149.9,
154.5, 160.0, 162.4 ppm; IR: n˜ =3219, 2972, 1763, 1731, 1708, 1613,
1509, 1217 cmꢀ1; HRMS-ESI m/z [M+H]+ calcd for C19H19N4O5F:
403.1412, found: 403.1414.
Ethyl
[2-amino-4-(4-fluorobenzylamino)phenyl](methyl)carba-
mate (5): Following general procedure A starting from 2 (303 mg)
gave a light-yellow solid (92 mg, 29%); purity: 100%; mp: 113–
115 8C (from CH2Cl2); 1H NMR (400 MHz, [D6]DMSO): d=1.04 and
1.21 (brs and brs, 2H and 1H, O-CH2-CH3), 2.95 (s, 3H, N1-CH3), 3.95
(brs, 2H, O-CH2), 4.15 (d, J=5.2 Hz, 2H, C(a’)H), 4.63 (s, 2H, N2H2),
5.79 (d, J=8.4 Hz, 1H, C(5)H), 5.90 (s, 1H, C(3)H), 5.99 (t, J=5.6 Hz,
1H, N4H), 6.57 (d, J=8.0 Hz, 1H, C(6)H), 7.13 (t, J=8.6 Hz, 2H,
C(3’)H and C(5’)H), 7.36 ppm (m, 2H, C(2’) and C(6’)H); 13C NMR
(100 MHz, [D6]DMSO): d=14.7, 36.8, 45.8, 60.4, 98.2, 101.7, 114.9,
N-Methyl-4-fluorbenzylamine (12): 4-Fluorobenzaldehyde (9;
10 mmol, 1056 mL) and methylamine (30 mmol, 3 mL, 40% solution
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 13
&
8
&
ÞÞ
These are not the final page numbers!

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
in MeOH) were diluted with MeOH (10 mL) and held at reflux over-
night. The solution was cooled to 08C, and NaBH4 (25 mmol, 1 g)
was slowly added. The mixture was allowed to warm to room tem-
perature and was stirred overnight. A small volume of H2O was
added to quench residual NaBH4, then the mixture was concentrat-
ed under reduced pressure, the residue diluted with CH2Cl2 (20 mL)
and washed with brine (2ꢂ20 mL). The organic layer was dried
over Na2SO4 and concentrated in vacuo to give the desired com-
yellow solution was reduced to dryness under a stream of air, then
the remaining crystals were washed twice with H2O and once with
a small amount of 2-propanol to give a yellow solid (2.58 g, 93%);
1
purity 99%; mp: 127–1288C (from 2-propanol); H NMR (400 MHz,
[D6]DMSO): d=2.71 (brs, 3H, CH3), 4.85 (brs, 2H, C(a’)H2), 6.23 (d,
1H, C(5)H), 7.17 (m, 2H, C(3’)H and C(5’)H), 7.33 (m, 2H, C(2’)H and
C(6’)H), 7.77 and 7.99 (brs and brs, 2H, NH2), 8.08 ppm (d, 1H,
C(4)H); 13C NMR (100 MHz, [D6]DMSO): d=35.9, 51.1, 98.7, 115.2,
118.0, 129.3, 133.6, 135.7, 154.4, 159.8, 160.1, 162.5 ppm; IR: n˜ =
3461, 3341, 1594, 1549, 1504, 1214; HRMS-ESI m/z [M+H]+ calcd
for C13H13N4O2F: 277.1095, found: 277.1082; Anal. calcd for
pound as
a
nearly colorless solution (1.2 g, 87%); 1H NMR
(400 MHz, [D6]DMSO): d=2.01 (brs, 1H, NH), 2.24 (s, 3H, CH3), 3.60
(s, 2H, CH2), 7.11 (m, 2H, C(3)H and C(5)H), 7.34 ppm (m, 2H, C(2)H
and C(6)H); 13C NMR (100 MHz, [D6]DMSO): d=35.4, 54.2, 114.8,
129.6, 136.9, 159.7, 162.1 ppm; IR: n˜ =2986, 2944, 2856, 1603,
C13H13N4O2F·1= H2O: C 55.6, H 4.85, N 19.95, found: C 55.4, H 4.75,
4
N 19.7.
1508, 1219, 821 cmꢀ1
.
N-Methyl-4-fluorbenzylidenamine (10): Intermediate of the syn-
thesis of 12 was isolated by taking a sample of the solution after
holding at reflux and concentrating it in vacuo, to give a colorless
liquid; 1H NMR (400 MHz, [D6]DMSO): d=3.43 (d, J=1.6 Hz, 3H,
CH3), 7.27 (m, 2H, C(3)H and C(5)H), 7.78 (m, 2H, C(2)H and C(6)H),
8.34 ppm (d, J=1.6 Hz, 1H, C(a)H); 13C NMR (100 MHz, [D6]DMSO):
d=47.4, 115.6, 129.8, 132.8, 160.8, 162.2, 164.6 ppm; IR: n˜ =2942,
General procedure C
Compound 16 (5 mmol, 1 g) was dissolved in MeCN (10 mL), Et3N
(14 mmol, 1.9 mL), and the appropriate 4-fluorobenzylamine (11 or
12; 6 mmol). The mixture was stirred overnight at room tempera-
ture, then more of the appropriate 4-fluorobenzylamine (6 mmol)
was added, and the solution was held at reflux for 24 h (if necessa-
ry even longer). After cooling, the reaction mixture and HCl solu-
tion (2.5 mL, 1m) were combined in a separation funnel and ex-
tracted with EtOAc (3ꢂ10 mL). The combined organic layers were
washed with brine (25 mL), dried over Na2SO4, and concentrated to
dryness under a stream of air.
