A vHTS approach for the identification of β-adrenoceptor ligands

Article abstract of DOI:10.1016/j.bmcl.2010.04.009

Using a vHTS based on a pharmacophore alignment on known β3-adrenoceptor ligands, a set of intriguing β-adrenoceptor ligands was identified, optimization of which resulted in a selective and potent human β2-AR antagonist.

Full text of DOI:10.1016/j.bmcl.2010.04.009

Bioorganic & Medicinal Chemistry Letters 20 (2010) 3399–3404
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
A vHTS approach for the identification of b-adrenoceptor ligands
a
a
a
a
Stefan Tasler a, , Roland Baumgartner , Andrea Aschenbrenner , Astrid Ammendola , Kristina Wolf ,
*
Tanja Wieber a, Josef Schachtner a, Marcus Blisse a, Udo Quotschalla b, Peter Ney b
a 4SC AG, Am Klopferspitz 19a, 82152 Planegg-Martinsried, Germany
b UCB Pharma S.A., Chemin du Foriest, 1420 Braine l’Alleud, Belgium
a r t i c l e i n f o
a b s t r a c t
Article history:
Using a vHTS based on a pharmacophore alignment on known b3-adrenoceptor ligands, a set of intriguing
b-adrenoceptor ligands was identified, optimization of which resulted in a selective and potent human
b2-AR antagonist.
Received 10 March 2010
Revised 30 March 2010
Accepted 5 April 2010
Available online 9 April 2010
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
Adrenoceptor
GPCR
Molecular modeling
Virtual high throughput screening
Many ligands for b-adrenergic receptors are currently used to
treat a range of disorders. These marketed drugs still address pur-
posely only the human b1- and b2-adrenergic receptor subtypes
(hb1- and hb2-AR). This is due to the fact that a third subtype,
hb3-AR, was not discovered and cloned until the late 80s1 and clin-
ical trials for the use of hb3-agonists to treat obesity and diabetes—
which represented the primary focus back then—were rather dis-
appointing.2–4 b-Blockers, most ideally b1-selective, are effective
in the treatment of angina pectoris, heart failure and hyperten-
sion.2,5 The inhalation of b2-adrenoceptor-selective agonists has
been established as therapy of asthma and other bronchospastic
conditions for some time.2,6 Furthermore, systemic b2-agonists
are used as tocolytics in preterm labor.6–8 With the hb3-AR being
predominantly located in adipose tissue but also playing a crucial
role in the gastrointestinal tract and the bladder smooth muscle,
stimulation of this receptor represents a potential therapeutic ap-
proach to treat irritable bowel syndrome or urinary inconti-
nence2,9,10—and still obesity, with a phase II clinical trial in this
indication ongoing for LY-377604 in combination with Sibutr-
amine.11 Upon the identification of the b3-AR in near-term human
myometrium, agonists of this receptor have also been discussed as
rescue therapy for the treatment of preterm labor.7
sin receptor was available as a representative of the target class of
GPCRs,12 with only around 40% sequence homology as compared to
hb3-AR. This changed not before 2007 when crystal structures
were resolved for the human b2-receptor13 and, shortly after-
wards, for the turkey b1-receptor.14 Consequently, a direct docking
approach on hb3-AR using a homology model was not selected as
method of choice back then. For an alignment approach, five differ-
ent literature-known hb3-AR agonists were selected as templates
based on their advancement in clinical development, their activity
on b3-receptors and/or for structural reasons (Fig. 1): two phase II
candidates at the given point in time (KUC-7322 and N-5984, dis-
playing EC50 values in the lower or even sub-nanomolar range in
different functional b3-AR assays),9,15 or structurally intriguing
compounds 1 (CP-331,679, full agonist, EC50(hb3) = 300 nM),3,4
2
(sub-nanomolar agonist on b3-AR)4 and the non-subtype-selective
isoquinoline 3 (one-digit nanomolar EC50 values on all three hu-
man b-adrenergic subtypes).16
Each of these templates was used for a vHTS run of a virtual li-
brary comprising 4.6 Mio commercially available compounds using
the 4SCanÒ technology.17 Top scoring 5000 molecules of each run
were combined and filtered by molecular weight (<600 g/mol),
TPSA (<200 Å2), number of rotatable bonds (<20) and violation of
Lipinski rules (max 1). Out of the resulting list of approximately
800 compounds of interest, about 210 were selected driven by
chemical diversity for biological screening in a binding assay on
hb3-AR.18 Three compounds were identified with Ki values
Focusing primarily on the identification of new hb3-adrenocep-
tor ligands, a pharmacophore alignment was chosen for a virtual
high throughput screening (vHTS) approach on hb3-AR. At the gi-
ven point in time, only the crystal structure of the bovine rhodop-
<20 lM, based on which a ‘back-screening’ was initiated: another
420 compounds were selected for biological evaluation according
to their structural similarity with the initial hit molecules. This
* Corresponding author. Tel.: +49 89 700763 0; fax: +49 89 700763 29.
