Benzyl prolinate derivatives as novel selective KCC2 blockers

Article abstract of DOI:10.1016/j.bmcl.2010.02.092

The discovery and optimization of a novel class of selective submicromolar KCC2 blockers is described. Details of synthesis and SAR are given together with ADME properties of selected compounds. A methylsulfone residue on the R¹ phenyl group im

Full text of DOI:10.1016/j.bmcl.2010.02.092

Bioorganic & Medicinal Chemistry Letters 20 (2010) 2542–2545
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Benzyl prolinate derivatives as novel selective KCC2 blockers
a
a
a
a
Cécile Pégurier a, , Nathalie Bosman , Philippe Collart , Marie-Laure Delporte , Karine Leclercq ,
Sébastien Lengelé a, Ananda Kumar Kanduluru b, Stéphane Meunier b, Nathalie Pacico a,
Lakshmana Rao Vadali b, Alain Wagner b, Christian Wolff a, Laurent Provins a
*
a UCB Pharma SA, UCB NewMedicines, Chemin du Foriest, B-1420 Braine-L’Alleud, Belgium
b Laboratoire des systèmes chimiques fonctionnels, UMR 7199 associé au CNRS, Faculté de Pharmacie, Université de Strasbourg, 74 route du Rhin, BP 24, F-67401 Illkirch, France
a r t i c l e i n f o
a b s t r a c t
Article history:
The discovery and optimization of a novel class of selective submicromolar KCC2 blockers is described.
Details of synthesis and SAR are given together with ADME properties of selected compounds. A methyl-
sulfone residue on the R1 phenyl group improved the overall general profile of these prolinate derivatives.
Ó 2010 Elsevier Ltd. All rights reserved.
Received 21 January 2010
Revised 23 February 2010
Accepted 24 February 2010
Available online 1 March 2010
In memory of Dr. Charles Mioskowski
Keywords:
KCC2
Epilepsy
Furosemide
KCC2 is a neuronal-specific electroneutral potassium-chloride
co-transporter, whose function is to regulate intracellular levels
of chloride ions in neurons. It has been shown that, under normal
conditions, KCC2 transports ClÀ ions out of neurons, keeping ClÀ
concentrations below the thermodynamic equilibrium potential,
so that GABAA channels, when opened, will allow ClÀ to enter neu-
rons, hyperpolarizing them and inhibiting firing.1,2 KCC2 therefore
acts as a modulator of inhibitory neurotransmission, both in the
brain and in the spinal cord, hence its growing interest as a new
target for neuronal hyperexcitability disorders such as epilepsy
or neuropathic pain.3
KCC2 belongs to a large family of cation-chloride co-transporters
consisting of seven well-characterized members (NCC, NKCC1-2,
and KCC1-4) for which the cation (sodium or potassium) transport
is always accompanied by the concomitant stoichiometric trans-
port of a chloride ion.4
Anticonvulsant properties of loop diuretics such as furosemide
1 were reported in the clinic.5 It is a known blocker of cation-chlo-
ride co-transporters but does not display satisfactory selectivity
levels to render it attractive as pharmacological tool to definitively
unveil the role of KCC2.
The first KCC2 blockers displaying higher selectivity versus
NKCC1 have been recently reported6 but without detailed SAR
and without assessment of the overall druggability profile of the
identified compounds. In this Letter, we report the identification
of novel potent selective KCC2 blockers displaying suitable proper-
ties for their in vivo pharmacological evaluation.
High throughput screening (HTS) of our corporate compound
collection using a Rb flux assay7 on the KCC2 co-transporter led
to the identification of benzyl 1-acetyl-2-benzylprolinate (+)-2 as
a hit with an IC50 of 0.32
found to be inactive in a Rb flux assay8 on the NKCC1 co-trans-
porter (14% inhibition at 100 M). This selectivity clearly differen-
lM. The compound was subsequently
l
tiates compound (+)-2 from the loop diuretics like furosemide.
Compound (+)-2 is the dextro enantiomer and, interestingly, the
levo enantiomer (À)-2 is about 100-fold less potent on KCC2
(IC50 = 50
lM). Unfortunately, the poor metabolic stability (Clint
rat microsomes = 392
lL/min/mg protein) of compound (+)-2 ren-
der its use as an in vivo pharmacological tool problematic. It ap-
peared nevertheless to be a good starting point to identify a
compound with an improved profile.