2845, 2773, 1650, 1602, 1507, 1228, 832 cmꢀ1
.
2,6-Dichloro-3-nitropyridine (14): 2,6-Dichloropyridine (13;
30 mmol, 4.5 g), HNO3 (35 mL, 65%) and conc. H2SO4 (90 mL) were
held at reflux overnight at ~100–1058C under an atmosphere of
argon. The mixture was then cooled to 08C and poured into ice (~
600 mL). The white precipitate was filtered off, rinsed several times
with H2O until neutral and dried for several days in a desiccator
under reduced pressure. Colorless solid (3.3 g, 57%); mp: 56–598C
(from H2O); 1H NMR (400 MHz, [D6]DMSO): d=7.89 (d, J=8.4 Hz,
1H, C(5)H), 8.65 ppm (d, J=8.4 Hz, 1H, C(4)H); 13C NMR (100 MHz,):
d=125.0, 138.5, 141.5, 143.8, 151.9 ppm; IR: n˜ =3059, 1561, 1524,
2-Dimethylamino-6-(4-fluorobenzylamino)-3-nitropyridine (18):
Following general procedure C starting from 11 (2ꢂ680 mL) yielded
a
brown oil (1.17 g, 81%); purity 98%; 1H NMR (400 MHz,
[D6]DMSO): d=2.90 (s, 6H, N2(CH3)2), 4.52 (d, J=5.6 Hz, 2H,
C(a’)H2), 5.99 (d, J=9.2 Hz, 1H, C(5)H), 7.15 (m, 2H, C(3’)H and
C(5’)H), 7.36 (m, 2H, C(2’)H and C(6’)H), 7.97 (d, J=8.8 Hz, 1H,
C(4)H), 8.28 ppm (brs, 1H, N6H); 13C NMR (100 MHz, [D6]DMSO): d=
40.1, 43.4, 100.4, 114.9, 120.1, 129.2, 135.7, 136.6, 155.1, 157.8,
159.8, 162.3 ppm; IR: n˜ =3344, 3074, 2922, 2800, 1584), 1528, 1506,
1217 cmꢀ1; HRMS-ESI m/z [M+H]+ calcd for C14H15N4O2F: 291.1252,
found: 291.1255.
1340 cmꢀ1
.
2-Amino-6-chloro-3-nitropyridine (15): Compound 14 (20 mmol,
3.86 g) was dissolved in 2-propanol (200 mL) and aqueous NH3
(10 mL, 25%), warmed in a stoppered round-bottom flask to 358C
and stirred for 3 days, resulting in a yellow precipitate. The precipi-
tate was filtered off, washed twice with H2O (20 mL) and dried for
several days under reduced pressure. Yellow solid (1.77 g, 51%);
mp: 180–1858C (from 2-propanol); 1H NMR (400 MHz, [D6]DMSO):
d=6.77 (d, J=8.8 Hz, 1H, C(5)H), 8.26 (brs, 2H, NH2), 8.39 ppm (d,
J=8.8 Hz, 1H, C(4)H); 13C NMR (100 MHz, [D6]DMSO): d=112.0,
126.1, 138.4, 153.5, 155.0 ppm; IR: n˜ =3443, 3279, 3157, 3098,
2-Dimethylamino-6-[(4-fluorobenzyl)(methyl)amino]-3-nitropyri-
dine (19): Following general procedure C starting from 12 (2ꢂ
800 mL) to yield a brown oil (1.38 g, 91%); purity 99%; 1H NMR
(400 MHz, [D6]DMSO): d=2.92 (s, 6H, N2(CH3)2), 3.12 (s, 3H, N6CH3),
4.83 (s, 2H, C(a’)H2), 6.21 (d, J=9.2 Hz, 1H, C(5)H), 7.16 (m, 2H,
C(3’)H and C(5’)H), 7.29 (m, 2H, C(2’)H and C(6’)H), 8.08 ppm (d, J=
9.2 Hz, 1H, C(4)H); 13C NMR (100 MHz, [D6]DMSO): d=36.3, 40.2,
51.9, 97.3, 115.3, 120.9, 129.0, 134.0, 137.7, 154.1, 157.5, 160.1,
162.5 ppm; IR: n˜ =3060, 2927, 2800, 1571, 1503, 1218 cmꢀ1; HRMS-
ESI m/z [M+H]+ calcd for C15H17N4O2F: 305.1408, found: 305.1405.
1632, 1556, 1496, 761 cmꢀ1
.