E-mail address: stefan.tasler@4sc.com (S. Tasler).
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.04.009

3400
S. Tasler et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3399–3404
OH
OH
H
N
H
N
O
O
H
O
OH
OH
HO
O
Cl
Cl
O
N-5984
KUC-7322
HO
HO
OH
O
N
OH
OH
H
N
H
N
NH
O
S
O
I
H2N
N
HO
O
O
O
HN
1
2
3
S
I
O
O
OH
H
N
OH
N
OH
O
N
O
N
O
N
O
S
O
N
O
N
N
4
Ki(hß3-AR) = 411 nM
5
Ki(hß3-AR) = 409 nM
6
Ki(hß3-AR) = 378 nM
Figure 1. Templates chosen for a pharmacophore alignment and vHTS hits 4–6.
approach yielded compounds displaying even better affinity, now
with sub-micromolar Ki values, the best of which are depicted in
Figure 1 (compounds 4–6).19
hß2-AR. With regard to functional activity, compound 4 possessed
antagonistic activity on all three b-AR subtypes, the strongest ef-
fect again on hb2-AR with an IC50 value of 377 nM. As mentioned
above, slight structural changes might already alter functional
activity significantly. This was confirmed by the structurally very
similar naphthalimide 5, which displayed weak partial agonism
on hb3-AR with an intrinsic activity of 65% as compared to the
standard isoproterenol. For thienopyrimidine 6, the sub-micromo-
lar affinity for hb3-AR as detected in the initial screening was con-
firmed. It already displayed a slight selectivity for hb3-AR over
hb2-AR based on binding data and proved to be a partial agonist
on hb3-AR as well, but now on an even better activity level with
an EC50 value of 120 nM. These results validated an intriguing
starting point for medicinal chemistry. SAR evaluations for the
naphthalimides will be described herein, whereas those for the thi-
enopyrimidines will be disclosed in due course.22
For a systematic evaluation of structural effects on selectivity
and functional activity within the naphthalimide series, the aryl-
oxypropanolamine array was kept constant and the substitution
pattern of the aryloxy portion as well as the aminopiperidine
spacer unit were subjected to alterations. Synthetically, such
derivatives were obtained as follows: required 2-(aryloxymeh-
tyl)oxiranes 9 were prepared by O-alkylation of phenols 7 using
epibromohydrine (8).23 4-Piperidinyl and piperidin-4-ylmethyl-
naphthalimides 12 were attained from the corresponding primary
amine 10 and a 1,8-naphthalic anhydride 11. For smaller imide
building blocks 15, deprotonation of phthalimide (14a) or glutar-
imide (14b) and subsequent N-alkylation with 4-(bromo-
methyl)piperidine 13 proved to be a better route as compared to
transformation of cyclic anhydrides with the corresponding
amines. Finally, reaction of oxiranes 9 with piperidines 12 or 15 oc-
curred in isopropanol at 60–80 °C.24 This procedure was likewise
effective with morpholine (cf. compound 35, Table 4), 1-methylpi-
peridin-4-amine (cf. compounds 36 and 37), 4-benzylpiperidines
(cf. compounds 41–43, Table 5) and 4-benzoylpiperidines (cf. com-
pounds 44–46). DIEA had to be added only if the respective amine
was used as its hydrochloride salt. For compounds bearing a
4-hydroxyaryloxy moiety (R = OH, Scheme 1), the sequence was
realized with an O-Bz protective group, which was saponified in