In this Letter, we wish to report structure–activity relation-
ships around our hit compound focusing on the modulation of
R1, R2, and R3 (Fig. 1) substituents. Compounds were obtained fol-
lowing the synthetic route depicted in Scheme 1. N-Boc L-proline
was reacted with benzyl bromide in the presence of potassium
carbonate to form the corresponding benzyl ester. The alkylation
step9 was performed by deprotonation with LiHMDS at À70 °C,
followed by addition of various benzyl or alkyl bromides (condi-
tions c). Racemization occurred under these conditions and all de-
scribed compounds are racemic mixtures unless otherwise stated.
* Corresponding author. Fax: +32 2386 2704.
E-mail address: cecile.pegurier@ucb.com (C. Pégurier).
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.02.092

C. Pégurier et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2542–2545
2543
R1
COOH
N
O
O
O
N
SO2NH2
N
R3
O
O
Cl
O
O
1
(+/-)-2
R2
Figure 1. Furosemide 1 reference compound, hit compound 2, and SAR modulations.
R1
a, b
O
OH
OH
O
N
N
N
O
R2
O
O
O
2-21
O
j
23-29
d, e
R1
g
R1=Ph
d
e, f
OR3
a, c
OH
N
O
O
N
O
N
N
O
N
H
O
O
O
O
O
O
O
O
O
34-43
22
32
i
h
H
N
R3
O
N
N
O
O
O
O
30-31
33
Scheme 1. Reagents and conditions: (a) benzylbromide (2 equiv), K2CO3 (2 equiv), KI cat., CH3CN, rt, 80%; (b) n-BuLi (1.1 equiv), THF, À70 °C, R1CH2Br (1.1 equiv) or EtI
(1.1 equiv), 8–30%; (c) LiHMDS (1.2 equiv), THF, À70 °C, R1CH2Br (1.3 equiv), 40–60%. (d) TFA, DCM, rt, 100%; (e) CH3COCl (2 equiv), Na2CO3 (10 equiv), CH2Cl2/H2O (1:1), rt,
50–100%; (f) NaOH, EtOH, 70 °C, 87%; (g) R3Br (1 equiv), K2CO3 (1.1 equiv), KI cat., CH3CN, 50 °C, 25–80%; (h) BnNH2 or BnNHMe (1.2 equiv), EDCI (1.5 equiv), HOBt (1.5 equiv),
DIPEA (1.5 equiv), CH2Cl2, rt, 70–90%; (i) SOCl2, MeOH, rt, 95%; (j) CH3SO2Cl (1.2 equiv), NaH (1.2 equiv), THF, 60 °C, 43% or ClCOOCH3 (2 equiv), Pyridine (2 equiv), CH2Cl2, rt,
65% or EtI (3 equiv), K2CO3 (1.5 equiv), CH3CN, rt, 40% or EtNCO (1 equiv), CH2Cl2, rt, 100% or for R2 = RCO: RCOCl (2 equiv), Na2CO3 (10 equiv), CH2Cl2/H2O (1:1), rt, 56–76%.
Enantiomers would then be separated by chiral chromatography
at the final step if required. After TFA deprotection, the interme-
diates were acetylated to provide compounds 2–21. A shorter
route was also used starting from N-acetyl proline but the yields
of the alkylation step were always lower (below 30%) in particu-
lar with the less reactive bromides like cyclohexylmethyl bromide
or phenethyl bromide (conditions b). Compounds 23–29 were ob-
tained by reacting intermediate 22 with various acyl chlorides,
sulfonyl chlorides, chloroformates or isocyanates. After acetyl
protection and saponification of 22, the acid 32 was a key inter-
mediate to provide, on one hand, amides 30–31 by peptidic cou-
pling with amines and, on the other hand, esters 34–43 by
alkylation with the corresponding bromides.
by modulating its electron density. Moving a cyano group around
the phenyl ring indicates the ortho substitution is clearly disfa-
vored (as in 16 and 21). The meta and para substitution are both
very favorable and can accommodate rather different types of sub-
stituents. The most potent compounds are found with bulky
groups such as CF3 in meta position (compounds 18–20, IC50 from
50 to 80 nM) but those compounds are also amongst the most lipo-
philic ones (ALog P > 4.2).
We then modulated the N-acetyl moiety (Table 2). Deletions of
the acetyl (22) or the carbonyl (23) groups lead to a complete loss
in activity. Increasing the length of the alkyl of one methylene
slightly reduces the potency (compound 27, IC50 = 1.2 lM). A
cyclopropyl chain is beneficial for activity contrary to a phenyl
group. Carbamates, ureas and sulfonamides were also evaluated;
they have broadly similar activities to the corresponding amides
in the same range of lipophilicity.