6-Chloro-2-dimethylamino-3-nitropyridine (16): Compound 14
(5 mmol, 1 g) was dissolved in MeCN (20 mL) and Et3N (7.5 mmol,
1 mL) and cooled to 08C. Then dimethylamine (6 mmol, 760 mL,
40% aqueous solution) was slowly added, the mixture stirred for
10 min and then allowed to warm to room temperature. After
30 min the mixture was reduced to dryness under a stream of air
and the crude yellow crystalline product was purified by silica gel
chromatography (solvent: toluene) to give a yellow solid (510 mg,
Ethyl {2-amino-6-[(4-fluorobenzyl)(methyl)amino]pyridine-3-yl}-
carbamate (20): Compound 17 (1 mmol, 276 mg) and Pd/C
(25 mg, 10% Pd) were carefully set under a hydrogen atmosphere,
diluted with EtOH (10 mL), and stirred vigorously overnight. Then
ethyl chloroformate (1.25 mmol, 119 mL) and Et3N (1.5 mmol,
207 mL) were added and stirred for another 4 h. The suspension
was filtered through a glass filter (por. 3) with a 1 cm layer of silica
60 H to remove all catalyst, and the filtrate was concentrated in
vacuo. H2O (1 mL) and 2-propanol (1 mL) were added to the oily
residue and after storing overnight at 2–88C, a colorless precipitate
formed, which was collected by filtration and dried under reduced
pressure. Colorless solid (208 mg, 65%); purity 99%; mp: 114–
116 8C (from H2O/2-propanol); 1H NMR (400 MHz, [D6]DMSO): d=
1.21 (brs, 3H, O-CH2-CH3), 2.89 (s, 3H, N6-CH3), 4.05 (q, J=6.8 Hz,
2H, O-CH2), 4.69 (s, 2H, C(a’)H2), 5.33 (s, 2H, N2H2), 5.80 (d, J=
1
51%); mp: 70–728C (from toluene); H NMR (400 MHz, [D6]DMSO):
d=2.97 (s, 6H, N(CH3)2), 6.83 (d, J=8.4 Hz, 1H, C(5)H) 8.25 ppm (d,
J=8.4 Hz, 1H, C(4)H); 13C NMR (100 MHz, [D6]DMSO): d=39.9,
111.4, 130.0, 138.9, 150.8, 151.7 ppm; IR: n˜ =3114, 3082, 2935,
2805, 1593, 1546, 1509 cmꢀ1
.
2-Amino-6-[(4-fluorobenzyl)(methyl)amino]-3-nitropyridine (17):
Compound 15 (10 mmol, 1.74 g) was suspended in 2-propanol
(150 mL), 12 (15 mmol, 2 mL) and Et3N (20 mmol, 2.8 mL) were
added, and the mixture was held at reflux overnight. The resulting
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 13
&
9
&
ÞÞ
These are not the final page numbers!

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
8.4 Hz, 1H, C(5)H), 7.11 (m, 3H, C(3’)H, C(5’)H and C(4)H), 7.25 (t,
2H, C(2’)H and C(6’)H), 8.25 ppm (brs, 1H, N3H); 13C NMR (100 MHz,
[D6]DMSO): d=14.6, 35.7, 51.3, 60.0, 93.2, 107.2, 115.0, 129.1, 135.5,
135.7, 152.6, 155.1, 155.2, 159.9, 162.3 ppm; IR: n˜ =3459, 3367,
2981, 2904, 1712, 1604, 1500, 1205 cmꢀ1); HRMS-ESI m/z [M+H]+
calcd for C16H19N4O2F: 319.1565, found: 319.1551; Anal. calcd for
C16H19N4O2F: C 60.4, H 6.0, N 17.6, found: C 60.5, H 6.0, N 17.4.
13C NMR (100 MHz, [D6]DMSO): d=14.7, 40.3, 43.7, 59.8, 97.9,
109.8, 114.7, 128.9, 137.4, 138.8, 154.9, 155.3, 155.7, 159.6,
162.1 ppm; IR: n˜ =3391, 2993, 2938, 2842, 1707, 1603, 1587, 1215,
821 cmꢀ1; HRMS-ESI m/z [M+H]+ calcd for C17H21N4O2F: 333.1721,
found: 333.1718.
Ethyl {2-amino-6-[(4-fluorobenzyl)(methyl)amino]pyridine-3-yl}-
(methyl)carbamate (23): Compound 20 (0.5 mmol, 159 mg) and
NaH (0.75 mmol, 30 mg, 65% solution in n-hexane) in THF (8 mL)
were stirred under an atmosphere of argon at 08C in the dark.
After 30 min CH3I (1 mmol, 62 mL) was added and stirred overnight
while allowing the reaction to reach room temperature. The reac-
tion mixture was concentrated under reduced pressure, and the
residue was diluted with HCl solution (10 mL, 1m) and washed
with Et2O (10 mL). The aqueous layer was then made alkaline with
NaOH solution (10 mL) and extracted with Et2O (2ꢂ10 mL). The
combined organic layers were concentrated in vacuo, and the re-
maining colorless oil was stored at 2–88C until a precipitate
formed (2–3 days). Colorless solid (40 mg, 24%); purity: 100%; mp:
73–768C (from Et2O); 1H NMR (400 MHz, [D6]DMSO): d=1.07 and
1.23 (brs and brs, 2H and 1H, O-CH2-CH3), 2.91 (s, 3H, N6-CH3), 2.99
(s, 3H, N3-CH3), 3.98 (brs, 2H, O-CH2), 4.71 (m, 2H, C(a’)H2), 5.50 (s,
2H, N2H2), 5.76 (d, J=8.4 Hz, 1H, C(5)H), 7.02 (d, J=8.4 Hz, 1H,
C(4)H), 7.12 (m, 2H, C(3’)H and C(5’)H) 7.28 ppm (m, 2H, C(2’)H and
C(6’)H); 13C NMR (100 MHz, [D6]DMSO): d=14.6, 35.7, 36.5, 51.2,
60.5, 93.2, 111.8, 114.9, 129.0, 135.4, 137.2, 153.7, 155.5, 156.1,
159.8, 162.2 ppm; IR: n˜ =3443, 3346, 2983, 2847, 1675), 1603, 1217,
773 cmꢀ1; HRMS-ESI m/z [M+H]+ calcd for C17H21N4O2F: 333.1721,
found: 333.1708; Anal. calcd for C17H21N4O2F: C 61.4, H 6.4, N 16.6,
found: C 61.5, H 6.1, N 16.6.