the final product using 3 N aq NaOH in dioxane (100 °C, 2 h). Same
As an alignment procedure is based on the generation of a
‘pseudo-receptor’ out of the template pharmacophore, application
of this method allows only for matching interaction possibilities of
the template used but does not provide any information about the
binding site itself. Subtle changes within a given structural scaffold
can result in drastic effects on selectivity and/or functional activity,
potentially covering the whole range from an agonist to an inverse
agonist, with such variations possibly not lying within the pseudo-
receptor area. Thus, such a modeling approach is very useful for the
identification of new structural entities as ligands for a target
group of high homology (e.g., the b-adrenoceptors), a prediction
of selectivity and/or functional activity, however, can not necessar-
ily be expected—even though the selection of templates might bias
the outcome. Consequently, a hit validation had to follow, includ-
ing the performance of binding and functional assays for all three
b-AR subtypes.20,21
For naphthalimide derivatives 4 and 5, sub-micromolar affini-
ties on hb3-AR as attained from the initial screening campaign
could not be confirmed. Both compounds displayed Ki values
slightly below 2 lM (Table 1). However, affinity for hb2-AR was
quite good with Ki values of 80 nM for both compounds, one dis-
playing already some selectivity over hb1-AR (compound 4, by fac-
tor 12). Compound 5 exhibited similar affinities for hb1-AR and
Table 1
Binding and functional data on b-AR subtypes for hit molecules out of vHTS
Compd Binding assay, Ki [nM]
Functional assay, EC50/IC50 [nM] (IA)a
hb2- hb3- hb1-AR hb2-AR hb3-AR
AR AR
80 1630 IC50 = 1375 IC50 = 377 IC50 = 1474
80
hb1-
AR
4
5
1000
80
1850 nd
460 nd
nd
nd
EC50 = 2400
(0.65)
6
>10,000 700
EC50 = 120
(0.6)
a
EC50 indicates agonistic, IC50 indicates antagonistic effects; IA = intrinsic activ-
ity, relative to isoproterenol (IA = 1.0); nd = not determined.

S. Tasler et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3399–3404
3401
Table 2
O
Binding data of naphthalimide derivatives on the three b-AR subtypes: variations of
OH
O
O
the aryloxy moiety19
i
R
R
+
Br
O
28-94%
OH
N
7
8
9
O
N
O
Ar
0,1
NH
2
O
Compd
Ar
Binding assay, Ki [nM]
hb2-AR
N
0,1
N
Boc
R'
ii, iii
hb1-AR
hb3-AR
10
+
11
HN
O
O
R'
O
4a
1000
80
80
1630
12
R' = OH
iv, 41%
(prior to iii)
O
R' = OMe
5a
80
1850
O
O
N
HN
HN
O
O
O
v, iii
Br
14a
14b
16
250
230
1030
1180
15
+
N
O
Boc
39-52%
N
HN
O
13
HN
O
O
17a
1360
3290
O
0,1
O
18a
530
300
3380
N
O
HN
R
+
O
O
9
12, 15
vi 12-86%
19
20
4530
27
1260
120
8500
1000
N
H
O
0,1
HN
OH
N
O
N
O
R
cf. Table entries
Scheme 1. Reagents and conditions: (i) K2CO3, butanone, 80 °C, 24–48 h; (ii) DMF,
160 °C, 1–3 h (R0 = H, OH), 64–80%; (iii) HCl in dioxane (4.0 M), 0 °C, 2 h, quant.; (iv)
conducted in between steps ii and iii: (a) NaH, DMF, 0 °C to rt, 30 min; (b) MeI, 0 °C
to rt, 30 min; (v) (a) NaH, DMF, 0 °C to rt, 30 min; (b) 13, 80–100 °C, 10–20 h; (vi)
depending on solubility: iPrOH or iPrOH/DMSO (up to 1:1), DIEA (optional), 60–
80 °C, 2–8 h; yields are not optimized.
