Discovery of potent and selective urea-based ROCK inhibitors and their effects on intraocular pressure in rats

Article abstract of DOI:10.1021/ml1000382

A series of urea-based Rho kinase (ROCK) inhibitors were designed and evaluated. The discovered compounds had excellent enzyme and cellular potency, high kinase selectivity, high aqueous solubility, good porcine corneal penetration, and appropriate DMPK profiles for topical applications as antiglaucoma therapeutics.

Full text of DOI:10.1021/ml1000382

pubs.acs.org/acsmedchemlett
Discovery of Potent and Selective Urea-Based ROCK
Inhibitors and Their Effects on Intraocular Pressure in Rats
†
†
†
†
†
†
€
Yan Yin, Michael D. Cameron, Li Lin, Susan Khan, Thomas Schroter, Wayne Grant,
†
†
‡
†
,†
€
Jennifer Pocas, Yen Ting Chen, Stephan Schurer, Alok Pachori, Philip LoGrasso,* and
Yangbo Feng*,†
†Translational Research Institute and Department of Molecular Therapeutics, 130 Scripps Way, #2A1, Jupiter, Florida 33458,
and ‡Department of Pharmacology and Center for Computational Science, University of Miami, Miami, Florida 33136
ABSTRACT A series of urea-based Rho kinase (ROCK) inhibitors were designed
and evaluated. The discovered compounds had excellent enzyme and cellular
potency, high kinase selectivity, high aqueous solubility, good porcine corneal
penetration, and appropriate DMPK profiles for topical applications as antiglau-
coma therapeutics.
KEYWORDS Glaucoma, ROCK, IOP, kinase inhibitor, urea, pyrazole
laucoma is a sight-threatening optic neuropathy dis-
ease, and it is estimated that over 60 million people
G
suffer from this disease worldwide.1 Primary open-
angle glaucoma, its most common form, is associated with
blockage of aqueous humor outflow, which results in ele-
vated intraocular pressure (IOP), leading to optic nerve
damage, progressive vision loss, and irreversible blindness.
Extensive efforts have been made to develop IOP-lowering
drugs that balance the rate of aqueous humor produced by
ciliary processes and the rate of aqueous outflow.2-4 How-
ever, current medical treatments are unable to efficiently
control IOP or lead to unacceptable side effects in many
patients. Therefore, the discovery of more effective and
noninvasive antiglaucoma medications is still needed.
Rho kinase (ROCK), a serine/threonine kinase, is a down-
stream effector of RhoA. Two isoforms of ROCK were
identified as ROCK-I (or ROCKβ) and ROCK-II (ROCKR),5
and both were shown to be expressed in human trabecular
meshwork (TM) and ciliary muscle (CM) cells.6 Upon activa-
tion by GTP-bound RhoA, ROCK phosphorylates the myosin
light chain (MLC) and myosin phosphatase, leading to
smooth muscle contraction.7 Recently, ROCK has attracted
numerous and diverse interest, especially as a new antiglau-
coma target.8-11 Some ATP-competitive small molecule
ROCK inhibitors including Y-27632,8 Y-39983,9 SNJ-1656,10
and H-1152P11 were reported to relax the TM cells and
increase the aqueous outflow facility and decrease IOP in
Figure 1. ROCK-II inhibitors.