General procedure D
The appropriate nitropyridine derivative (2 mmol 18 or 19) and Pd/
C (150 mg, 10% Pd) were diluted with EtOH (15 mL) and stirred
vigorously overnight under an atmosphere of hydrogen. Then
ethyl chloroformate (2.5 mmol, 238 mL) and Et3N (3 mmol, 414 mL)
were added and stirred for another 8 h. The suspension was fil-
tered through a glass filter (por. 3) with a 1 cm layer of silica 60 H
to remove the catalyst, and the filtrate was concentrated in vacuo.
The residue was diluted with CH2Cl2 (10 mL) and HCl solution
(2 mL, 1m) and in a separation funnel washed with brine (2ꢂ
5 mL). The combined organic layers were again concentrated in
vacuo, and the residue was diluted with EtOAc (8 mL). Insoluble
residues were removed by decanting the solution. n-Hexane (2 mL)
was added to the solution, resulting in the precipitation of crystals,
which were collected by filtration.
Ethyl [2-(dimethylamino)-6-(4-fluorobenzylamino)pyridine-3-yl]-
carbamate hydrochloride (21·HCl): Following general procedure D
starting from 18 (580 mg) yielded a red–brown solid (75 mg, 10%);
purity 95%; mp: 155–1588C (from n-hexane/EtOAc); 1H NMR
(400 MHz, [D6]DMSO): d=1.20 (brs, 3H, O-CH2-CH3), 2.82 and 3.01
(m and brs, 1H and 5H, N2(CH3)2), 4.07 (q, J=6.8 Hz, 2H, O-CH2),
4.46 (s, 2H, C(a’)H2), 6.24 (brs, 1H, C(5)H), 7.12 (m, 3H, C(3’)H,
C(5’)H and N6H), 7.40 (m, 3H, C(2’)H, C(6’)H and C(4)H), 8.82 (brs,
1H, N3H), 12.6 ppm (brs, 1H, HCl); 13C NMR (100 MHz, CDCl3): d=
14.7, 42.4, 45.9, 62.2, 116.0, 128.8, 132.2, 152.3, 155.1, 161.3,
163.7 ppm; IR: n˜ =3191, 3157, 3056, 2987, 2916, 2359, 1718, 1608,
1508, 1219 cmꢀ1; HRMS-ESI m/z [M+H]+ calcd for C17H21N4O2F:
2-(4-Fluorobenzylamino)-5-nitropyridine (25): 2-Chloro-5-nitropyr-
idine (24; 5 mmol, 0.8 g) was diluted with MeCN (10 mL), Et3N
(15 mmol, 2.1 mL), and 11 (6 mmol, 680 mL), and the mixture was
stirred overnight at room temperature. More of compound 11
(6 mmol) was then added, and the solution was held at reflux for
24 h. The reaction mixture was cooled to initiate crystallization, the
suspension was diluted with HCl (12 mL, 1m), and the precipitating
light-brown crystals were filtered off, washed with H2O and dried
under reduced pressure. Light-yellow–brown solid (1.09 g, 88%);
purity 97%; mp: 167–1688C (from MeCN); 1H NMR (400 MHz,
[D6]DMSO): d=4.61 (s, 2H, C(a’)H2), 6.63 (m, 1H, C(3)H), 7.17 (m,
2H, C(3’)H and C(5’)H), 7.37 (m, 2H, C(2’)H and C(6’)H), 8.14 (dd, J=
9.2 Hz, J=2.4 Hz, 1H, C(4)H), 8.59 (brs, 1H, NH) 8.92 ppm (d, J=
2.8 Hz, 1H, C(6)H); 13C NMR (100 MHz, [D6]DMSO): d=43.4, 108.6,
115.1, 129.3, 131.8, 134.5, 135.0, 146.7, 160.0, 161.1, 162.4 ppm; IR:
333.1721, found: 333.1721; Anal. calcd for C17H22N4O2FCl·1= EtOAc:
4
C 55.3, H 6.2, N 14.3, found: C 55.2, H 5.8, N 14.4.
Ethyl {2-(dimethylamino)-6-[(4-fluorobenzyl)(methyl)amino]pyri-
dine-3-yl}carbamate hydrochloride (22·HCl): Following general
procedure D starting from 19 (608 mg) yielded a light colorless
solid (200 mg, 26%); purity 92%; mp: 142–1478C (from n-hexane/
EtOAc); 1H NMR (400 MHz): d=1.22 (brs, 3H, O-CH2-CH3), 2.96 (s,
6H, N2(CH3)2), 3.07 (s, 3H, N6CH3), 4.08 (q, J=7.2 Hz, 2H, O-CH2);
4.76 (s, 2H, C(a’)H2), 6.61 (brs, 1H, C(5)H), 7.13 (m, 2H, C(3’)H and
C(5’)H), 7.29 (m, 2H, C(2’)H and C(6’)H), 7.48 (d, J=5.6 Hz, 1H,
C(4)H), 8.94 ppm (s, 1H, N3H); 13C NMR (100 MHz, [D6]DMSO): d=
14.5, 36.4, 42.9, 52.0, 60.7, 112.9, 115.2, 129.0, 134.7, 139.2, 147.0,
155.0 160.0, 162.4 ppm; IR: n˜ =3155, 3049, 2995, 2870, 1709, 1211,
811 cmꢀ1; HRMS-ESI m/z [M+H]+ calcd for [M+H]+ C18H24N4O2F:
347.1878, found: 347.1877; Anal. calcd for C18H25N4O2FCl: C 56.5, H
6.4, N 14.6, found: C 56.3, H 6.1, N 14.3.
n˜ =3220 (NH), 3097, 3048, 2991, 2935, 2881, 1600, 1498 cmꢀ1
;
HRMS-ESI m/z [M+H]+ calcd for C12H10N3O2F: 248.0830, found:
248.0837; Anal. calcd for C12H10N3O2F: C 58.3, H 4.1, N 17.0, found:
C 57.95, H 4.0, N 16.8.