HN
21
7
3
225
inactive (>10 µM) at hβ3-AR:
a
O
R = OH, OMe,
Cl,a F,a
conditions were applied for the conversion of a benzoic ester deriv-
ative into the corresponding acid (R = COOMe to R = COOH,
Table 2).
HO
O
COOH,
R
COOMe,
COMe,a
H
Within the set of differently substituted aryloxy units, a few
patterns were also realized known from other b-adrenoceptor li-
gands: p-hydroxyphenyloxy (Ritodrine, b2-agonist; KUL-7211,
b2/b3-agonist),9 carbazolyloxy (Carvedilol, b1/b2/b3-antagonist;
Carazolol, b1/b2-antagonist; LY-377604, b3-agonist), 4-indolyloxy
(pindolol, b1/b2-antagonist), 2-allylphenyloxy (Alprenolol, b1/b2-
antagonist; SR58894, b3-antagonist; already present in hit mole-
cule 5)2,9,25 and 4-allyl-2-methoxyphenyloxy (b1/b2-antagonist).2
As can be seen from Table 2, none of the aryloxy variations resulted
in a shift of selectivity towards hb3-AR. The best Ki on this receptor
was observed for the carbazolyl derivative 21 with 225 nM, which,
however, also experienced a significant increase of affinity for hb1-
and hb2-AR, now being in the one-digit nanomolar region. Gener-
ally, these naphthalimide derivatives displayed best affinities for
hb2-AR. The only compound with some selectivity for hb1-AR
a
Compound from a commercial source.
within a decent affinity range was 4-indolyl derivative 20 with a
Ki of 27 nM. 2-Allyl and 2-isopropyl derivatives 5 and 16 displayed
similar affinity for hb1- and hb2-AR. Incorporation of an additional
methyl group, however, resulted in some hb2-AR selectivity (com-
pound 4 vs 16, factor 12 towards hb1-AR, factor 20 towards hb3-
AR). Attachment of oxygen-based polar groups was generally det-
rimental to affinity for either receptor (compound 17 and summary
of inactive derivatives, Table 2; to a lesser extent also 18). Not even
respective aromatics deduced from reference ligands were able to

3402
S. Tasler et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3399–3404
Table 3
Table 5
Binding data of naphthalimide derivatives on the three b-AR subtypes: variations of
Binding data of carbazol-4-yloxy derivatives on the three b-AR subtypes: variations of
the spacer unit19
the imide part19
O
X
OH
N
OH
HN
R
spacer O
O
O
N
Compd
X
R0
Binding assay, Ki [nM]
Compd
Spacer
R
Binding assay, Ki [nM]
hb1-AR
hb2-AR
hb3-AR
hb1-AR
hb2-AR
hb3-AR
21
38
39
H
OH
OMe
7
250
0.1
3
37
8
225
590
410
4
5
16
2-iPr-5-Me
2-Allyl
2-iPr
1000
80
250
80
80
230
1630
1850
1030
O
R'
N
N
N
O
22a
2-iPr-5-Me
2-Allyl
2-iPr
Inactive
293
740
110
150
70
>10,000
3100
1480
O
N
23a
N
24a
25
26
2-Allyl
2-iPr
1050
240
40
13
2340
3820
O
40
76
14
57
98
N
a
Compound from a commercial source.
41
42
H
OMe
0.5
0.2
230
455
3.4
R'
R'
give optimized affinities (p-hydroxyphenoxy, and 4-allyl-2-meth-
oxyphenyloxy, compound 17).
43
1.0
16
13
32
0.1
72
Next, the influence of spacer length in between the propanol-
amine and the naphthalimide unit on affinity and selectivity was
investigated (Table 3). In all three variants, the spacer was based
on a six-membered heterocyclic ring (piperidine or piperazine),
and the alkyl chain (CH2)n attached to the cyclic amine was varied
from n = 0 to n = 2. In order to allow for a thorough SAR conclusion,
these variants were incorporated into three different sub-series de-
fined by the substitution pattern at the aryloxy portion (2-isopro-
pyl-5-methyl vs 2-allyl vs 2-isopropyl).