ROCK-II appears to be one of the key elements for its high
potency.12 Its high selectivity was proposed to be mainly due
to the H-bond between the protonated tertiary amine of
the dimethylaminoethoxy moiety and the carboxylate side
chain of Asp176 of ROCK-II.12
To develop diversified chemotypes as novel ROCK-II
inhibitors for the treatment of glaucoma, a urea-based ROCK
inhibitor 1a (Figure 1) was recently identified in our group. In
this compound, an achiral benzyl amine moiety was used to
replace the chiral 3-substituted benzodioxane group in
SR3677. We anticipated that the benzyl group would pick
up the hydrophobic interactions between the benzodioxane
group and the hydrophobic pocket under the P-loop of ROCK
since the distance between the carbonyl group and the
terminal phenyl ring of 1a is similar to that in SR3677
(three bonds). As shown in Figure 1, 1a exhibited good
enzyme and cellular potency (IC50 = 18 and 113 nM, respec-
tively). However, the corresponding tetrahydroisoquino-
line (instead of benzylamine)-based analogue of 1a was a
modest ROCK inhibitor (IC50 = 280 nM) despite a resem-
blance to the benzodioxane group. Herein, we report the
optimization of 1a to obtain potent and selective ROCK
animal models or clinical trials. SR3677 (Figure 1),12
a
potent ROCK-II inhibitor developed by our group for the
treatment of glaucoma, exhibited excellent enzyme and cell-
based MLC bisphosphorylation13,14 (ppMLC) potency and
high selectivity. Pharmacology studies showed that SR3677
inhibited TM MLC phosphorylation and significantly incre-
ased ex vivo aqueous humor outflow in porcine eye.12 The
hydrophobic interaction between the benzodioxane moiety
of SR3677 and the hydrophobic pocket under the P-loop of
Received Date: February 18, 2010
Accepted Date: March 27, 2010
Published on Web Date: April 05, 2010
r
2010 American Chemical Society
175
DOI: 10.1021/ml1000382 ACS Med. Chem. Lett. 2010, 1, 175–179
|

pubs.acs.org/acsmedchemlett
Scheme 1a
Table 2. SAR of the Hinge-Binding Moiety
IC50a (nM)
compd
Ar
ROCK-II ppMLC PKA
1i
1j
3-methyl-1H-pyrazol-4-yl
pyridin-4-yl
2
5
90
816
348 2304
397 1031
a Reagents and conditions: (a) Amine or alcohol nucleophile, Cs2CO3,
1k
1l
2-aminopyrimidine-4-yl
2-aminopyridin-4-yl
13
44
188
DMF. (b) SnCl2 H2O, EtOAc. (c) (i) Triphosgene, NaHCO3, CH2Cl2, 0 °C;
3
>2700 5142
(ii) benzylamine derivatives, CH2Cl2, 0 °C. (d) Boronic acid pinacol
ester, Ph(PPh3)4, K2CO3, dioxane, H2O, 95 °C or (i) bis-pinacolato-
diboron, PdCl2(dppf), KOAc, dioxane, reflux; (ii) Ar-Cl, Ph(PPh3)4,
dioxane, K2CO3, H2O, 95 °C.
1m
7H-pyrrolo[2,3-d]pyrimidin-4-yl
>2700
744
a IC50 values were means of two or more experiments with errors
within 40% of the mean.
Table 1. SAR of Substitutions on the Benzyl Amine Phenyl Ring
(Table 1) and higher selectivity against PKA as compared to
1a (440- and 33-fold for 1c and 1a, respectively). Substitution
by groups other than methoxy at the 3-position was also
studied. All substituents investigated (NH2, Me, Cl, and F)
could be tolerated in the enzyme assays. However, the 3-Me
(1e) and 3-NH2 (1f) analogues showed reduced potency in
cell assays. One potential explanation for the low cell
potency of 1b (as compared to 1c) was likely due to its lower
aqueous solubility caused by intramolecular H-bonding
between the methoxy group and the benzyl NH group
(6-membered ring). This intramolecular H-bond was not
capable in the 3-position isomer (1c) and hence the better
aqueous solubility.15 Another possible explanation for
decreased cell potency could be due to different albumin
binding between 1b and 1c. The 3-Cl (1g) and 3-F (1h)
analogues had good cell potency, but their selectivity against
PKA (27- and 107-fold, respectively) was still lower than that
of 1c. Therefore, the 3-methoxyl-benzyl amine moiety was
used in subsequent SAR studies.
IC50a (nM)
compd
R1
ROCK-II
ppMLC
PKA
1b
1c
1d
1e
1f
2-OCH3
3-OCH3
4-OCH3
3-CH3
3-NH2
3-Cl
26
2
>2700
4
NDb
>2700
360
41
654
882
NDb
2630
405
2750
33
3
1g
1h
17
3
454
3-F
20
322
a IC50 values were means of two or more experiments with errors
within 40% of the mean. b Not determined.
inhibitors with appropriate properties to be used as potential
antiglaucoma therapeutics.