2-[(4-Fluorobenzyl)(methyl)amino]-5-nitropyridine (26): 2-Chloro-
5-nitropyridine (24; 5 mmol, 0.8 g) was diluted with MeCN (10 mL),
Et3N (15 mmol, 2.1 mL), and 12 (6 mmol, 800 mL), and the mixture
was stirred overnight at room temperature. More of 12 (6 mmol)
was then added, and the solution was held at reflux for 24 h. The
reaction mixture was cooled and diluted with HCl (12 mL, 1m) and
extracted with EtOAc (3ꢂ10 mL). The combined organic layers
were washed with brine (25 mL), dried over Na2SO4 and concen-
trated in vacuo, slowly leading crystallization of the residue. Light-
yellow solid (1.16 g, 89%); purity 99%; mp: 76–788C (from EtOAc);
1H NMR (400 MHz, [D6]DMSO): d=3.18 (s, 3H, CH3), 4.93 (s, 2H,
C(a’)H2), 6.80 (d, J=9.6 Hz, 1H, C(3)H), 7.17 (m, 2H, C(3’)H and
C(5’)H), 7.29 (m, 2H, C(2’)H and C(6’)H), 8.24 (dd, J=9.6 Hz, J=
2.8 Hz, 1H, C(4)H), 8.99 ppm (d, J=2.8 Hz, 1H, C(6)H); 13C NMR
Ethyl [2-(dimethylamino)-6-(4-fluorobenzylamino)pyridine-3-yl]-
carbamate (21): Compound 21·HCl (0.05 mmol, 20 mg) was added
to a NaOH solution (2 mL, 2m) and CH2Cl2 (2 mL) and vigorously
extracted. The organic layer was separated, dried over Na2SO4, and
concentrated in vacuo. Red–brown slurry (16 mg, 89%); purity
1
89%; H NMR (400 MHz, [D6]DMSO): d=1.17 (brs, 3H, O-CH2-CH3),
2.77 (s, 6H, N2(CH3)2), 4.02 (q, J=7.2 Hz, 2H, O-CH2), 4.39 (d, J=
6.0 Hz, 2H, C(a’)H2), 5.88 (d, J=8.0 Hz, 1H, C(5)H), 6.82 (t, J=
6.0 Hz, 1H, N6H), 7.01 (d, J=7.6 Hz, 1H, C(4)H), 7.11 (m, 2H, C(3’)H
and C(5’)H), 7.35 (m, 2H, C(2’)H and C(6’)H), 8.22 ppm (s, 1H, N3H);
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 13
&10
&
ÞÞ
These are not the final page numbers!

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
(100 MHz, [D6]DMSO): d=36.6, 52.0, 105.2, 115.3, 129.1, 132.8,
133.4, 134.4, 146.1, 160.2, 160.4, 162.6 ppm; IR: n˜ =3088, 2934,
ous suspension of aluminum oxide powder (particle size: 0.3 mm)
for several minutes followed by sonication in deionized water.
1519, 1219, 809 cmꢀ1
;
HRMS-ESI m/z [M+H]+ calcd for
C13H12N3O2F: 262.0986, found: 262.0980; Anal. calcd for
C13H12N3O2F: C 59.8, H 4.6, N 16.1, found: C 59.9, H 4.5, N 16.0.
KV7.2/3-channel opening assay
KCNQ2/3-expressing HEK293 cells (cultivated with DMEM/F12, 10%
FCS, 100 UmLꢀ1 penicillin G sodium, 100 mgmLꢀ1 streptomycin sul-
fate) were plated in 384-well plates (black wall, flat, clear-bottom
microtiter plates, BD Biocoat poly-d-lysine multiwell cell culture
plate) incubated overnight (378C) to achieve near confluent mono-
layer (10000–30000 cells per well). Treatment with loading buffer,
assay buffer, and stimulus buffer was carried out according to the
FluxOR Potassium Ion Channel Assay protocol (Life Technolo-
gies).[35] Cells were incubated with test compounds dissolved in
assay buffer (concentrations: 0.03, 0.1, 0.3, 1, 3, 10, 30, and 100 mm)
for 30 min at room temperature prior to stimulus buffer addition.
Measurement of relative fluorescence intensity (stimulus buffer
alone: 0%, 100 mm flupirtine maleate: 100%, excitation/emission
wavelengths: 460–490/520–540 nm) was done by using a fluores-
cence imaging plate reader (FLIPRTETRA).
Ethyl [2-(4-fluorobenzylamino)pyridine-5-yl]carbamate hydro-
chloride (27·HCl): Compound 25 (2 mmol, 494 mg) and Pd/C
(150 mg, 10% Pd) were diluted with EtOH (15 mL) and stirred vigo-
rously overnight under a hydrogen atmosphere. Then ethyl chloro-
formate (2.5 mmol, 238 mL) and Et3N (3 mmol, 414 mL) were added
to the reaction and stirred for another 8 h. The suspension was fil-
tered through a glass filter (por. 3) with a 1 cm layer of silica 60 H
to remove the catalyst, and the filtrate was concentrated in vacuo.