Concerning the ester side chain (Table 3), we found that the
presence of the phenyl ring is essential for activity (the acid 22
and the methyl ester 23 are inactive). Moreover, this phenyl ring
in R4 can not be replaced by a cyclohexyl (36) or a pyridyl (37).
The substitution of this phenyl ring is also clearly detrimental to
We investigated the nature of the aromatic group R1 (Table 1).
The presence of the phenyl ring is essential for activity on KCC2
as its deletion gave compound 4 which is 100-fold less potent,
but it can be replaced by heteroaromatics (thienyl in 7, pyridyl in
6) or cycloalkyls (cyclohexyl in 5) with a minor loss of activity.
The increase in the chain length of one methylene provided com-
pound 3 with reduced activity (IC50 = 8
lM). We also studied the
effects of various substituents on the phenyl ring in R1 (Table 1)

2544
C. Pégurier et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2542–2545
Table 1
Table 3
KCC2 activities for compounds 2–21: 2-benzyl modulation
KCC2 activities for compounds 30–43: ester modulation
R1
O
N
X
R3
O
N
O
O
2-21
O
30-43
Compd
R1
IC50
(l
M)a
ALog Pb
R3
IC50
(
l
M)a
ALog Pb
2
Ph
Ph
Ph
CH2Ph
0.79
0.32
50
8
80
3.33
3.33
3.33
3.79
2.3
Compd
X
(+)-2
(À)-2
3
4
5
6
7
8
9
10
11
12
13
(+)-13
(À)-13
14
15
16
17
18
19
20
21
2
30
31
32
33
34
35
36
37
38
39
40
41
42
43
O
CH2Ph
CH2Ph
CH2Ph
H
CH3
CH(CH3)Ph
0.79
100
32
>100
>100
4
3.33
2.69
2.89
NA
1.75
3.71
3.65
3.94
NA
4
4
4
3.82
4.28
2.86
NH
NCH3
O
O
O
O
O
O
O
O
O
O
O
CH3
Cyclohexyl
4-Pyridyl
3-Thienyl
4-CH3-Ph
4-OCH3-Ph
4-CF3-Ph
4-F-Ph
1
2
4
NA
0.79
0.79
0.4
0.63
0.20
0.32
1
0.40
40
0.16
0.32
5
0.25
0.08
0.06
0.05
4
2.99
3.82
3.32
4.28
3.54
4
2.86
2.86
2.86
3.21
3.21
3.21
2.86
4.28
4.48
4.26
4.48
CH2CH2Ph
CH2cHex
10
63
CH2(4-pyridyl)
CH2(2-ClPh)
CH2(3-ClPh)
CH2(4-ClPh)
CH2(4-CH3Ph)
CH2(4-CF3Ph)
>100
6
2.5
1.3
0.79
5
4-Cl-Ph
4-(SO2Me)-Ph
4-(SO2Me)-Ph
4-(SO2Me)-Ph
4-CN-Ph
3-CN-Ph
2-CN-Ph
3-(SO2Me)-Ph
3-CF3-Ph
(3-CF3, 4-F)-Ph
(3-CF3, 4-OCH3)-Ph
(2-CF3, 4-F)-Ph
O
CH2(4-(SO2CH3)Ph)
>100
a
Rb influx assay on rKCC2.
b
Atom-based method to calculate the octanol–water partition coefficient for
non-ionisable compounds.10
The most potent KCC2 blockers from the benzyl prolinate series
were tested for their selectivity versus NKCC1, their aqueous solu-
bility and in vitro metabolic stability. Data are reported in Table 4.
They are all much less potent in inhibiting NKCC1 activity than
furosemide (1). Most compounds are oils and display low solubil-
ity. Moreover, compounds are generally extensively metabolized
on rat microsomes and hepatocytes. Surprisingly, compound 13
bearing a para methylsulfone in R2 displays the best profile with
a good aqueous solubility (0.5 mg/mL), a low clearance on rat
a
Rb influx assay on rKCC2.