A short route to synthesize 1a and its analogues is shown
in Scheme 1. Starting from 2-fluoro-4-bromonitrobenzene
(2), a nucleophilic aromatic substitution reaction was app-
lied to introduce a substitution to give 4-bromonitrobenzene
3. Compound 3 was then reduced to aniline 4 using
Several heterocycles were applied as the hinge-binding
moiety to replace the 4-pyrazole in 1c. As shown in Table 2,
the addition of a methyl group to the 3-position of the
pyrazole ring resulted in a compound (1i) with similar
ROCK-II potency and PKA selectivity but with lower cellular
potency (IC50 = 90 vs 4 nM in 1c). Replacement of the
pyrazole by 6-membered heterocycles, including pyridine,
2-aminopyrimidine, and 2-aminopyridine (1j-l), also gave
SnCl2 2H2O. Urea 5 was obtained by adding a benzylamine
3
derivative (the secondary amine was prepared by a reduc-
tive amination reaction using a substituted benzaldehyde
and a primary amine and using NaCNBH3 as the reducing
agent in methanol at room temperature) to a mixture of 4,
triphosgene, and saturated NaHCO3 in dichloromethane at
0 °C. Finally, the targeted inhibitor 1 was prepared through a
Suzuki reaction or through a two-step borylation/Suzuki
coupling reaction.
potent ROCK inhibitors, but again, the cellular potency (IC50 =
348, 397, and >2700 nM, respectively) of these compounds
was much worse than that of 1c. Application of the pyrrolo-
pyrimidine (1m) group decreased not only the enzymatic
potency (IC50 = 188 nM) but also the selectivityagainst PKA.
As a result, compound 1c was still the lead for further
optimizations.
SAR studies on the central phenyl ring and on the urea
bond were performed to improve the pharmaceutical pro-
perties of these urea-based ROCK inhibitors (Table 3). Thus,
both electron-donating groups (dimethyl amino 1n and
methoxy 1o) and electron-withdrawing groups (chloro 1p
and fluoro 1q) were introduced to the central phenyl ring at
the position ortho to the urea group. Notably, all resulting
compounds 1n-q exhibited excellent ROCK-II potency
The structure-activity relationship (SAR) of substitutions
on the benzyl ring was first investigated. As shown in Table 1,
mapping the 2-, 3-, and 4-positions (compounds 1b-d)
demonstrated that a methoxy group at the 2- (1b) and
3-positions (1c) was well-tolerated, while substitution at
the 4-position was not. The 3-position substitution was
optimal with 1c having improved enzyme and cell potency
r
2010 American Chemical Society
176
DOI: 10.1021/ml1000382 ACS Med. Chem. Lett. 2010, 1, 175–179
|

pubs.acs.org/acsmedchemlett
Table 3. SAR of Side Chains on the Central Phenyl Ring
Figure 2. Compound 1r docked to the ROCK-II ATP-binding
pocket.
selected cytochrome P450 isoforms, 1A2, 2C9, 2D6, and 3A4,
was also studied, and results showed that inhibition of P450
was not a concern for these compounds with the highest
inhibition being ∼80% at 10 μM for CYP2D6 on compounds
1x and 1z (Table S1 of the Supporting Information).
Inhibitor 1r was docked into the catalytic domain of
human ROCK-II by methods described previously.12 The
obtained binding motif of the enzyme-ligand complex with
the highest docking score was shown in Figure 2, and the
docking pose explained very well the observed SAR. There
were four H-bonds between the enzyme and the ligand: two
between the pyrazole headgroup and the Glu170 and
Met172 in the hinge-binding region, one between the urea
carbonyl group and the side-chain N-H of Lys121 in the
phosphate binding site, and the last H-bond between the pro-
tonated tertiary amine of the dimethylaminoethoxy moiety
of the ligand and the side chain carboxylate of Asp176 in the
ribose site. In addition to hydrogen-bonding interactions,
another important contributor to the high potency and
selectivity of these inhibitors was the hydrophobic inter-
action between the benzyl amine phenyl moiety of ligand
and the hydrophobic pocket under the P-loop of the enzyme,
which was composed of the side chains of Phe103, Leu123,
and Phe136.
High aqueous solubility and high corneal penetration are
critical for these compounds to be used in glaucoma therapy
by topical administration.17 Thus, these properties of com-
pounds 1r and 1x-z were evaluated. Results showed that
these compounds had excellent aqueous solubility at pH 5.5
(1844 μg/mL for 1x, 3147 μg/mL for 1y, and 2847 μg/mL for
1z), and the porcine cornea penetration is similar to that of
the IOP-reducing drug, timolol.18
a IC50 values were means of two or more experiments with errors
within 40% of the mean. b Not determined.