The residue was diluted with CH2Cl2 (10 mL) and HCl (2 mL, 1m),
and the precipitated crystals were collected by filtration, washed
with a small volume of H2O and dried under reduced pressure. Col-
orless solid (215 mg, 33%); purity 100%; mp: 187–1898C (from
CH2Cl2); 1H NMR (400 MHz, [D6]DMSO): d=1.24 (t, J=7.2 Hz, 3H,
CH3), 4.13 (q, J=7.2 Hz, 2H, O-CH2), 4.60 (brs, 2H, C(a’)H2), 7.10 (d,
J=9.6 Hz, 1H, C(3)H), 7.21 (m, 2H, C(3’)H and C(5’)H), 7.47 (m, 2H,
C(2’)H and C(6’)H), 7.86 (dd, J=2.4 Hz, J=9.2 Hz, 1H, C(4)H), 8.10
(brs, 1H, C(6)H), 8.95 (brs, 1H, N5H), 9.87 (s, 1H, N2H), 13.96 ppm
(brs, 1H, HCl); 13C NMR (100 MHz, [D6]DMSO): d=14.3, 44.4, 60.7,
113.7, 115.2, 123.6, 126.3, 129.8, 132.9, 136.0, 149.2, 153.5, 160.3,
162.7 ppm; IR: n˜ =3305, 3167, 3091, 3046, 2989, 2920, 2739, 1692,
1216, 822 cmꢀ1; HRMS-ESI m/z [M+H]+ calcd for C15H16N3O2F:
Cytotoxicity assessment in TAMH cells
TGF-a transgenic mouse hepatocytes (TAMH) were obtained from
Sidney D. Nelson (Department of Medicinal Chemistry, University of
Washington, Seattle, USA). Cultivation conditions for this cell line
and the MTT assay were described previously.[43] Briefly, cells be-
tween passages 31 and 40 at ~90% confluence were seeded into
96-well microtiter plates (Sarstedt) in 1:1 mixture of DMEM/F-12
culture medium containing ITS (5 mgmLꢀ1 insulin, 5 mgmLꢀ1 trans-
ferrin, 5 ng selenium), 0.1 mm dexamethasone, 10 mm nicotinamide,
and 50 mgmLꢀ1 gentamicin. After cells became confluent (usually
3–5 days following seeding), the medium was aspirated off and re-
placed with 200 mL of the same medium containing test com-
pound and 1% DMSO. Typically, 4–8 serial dilutions of test com-
pound were used for determination of the LD50 value. Untreated
control rows were prepared with medium containing 1% DMSO
alone. After 24 h exposure to drug, medium was aspirated off and
replaced with 100 mL medium, as well as 20 mL per well of
a 2.5 mgmLꢀ1 solution of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-
tetrazolium bromide (MTT) in phosphate-buffered saline (PBS).
Plates were returned to the incubator for 4 h. To each well was
then added 100 mL 0.04n HCl in 2-propanol. Plates were sonicated
until all crystals had dissolved before the optical densities were
measured on a plate reader (Sunrise, Tecan) at l=570 nm. LD50
values were calculated by linear regression of the T/C (%) versus
log(concentration) curves.
290.1299, found: 290.1299; Anal. calcd for C15H17N3O2FCl·1= H2O: C
4
54.55, H 5.3, N 12.7, found: C 54.5, H 5.0, N 12.7.
Ethyl {2-[(4-fluorobenzyl)(methyl)amino]pyridine-5-yl}carbamate
hydrochloride (28·HCl): Following general procedure D starting
from 523 mg 26 yielded a colorless solid (155 mg, 23%); purity:
98%; mp: 180–1858C (from n-hexane/EtOAc); 1H NMR (400 MHz,
[D6]DMSO): d=1.24 (t, J=7.2 Hz, 3H, CH3), 3.24 (s, 3H, N2CH3), 4.14
(q, J=7.2 Hz, 2H, O-CH2), 4.90 (s, 2H, C(a’)H2), 7.19 (m, 2H, C(3’)H
and C(5’)H), 7.29 (d, J=9.6 Hz, 2H, C(3)H), 7.36 (m, 2H, C(2’)H and
C(6’)H), 7.97 (d, J=9.2 Hz, 1H, C(4)H), 8.27 (s, 1H, C(6)H) and
9.97 ppm (s, 1H, N5H); 13C NMR (100 MHz, [D6]DMSO): d=14.4,
38.0, 53.4, 60.8, 112.2, 115.6, 125.4, 126.4, 129.2, 132.1, 135.8, 149.2,
153.6, 160.4, 162.8 ppm; IR: n˜ =2984, 2776, 1716, 812 cmꢀ1; HRMS-
ESI m/z [M+H]+ calcd for C16H18N3O2F: 304.1456, found: 304.1442;
Anal. calcd for C16H19N3O2FCl·1= H2O: C 55.1, H 5.8, N 12.05, found:
2
C 54.9, H 5.55, N 11.9.
N-(4-Hydroxyphenyl)-N-methylacetamide (29): This compound
was synthesized as described by Fernando et al.[45] to yield a color-
less solid (1.68 g; 70%); mp: 243–2448C (from CH2Cl2/MeOH);
1H NMR (400 MHz, [D6]DMSO): d=1.71 (s, 3H, CO-CH3), 3.07 (s, 3H,
N-CH3), 6.79 (d, J=8.4 Hz, 2H, C(3’)H and C(5’)H), 7.10 (d, J=8.4 Hz,
2H, C(2’)H and C(6’)H), 9.62 ppm (s, 1H, OH); 13C NMR (100 MHz,
[D6]DMSO): d=21.9, 36.6, 115.8, 128.1, 135.7, 156.4, 169.3 ppm; IR:
Acknowledgements
n˜ =3098, 3016, 2907, 2804, 1622, 1587, 1239, 837 cmꢀ1
.