b
Atom-based method to calculate the octanol–water partition coefficient for
non-ionisable compounds.10
Table 2
KCC2 activities for compounds 22–29: N-acyl modulation
microsomes (16
Caco-2 cells (Papps = 310 nm/s) and a low cytochrome P450 iso-
forms inhibition (<30% inhibition at 50 M for human Cyps 3A4,
2D6, 2C9, and 2C19). This compound 13 is one of the less lipophilic
of the series (ALog P = 2.86) and the polarity introduced by the
methylsulfone11 can partially explain this overall beneficial effect,
in particular, on the metabolic stability. The blockade of a potential
metabolic position could also contribute as the para CF3 analogue
10 and the para CN analogue 14 are also slightly more stable than
the original hit 2. Moreover, the meta methylsulfone analogue 17
lL/min/mg protein), a good permeability on
l
O
N
R2
O
22-29
Compd
R2
IC50
(l
M)a
ALog Pb
2
22
23
24
25
26
27
28
29
COCH3
H
0.79
3.33
NA
>100
>100
1.6
0.4
2
1.2
0.32
4
CH2CH3
SO2CH3
CO2CH3
CONHCH2CH3
COCH2CH3
COcPr
4.42
3.12
3.98
3.68
4
Table 4
NKCC1 activity, solubility and in vitro metabolic stability of selected compounds
Compd Inhibition%
NKCC1a
Solubilityb Clint on rat
microsomesc
Clint on rat
4.09
5
hepatocytesd
COPh
1
2
5
100
44
39
59
53
35
31
25
24
<20
nt
nt
nt
28
nt
23
20
12
31
24
20
23
a
Rb influx assay on rKCC2.
Atom-based method to calculate the octanol–water partition coefficient for
0.002
0.003e
0.01e
0.002e
0.5
385
382
333
128
16
123
107
356
400
b
non-ionisable compounds.10
9
10
13
14
17
19
24
0.02
nt
the activity, in particular in the ortho position (38) or with large
substituents (42, 43). The phenethyl ester 35 is 12-fold less active
suggesting that the lipophilic pocket is not very deep. The substitu-
tion of the benzylic position with a methyl causes a decrease in
affinity (compound 34, IC50 = 4 lM). Interestingly, replacing the es-
ter function by amides is detrimental to the KCC2 activity, the NH-
0.0005
0.001e
a
b
c
Rb flux assay on NKCC1 transporter at 100
In mg/mL in PBS buffer at pH = 7.4.
In
In
lM.
l
l
L/min/mg protein.
L/min/106 cells.
d
e
benzyl 30 and N-methyl-N-benzyl 31 analogues displaying IC50s of
100 and 32 lM, respectively.
Samples obtained as oils; nt = not tested.

C. Pégurier et al. / Bioorg. Med. Chem. Lett. 20 (2010) 2542–2545
2545
Table 5
zure models, which may yield more conclusive results regarding
the anticonvulsant potential of KCC2 inhibition.
Compared in vivo rat PK profile of hit and optimized compound
Compd t1/2a (h) Cla (mL/min/kg) AUCb (ng h/mL) Cmax (ng/mL)
F
b
(%)
Acknowledgments
2
13
2.4
0.3
61
26
8.6
726
11.9
457
0.5
18
This work was financed in part by the Walloon Region (Bel-
gium) under the Convention n°5296. The authors thank Véronique
Pinilla and her team for performing chiral chromatography and Be-
noit Mathieu for measuring solubility.
Compounds dosed at a1 mg/kg iv and b6 mg/kg po in male Wistar rats.
does not display similar low clearance such as 13. Compound 13
was also tested in an extended receptogram of more than 50 CNS
receptors, channels or enzymes and no activity was observed at
References and notes
1. Song, L.; Mercado, A.; Vazquez, N.; Xie, Q.; Desai, R.; George, A. L.; Gamba, G.;
Mount, D. B. Mol. Brain Res. 2002, 103, 91.
10 lM. Enantiomers of 13 were also prepared and a 100-fold dif-
ference in potency was found as for the hit compound 2 (compare
(+)-13 and its enantiomer (À)-13).
2. Rivera, C.; Voipio, J.; Payne, J. A.; Ruusuvuori, E.; Lahtinen, H.; Lamsa, K.;
Pirvola, U.; Saarma, M.; Kaila, K. Nature 1999, 397, 251.
3. De Koninck, Y. Curr. Opin. Pharmacol. 2007, 7, 93.
With its improved properties, compound 13 was investigated
further and the in vivo PK profile was evaluated in rat (Table 5).
Compound 13 showed in rat a lower clearance, higher plasma
exposures and an improved oral bioavailability compared to com-
pound 2. Plasma and brain exposure were also evaluated in NMRI
mice in order to assess its potential as tool for in vivo pharmacol-
ogy evaluation. At 5 minutes after ip dosing at 43.3 mg/kg, com-
4. Gamba, G. Physiol. Rev. 2005, 85, 423.
5. Hesdorffer, D. C.; Stables, J. P.; Hauser, W. A.; Annegers, J. F.; Cascino, G. Ann.
Neurol. 2001, 50, 458.