(1 nM or below) and high cellular activity. However, these
compounds were also potent inhibitors for PKA with the
exception of 1q (IC50 = 1821 nM). The introduction of a
large tertiary amine-containing side chain was applied to
reduce PKA inhibition and to enhance the aqueous solubility.
As shown in Table 3, compounds 1r-t exhibited not only
excellent ROCK inhibition (IC50 values below 1 nM) and cell-
based potency (<4, <4, and 6 nM, respectively) but also
very low PKA inhibitions (IC50 = 3900, 2727, and 3573 nM,
respectively). Alkylation of the urea NH of 1r-t, which
decreases the H-bonding numbers of the inhibitor, was also
investigated. Both methylation (inhibitors 1u-w) and ethy-
lation (1x-z) were applied on the free NH of the benzyl
amine group. These compounds exhibited excellent enzyme
and cellular potency and reduced PKA activity with the
exception of compounds 1w and 1z, where R1 is N,N,N0-
trimethylethyldiamine. The biochemical potency of all of the
inhibitors in Table 3 exceeds that of all well-known ROCK
inhibitors such as Y-27632, Fasudil, or H-1152P. Indeed, 1v
and 1y are >110-fold more potent than Y-27632 in biochem-
ical assays and >100-fold more potent in cell assays.16
Similarly, Fasudil and H-1152P are both weak inhibitors in
cell-based assays with IC50 values in the micromolar range.16
The selectivity of inhibitors 1x-z against a few other
kinases was evaluated (see Table S1 in the Supporting Infor-
mation). These compounds were generally inactive against
JNK3 and p38, in addition to PKA, and only moderately
active against MRCKR. Screening against ROCK-I demon-
strated that these compounds were pan-ROCK inhibitors,
although the ROCK-I activity was lower. Inhibition to four
A large fraction of the administrated dose from ophthal-
mic drops is absorbed via the tear duct and enters systemic
circulation via the nasolacrimal duct. Systemic exposure of
ROCK inhibitors has the potential for cardiovascular or other
systemic side effects.16 ROCK inhibitors developed as anti-
glaucoma therapeutics are expected to be rapidly metabolized
upon entering general circulation to relieve these safety con-
cerns. The rat PK properties of compounds 1r, 1u, and 1x-z
(see Table S2 in the Supporting Information) demonstrated
r
2010 American Chemical Society
177
DOI: 10.1021/ml1000382 ACS Med. Chem. Lett. 2010, 1, 175–179
|

pubs.acs.org/acsmedchemlett
compounds 1a-1z. This material is available free of charge via
the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author: *To whom correspondence should be
addressed. (Y.F.) Tel: 561-228-2201. E-mail: yfeng@scripps.edu.
(P.L.) Tel: 561-228-2230. E-mail: lograsso@scripps.edu.
ACKNOWLEDGMENT We thank Professors Patrick Griffin and
William R. Roush for their support. We are grateful to Dr. Derek
Duckettand Weimin Chen for p38 and JNKassays and to Dr. Michael
Chalmers for HRMS. Experimental help from Claudia Ruiz and
Shannon Clapp are also greatly appreciated.
Figure 3. Rat IOP effects of 1y (IOP decreases relative to vehicle).
Decreases at 1 and 4 h were statistically significant with p < 0.05.
REFERENCES
(1)
(2)
(3)
(4)
Quigley, H. A.; Broman, A. T. The number of people with
glaucoma worldwide in 2010 and 2020. Br. J. Ophthalmol.
2006, 90, 262–267.
Coakes, R. L.; Brubaker, R. F. The mechanism of timolol in
lowering intraocular pressure in the normal eye. Arch.
Ophthalmol. 1978, 96, 2045–2048.
Toris, C. B.; Gleason, M. L.; Camras, C. B.; Yablonski, M. E.
Effects of brimonidine on aqueous humor dynamics in
human eyes. Arch. Ophthalmol. 1995, 113, 1514–1517.
Lippa, E. A.; Carlson, L. E.; Ehinger, B.; Eriksson, L. O.;
Finnstrom, K.; Holmin, C.; Nilsson, S. E.; Nyman, K.; Raitta,
C.; Ringvold, A. Dose response and duration of action of
dorzolamide, a topical carbonic anhydrase inhibitor. Arch.