FLIPR assays were carried out by ChanTest Corporation (Cleve-
land, OH, USA). A scholarship to C.J.L. from the Dr. Hilmer Stif-
tung is gratefully acknowledged. P.J.B. thanks Dr. Kevin J. Coe
and Dr. Peter Rademacher, University of Washington, USA, for
helpful instructions in the use of TAMH cells.
Cyclic voltammetry studies
The voltammeter consisted of a glassy carbon working electrode,
a platinum auxiliary electrode, and an Ag/AgCl reference electrode.
Solutions of compound were prepared in 100 mm Tris buffer
(pH 7.4) at a concentration of 1.0 mm, and were purged with N2
for 3 min prior to analysis. Cycles were swept through ꢀ1.0 to
1.0 V (displaying only ꢀ0.5 to 1.0 V) at a sweep rate of 0.5 Vsꢀ1. Be-
tween measurements, the working electrode required hand polish-
ing, which was done by using a polishing cloth soaked in an aque-
Keywords: alkylation · cyclic voltammetry · flupirtine · KCNQ ·
KV7
[1] A. Douros, E. Bronder, F. Andersohn, A. Klimpel, M. Thomae, H.-D. Orze-
chowski, R. Kreutz, E. Garbe, Eur. J. Clin. Pharmacol. 2014, 70, 453–459.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 13
&
11
&
ÞÞ
These are not the final page numbers!

CHEMMEDCHEM
FULL PAPERS
www.chemmedchem.org
[2] I. Szelenyi, Inflammation Res. 2013, 62, 251–258.
[26] G. Jas, I. Freifeld, (Corden PharmaChem GmbH, Mannheim, Germany),
Ger. Pat. No. DE102009023162A1, 2011.
[27] J. R. Merritt, J. Liu, E. Quadros, M. L. Morris, R. Liu, R. Zhang, B. Jacob, J.
Postelnek, C. M. Hicks, W. Chen, E. F. Kimble, W. L. Rogers, L. O’Brien, N.
White, H. Desai, S. Bansal, G. King, M. J. Ohlmeyer, K. C. Appell, M. L.
Webb, J. Med. Chem. 2009, 52, 1295–1301.
[28] W. K. Anderson, A. S. Milowsky, J. Med. Chem. 1986, 29, 2241–2249.
[29] W. D. Brown, C. Jessen, D. Stroebaek, (Neurosearch A/S, Ballerup, Den-
mark), Int. PCT Pub. No. WO2009074591A1, 2009.
[3] G. Amato, R. Roeloffs, G. C. Rigdon, B. Antonio, T. Mersch, G. McNaugh-
ton-Smith, A. D. Wickenden, P. Fritch, M. J. Suto, ACS Med. Chem. Lett.
2011, 2, 481–484.
[4] Q. Li, S. Wanderling, M. Paduch, D. Medovoy, R. E. Hulse, A. Singharoy,
R. McGreevy, C. A. Villalba-Galea, B. Roux, A. Kossiakoff, E. Perozo, K.
Schulten, Nat. Struct. Mol. Biol. 2014, 21, 244–252.
[5] K. Padilla, A. D. Wickenden, A. C. Gerlach, K. McCormack, Neurosci. Lett.
2009, 465, 138–142.
[6] H. Yu, M. Wu, S. D. Townsend, B. Zou, S. Long, J. S. Daniels, O. B. McMa-
nus, M. Li, C. W. Lindsley, C. R. Hopkins, ACS Chem. Neurosci. 2011, 2,
572–577.
[7] G. Munro, W. Dalby-Brown, J. Med. Chem. 2007, 50, 2576–2582.
[8] H. Wulff, N. A. Castle, L. A. Pardo, Nat. Rev. Drug Discovery 2009, 8, 982–
1001.
[9] S. K. Bagal, A. D. Brown, P. J. Cox, K. Omoto, R. M. Owen, D. C. Pryde, B.
Sidders, S. E. Skerratt, E. B. Stevens, R. I. Storer, N. A. Swain, J. Med.
Chem. 2013, 56, 593–624.
[10] F. Puls, C. Agne, F. Klein, M. Koch, K. Rifai, M. P. Manns, J. Borlak, H. H.
Kreipe, Virchows Arch. 2011, 458, 709–716.
[30] C. Jessen, W. D. Brown, D. Stroebaek, (Neurosearch A/S, Ballerup, Den-
mark), Int. PCT Pub. No. WO2009074594A1, 2009.
[31] C. Jessen, W. D. Brown, D. Stroebaek, (Neurosearch A/S, Ballerup, Den-
mark), Int. PCT Pub. No. WO2009074592A1, 2009.
[32] W. D. Brown, C. Jessen, D. Stroebaek, L. Borup, (Neurosearch A/S, Baller-
up, Denmark), Int. PCT Pub. No. WO2009074593A1, 2009.
[33] K. G. Madsen, J. Olsen, C. Skonberg, S. H. Hansen, U. Jurva, Chem. Res.
Toxicol. 2007, 20, 821–831.
[34] K. G. Madsen, C. Skonberg, U. Jurva, C. Cornett, S. H. Hansen, T. N. Jo-
hansen, J. Olsen, Chem. Res. Toxicol. 2008, 21, 1107–1119.
[35] Life Technologies, tools.lifetechnologies.com/content/sfs/manuals/fluxor
potassium ion channel assay man.pdf (accessed August 13, 2013).