6. Delpire, E.; Days, E.; Lewis, L. M.; Mi, D.; Kim, K.; Lindsley, C. W.; Weaver, C. D.
PNAS 2009, 106, 5383.
7. The rat isoform of KCC2 was cloned and transfected into the human liver
adenocarcinoma cell line SK-Hep. A stable clone was selected by a 86Rb+ influx
assay. The KCC2 assay consists in the activation of the co-transporter by the
alkylating agent N-ethylmaleimide (which induces dephosphorylation of KCC2)
in the presence of the test compound and to induce the Rb+ influx by adding
5 mM RbCl to the extracellular medium. The Rb influx is stopped after 10 min
by several washing steps and the amount of intracellular Rb+ is determined by
atomic absorption spectroscopy.
pound 13 reached free brain levels close to 6 lM, which was
sixfold higher than the in vitro IC50 for KCC2 and then sufficient
to be evaluated in an in vivo model of epilepsy, the audiogenic
mouse seizure test.12 Compound 13 is inactive in this test up to
75 mg/kg (ip injection 5 min before stimulation). Thus, the block-
ade of KCC2 co-transporter doesn’t protect against sound-induced
seizure in the audiogenic mouse. This observation may be consis-
tent with the fact that furosemide shows limited efficacy in this
model only at doses exceeding 100 mg/kg.14 Furthermore, furose-
mide displays anti-epileptiform activity in vitro only at mM con-
centrations13, which are much higher than its IC50s for KCC2 and
8. The NKCC1 assay was developed on a MDCK (Madin Darby Canine Kidney) cell
line expressing an endogenous canine NKCC1 isoform. The NKCC1 co-
transporter is activated by incubating the cells during 2 h in
a
ClÀ free
medium (which induces phosphorylation of the co-transporter) and Rb+ influx
is stimulated during 45 min by adding 20 mM RbCl to the extracellular
solution. The amount of intracellular Rb+ is determined by atomic absorption
spectroscopy.
9. Confalone, P. N.; Huie, E. M.; Ko, S. S.; Cole, G. M. J. Org. Chem. 1988, 53, 482.
10. Ghose, A. K.; Crippen, G. M. J. Comput. Chem. 1986, 7, 565.
11. Leach, A. G.; Jones, H. D.; Cosgrove, D. A.; Kenny, P. W.; Ruston, L.; MacFaul, P.;
Wood, J. M.; Colclough, N.; Law, B. J. Med. Chem. 2006, 49, 6672.
12. Male genetically sound-sensitive mice, weighing 18–30 g, are used. They are
submitted to a first challenge for selection of responding animals. An acoustic
stimulus (85 dB, 10–20 kHz) is delivered for 30 s. The mice are observed and
the presence of the three phases of the seizure activity (wild running, clonic
and tonic convulsions) is noted. The next day, compounds are administered
before induction of the acoustic stimulus and the proportion of animals
protected against the three phases is calculated in each group (N = 10 animals/
tested dose).
13. Reid, K. H.; Guo, S. Z.; Iyer, V. G. Brain Res. 2000, 864, 134.
14. Margineanu, D. G.; Klitgaard, H. Epilepsy Res. 2006, 69, 93.
15. (a) Holtkamp, M.; Matzen, J.; Buchheim, K.; Walker, M. C.; Meierkord, H.
Epilepsia 2003, 44, 1141; (b) Luszczki, J. J.; Sawicka, K. M.; Kozinska, J.;
Borowicz, K. K.; Czuczwar, S. J. Epilepsy Res. 2007, 76, 66.
NKCC1, 25 lM and 3 lM, respectively. Thus, anticonvulsant activ-
ity of furosemide appears to be model dependent15 and difficult to
link specifically with KCC2 inhibition.
In conclusion, we describe the discovery of a novel class of
selective submicromolar KCC2 blockers: the benzyl N-acyl 2-ben-
zylprolinate derivatives. SAR modulations allowed for the identifi-
cation of an analogue 13 with drug-like properties and good brain
exposure in mice. This compound was found inactive in an in vivo
model of seizures, which could argue against the concept of using
KCC2 blockers for the treatment of epilepsy. However, further
characterization of compound 13 will be performed in other sei-