Ophthalmol. 1992, 110, 495–499.
Nakagawa, O.; Fujisawa, K.; Ishizaki, T.; Saito, Y.; Nakao, K.;
Narumiya, S. ROCK-I and ROCK-II, two isoforms of Rho-
associated coil-coil forming protein serine/threonine kinase
in mice. FEBS Lett. 1996, 392, 189–193.
Nakajima, E.; Nakajima, T.; Minagawa, Y.; Shearer, T. R.;
Azuma, M. Contribution of ROCK in contraction of trabecular
meshwork: Proposed mechanism for regulating aqueous
outflow in monkey and human eyes. J. Pharm. Sci. 2005,
47, 701–708.
Uehata, M.; Ishizaki, T.; Satoh, H.; Ono, T.; Kawahara, T.;
Morishita, T.; Tamakawa, H.; Yamagami, K.; Inui, J.; Maekawa,
M.; Narumiya, S. Calcium sensitization of smooth muscle
mediated by a Rho-associated protein kinase in hypertension.
Nature 1997, 389, 990–994.
Waki, M.; Yoshida, Y.; Oka, T.; Azuma, M. Reduction of
intraocular pressure by topical administration of an inhibitor
of the Rho-associated protein kinase. Curr. Eye Res. 2001, 22,
470–474.
Tokusbige, H.; Inatani, M.; Nemoto, S.; Sakaki, H.; Katayama,
K.; Uebata, M.; Tanibara, H. Effects of topical administration
of Y-39983, a selective Rho-associated protein kinase inhi-
bitor, on ocular tissues in rabbits and monkeys. Invest.
Ophthalmol. Vis. Sci. 2007, 48, 3216–3222.
that these ROCK inhibitors have high clearance (Cl), low oral
Cmax, and very low bioavailability (F), which we believe are
desirable for topical antiglaucoma applications.
To test the effect of these novel urea-based ROCK inhi-
bitors on lowering the IOP, compound 1y was applied to rat
eyes. As shown in Figure 3, statistically significant decreases
in IOP were detected when 40 μg (2 ꢀ 20 μL drops of a 0.1%
solution) of 1y was applied to the eyes of Brown Norway rats
(n = 7/group) housed under constant low light conditions.
Initial IOP was 29 mmHg. This elevated IOP model has been
suggested to be akin to the magnitude changes seen in
glaucoma patients,19,20 making it a reasonable experimental
model reflective of clinical conditions and therefore helpful
in compound development. The maximal IOP decrease was
∼7 mmHg relative to that of vehicle from 1 to 4 h, returning
to baseline at 8 h as compared to the vehicle. These results
demonstrated that these urea-based ROCK inhibitors are
good candidates for further IOP studies. One challenge will
be to improve upon the duration of action for this class of
compounds, which will be future optimization focus driven
by ocular stability assays and insights of the inhibitors'
physicochemical properties.
In conclusion, we have developed a series of potent and
selective urea-based novel ROCK-II inhibitors. These inhibi-
tors exhibited low inhibitions over PKA, MRCKR, JNK3, p38,
and selected cytochrome P450 isoforms, had good aqueous
solubility and corneal penetration, and possessed high sys-
temic clearance and low oral bioavailability. All of these
properties are desirable for these ROCK inhibitors to be used
in the topical administration as antiglaucoma agents. Signi-
ficant IOP-lowering effects on rat eyes were observed for
compound 1y, further demonstrating the effectiveness of
these urea-based ROCK inhibitors as potential antiglaucoma
therapeutics. Future work will be directed at improving the
duration of action for these compounds so that IOP lowering
could be accomplished by aq.d dosing regimen. Once these
properties are achieved, ocular toxicology studies that exa-
mine conjunctival hyperemia will be investigated in appro-
priate species as other ROCK inhibitors have shown hyper-
emia in rabbits and monkeys after long-term dosing.9
(5)
(6)
(7)
(8)
(9)
(10) Tanihara, H.; Inatani, M.; Honjo, M.; Tokushige, H.; Azuma, J.;
Araie, M. Intraocular pressure-lowering effects and safety of
topical administration of a selective ROCK inhibitor, SNJ-
1656, in healthy volunteers. Arch. Ophthalmol. 2008, 126,
309–315.