[36] L. Tatulian, P. Delmas, F. C. Abogadie, D. A. Brown, J. Neurosci. 2001, 21,
5535–5545.
[37] J. C. Wu, G. Merlino, K. Cveklova, B. Mosinger, Jr., N. Fausto, Cancer Res.
1994, 54, 5964–5973.
[38] J. K. Seydel, K.-J. Schaper, E. A. Coats, H. P. Cordes, P. Emig, J. Engel, B.
Kutscher, E. E. Polymeropoulos, J. Med. Chem. 1994, 37, 3016–3022.
[39] P. J. Harvison, A. J. Forte, S. D. Nelson, J. Med. Chem. 1986, 29, 1737–
1743.
[11] N. Anderson, J. Borlak, PLoS One 2011, 6, 1–11.
[12] German Medical Association, www.akdae.de/Arzneimittelsicherheit/
DSM/Archiv/2013-13.html (accessed February 28, 2014).
[13] K. Methling, P. Reszka, M. Lalk, O. Vrana, E. Scheuch, W. Siegmund, B.
Terhaag, P. J. Bednarski, Drug Metab. Dispos. 2009, 37, 479–493.
[14] K. Obermeier, G. Niebch, K. Thiemer, Arzneim.-Forsch. 1985, 35, 60–67.
[15] P. Hlavica, G. Niebch, Arzneim.-Forsch. 1985, 35, 67–74.
[16] D. A. Andersson, C. Gentry, L. Alenmyr, D. Killander, S. E. Lewis, A. Ander-
sson, B. Bucher, J.-L. Galzi, O. Sterner, S. Bevan, E. D. Hogestatt, P. M.
Zygmunt, Nat. Commun. 2011, 2, 1–11.
[17] J. G. M. Bessems, N. P. E. Vermeulen, Crit. Rev. Toxicol. 2001, 31, 55–138.
[18] L. P. James, P. R. Mayeux, J. A. Hinson, Drug Metab. Dispos. 2003, 31,
1499–1506.
[19] C. Springer, I. Niculescu-Duvaz, R. Marais, D. Niculescu-Duvaz, A.
Zambon, D. Menard, (Cancer Research Technology Ltd., London, UK),
Int. PCT Pub. No. WO2011092469A1, 2011.
[40] N. Sahoo, T. Hoshi, S. H. Heinemann, Antioxid. Redox Signaling 2014, 21,
933–952.
[41] N. Gamper, O. Zaika, Y. Li, P. Martin, C. C. Hernandez, M. R. Perez, A. Y. C.
Wang, D. B. Jaffe, M. S. Shapiro, EMBO J. 2006, 25, 4996–5004.
[42] R. H. Pierce, C. C. Franklin, J. S. Campbell, R. P. Tonge, W. Chen, N.
Fausto, S. D. Nelson, S. A. Bruschi, Biochem. Pharmacol. 2002, 64, 413–
424.
[20] J. W. Clark-Lewis, M. J. Thompson, J. Chem. Soc. 1957, 442–446.
[21] W. von Bebenburg, (Degussa AG, Frankfurt, Germany), Ger. Pat. No.
DE2241575A1, 1973.
[43] K. J. Coe, Y. Jia, H. K. Ho, P. Rademacher, T. K. Bammler, R. P. Beyer, F. M.
Farin, L. Woodke, S. R. Plymate, N. Fausto, S. D. Nelson, Chem. Res. Toxi-
col. 2007, 20, 1277–1290.
[22] D. Zhang, X. Song, J. Su, J. Pharm. Biomed. Anal. 2014, 90, 27–34.
[23] D. S. Mortensen, M. M. D. Mederos, J. J. Sapienza, R. J. Albers, B. G. Lee,
R. L. Harris III, G. I. Shevlin, D. Huang, K. L. Schwarz, G. K. Packard, J. S.
Parnes, P. W. Papa, L. R. Tehrani, S. Perrin-Ninkovic, (Signal Pharmaceuti-
cals, LLC, San Diego, USA), Int. PCT Pub. No. WO2008051493A2, 2008.
[24] W. von Bebenburg, (Degussa AG, Frankfurt, Germany), Ger. Pat. No.
DE3133519A1, 1982.
[44] Y. L. Teo, M. Saetaew, S. Chanthawong, Y. S. Yap, E. C. Y. Chan, H. K. Ho,
A. Chan, Breast Cancer Res. Treat. 2012, 133, 703–711.
[45] C. R. Fernando, I. C. Calder, K. N. Ham, J. Med. Chem. 1980, 23, 1153–
1158.
Received: October 14, 2014
[25] Y. M. Choi, N. Kucharczyk, R. Duane, J. Labelled Compd. Radiopharm.
1987, 24, 1–14.
Published online on && &&, 0000
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 13
&12
&
ÞÞ
These are not the final page numbers!

FULL PAPERS
Oxidation–activation–toxicity? The
connection between activity, toxicity,
and oxidation potential of the pain killer
flupirtine was investigated. N-Modified
analogues of flupirtine displayed vary-
ing oxidation potentials that appeared
to be linked to potassium channel
opening activity but not to in vitro hep-
atotoxicity. These data suggest that oxi-
dative metabolites of flupirtine contrib-
ute to the mechanism of action, but not
to liver toxicity.
C. J. Lemmerhirt, M. Rombach, A. Bodtke,
P. J. Bednarski, A. Link*
&& – &&
Oxidation Potentials of N-Modified
Derivatives of the Analgesic Flupirtine
Linked to Potassium KV7 Channel
Opening Activity But Not Hepatocyte
Toxicity
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 13
&13
&
ÞÞ
These are not the final page numbers!