(11) Nishio, M.; Fukunaga, T.; Sugimoto, M.; Ikesugi, K.; Sumi, K.;
Hidaka, H.; Uji, Y. The effect of the H-1152p, a potent
Rho-associated coiled coil-formed protein kinase inhibitor,
in rabbit normal and ocular hypertensive eyes. Curr. Eye Res.
2009, 34, 282–286.
SUPPORTING INFORMATION AVAILABLE Details for syn-
thetic procedures, biological studies, and analytical data for
r
2010 American Chemical Society
178
DOI: 10.1021/ml1000382 ACS Med. Chem. Lett. 2010, 1, 175–179
|

pubs.acs.org/acsmedchemlett
(12) Feng, Y.; Yin, Y.; Weiser, A.; Griffin, E.; Cameron, M.; Lin, L.;
€
€
Ruiz, C.; Schurer, S. C.; Inoue, T.; Rao, V.; Schroter, T.;
LoGrasso, P. Discovery of Substituted 4-(Pyrazol-4-yl)-phenyl-
benzodioxane-2-carboxamides as Potent and Highly Selec-
tive Rho Kinase (ROCK-II) Inhibitors. J. Med. Chem. 2008, 51,
6642–6645.
€
(13) Schroter, T.; Minond, D.; Weiser, A.; Dao, C.; Habel, J.; Spicer,
€
T.; Chase, P.; Baillargeon, P.; Scampavia, L.; Schurer, S.;
Chung, C.; Mader, C.; Southern, M.; Tsinoremas, N.;
LoGrasso, P.; Hodder, P. Comparison of miniaturized time-
resolved fluorescence resonance energy transferand enzyme-
coupled luciferase high-throughput screening assays to
discovery inhibitors of Rho-kinase II (ROCK-II). J. Biomol.
Screening 2008, 13, 17–28.
€
(14) Schroter, T.; Griffin, E.; Weiser, A.; Feng, Y.; LoGrasso, P.
Detection of myosin light chain phosphorylation-Acell-based
assay for screening Rho-kinase inhibitors. Biochem. Biophys.
Res. Commun. 2008, 374, 356–360.
(15) As indicated in the Supporting Information, in enzyme
assays, compounds were diluted directly from 90% DMSO/
H2O media into the assay plates, while in ppMLC cell assays,
compounds were diluted from 20% DMSO/H2O or aqueous
buffer media into the assay plates. Therefore, there was no
solubility issue for compound 1b in the enzyme assays, but it
could be a problem in the cell assay. For details, see refs 13
and 14.
(16) Ikenoya, M.; Hidaka, H.; Hosoya, T.; Suzuki, M.; Yamamoto,
N.; Sasaki, Y. Inhibition of rho-kinase-induced myristoylated
alanine-rich C kinase substrate (MARCKS) phosphorylation
in human neuronal cells by H-1152, a novel and specific Rho-
kinase inhibitor. J. Neurochem. 2002, 81, 9–16.
(17) Shirasaki, Y. Molecular design for enhancement of ocular
penetration. J. Pharm. Sci. 2007, 97, 2462–2496.
(18) Corneal permeability was tested in a three-chamber dialysis
system with porcine cornea acting as the partitions between
the chambers. The cornea was orientated so that the exterior
surface was facing the central compound compartment. All
three chambers contained pH 7.4 buffer, and the central
chamber also contained 100 μM compound. Compound
levels were determined in the two terminal chambers at
4 h. Values are the means of the two chambers. The glaucoma
drug timolol was used as a reference. Concentrations in the
buffer chambers at the 4 h time point were 1.6 μM for timolol,
1.1 μM for 1x, 1.5 μM for 1y, and 1.0 μM for 1z.
(19) Pang, I. H.; Clark, A. F. Rodent models for glaucoma retino-
pathy and optic neuropathy. J. Glaucoma 2007, 16, 483–505.
(20) Pang, I. H.; Wang, W. H.; Clark, A. F. Acute effects of glaucoma
medications on rat intraocular pressure. Exp. Eye Res. 2005,
80, 207–214.
r
2010 American Chemical Society
179
DOI: 10.1021/ml1000382 ACS Med. Chem. Lett. 2010, 1, 175–179
|