N
44
45
46
H
OMe
F
0.2
0.2
0.3
230
9180
250
O
accompanied by a piperidine exchange to piperazine for the ex-
tended n = 2 version, though) enhanced selectivity for hb2-AR. This
effect was caused by a general decrease of affinity for b3-AR and—
with only one exemption (compound 26 vs 16)—for hb1-AR. With-
For all three subsets of aryloxy variants, extension or diminu-
tion of the spacer unit (CH2)n by one methylene unit (which was
Table 4
Binding data of naphthalimide derivatives on the three b-AR subtypes: variations of the imide part19
R
OH
O
X
Compd
X
R
R0
Binding assay, Ki [nM]
hb2-AR
hb1-AR
hb3-AR
5
27
28
16
29
Allyl
Allyl
Allyl
iPr
H
OH
OMe
H
OMe
80
170
300
250
209
80
135
78
230
38
1850
2200
1510
1030
698
O
R'
N
N
N
O
iPr
30
31
Allyl
iPr
190
720
50
17
3400
1480
O
N
N
O
32
33
Allyl
iPr
440
245
80
9
11,000
2900
O
N
N
O
34a
35
Allyl
iPr
114
12
7
1.2
4960
1770
O
36
37
Allyl
iPr
260
89
2.4
17
6225
1060
H
N
N
a
Compound from a commercial source.

S. Tasler et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3399–3404
3403
in the shortened piperidine series (n = 0; compounds 25 and 26),
affinity for hb2-AR was simultaneously increased, resulting in a
good Ki of 13 nM for the 2-iPr derivative 26.
excellent affinity for hb1-AR was attained (compound 39;
Ki = 0.1 nM), displaying a decent selectivity over hb2-AR. Only
shortening the spacer unit by one methylene group as realized in
compound 40 resulted in a stronger emphasis of affinity for hb3-
AR by reducing affinities for hb1- and hb2-AR. The overall outcome
was a mediocre Ki(hb3) of 98 nM and a comparable affinity level
for both of the other b-subtypes. A replacement of the naphthalim-
idomethylene by either benzyl or benzoyl derivatives eventually
shifted the focus completely: excellent picomolar affinities were
attained on hb2-AR. By variation of the substitution pattern at
the aromatic ring, selectivity over hb1-AR seems to be tunable,
with the para-fluoro derivative 46 displaying good selectivity fac-
tors of 105 over hb1-Ar and 830 over hb3-AR (based on binding
data). For this compound 46 and derivative 38, functional data
was acquired (Table 6): both displayed a moderate partial agonistic
activity at hb1-AR, a mediocre full agonistic effect on hb3-AR and
good antagonistic activity at hb2-AR, with compound 46 giving rise
to an IC50 of 32 nM. For the ‘best’ hb3-AR ligand out of this series,
compound 40, agonistic activity on hb3-AR was rather poor with
When removing some bulk from the naphthalimide portion,
which is known to potentially cause DNA intercalation26 and was
thus sought to be avoided, affinity for hb3-AR was further de-
creased for phthalimides to glutarimides within both, the allyl
and isopropyl series (compounds 5 vs 30 vs 32 and 16 vs 31 vs
33; Table 4). For both of these series, a clearly optimized hb2-AR
selectivity was achieved: for allyl derivatives, affinity for hb1-AR
decreased as well from naphthalimides over phthalimides to the
glutarimides accompanied by a nearly constant affinity for hb2-
AR; for isopropyl compounds, the affinity for hb1-AR remained
within the same range along with an increase of affinity for hb2-
AR. Leaving out any imide array and thus rather mimicking the
Alprenolol scaffold (compounds 34 and especially 36), ligands with
good affinities for hb2-AR were attained with Ki values of 7 and
2.4 nM, respectively—and even with good selectivity by a factor
of 110 over hb1- and above 2000 over hb3-AR (compound 36).
The corresponding isopropyl derivatives 35 and 37 possessed an
increased affinity for hb1- and hb3-AR as compared to the 2-allyl
variants 34 and 36, thus resulting in an overall decreased selectiv-
ity for hb2-AR. However, an even better Ki of 1.2 nM was detected
for compound 35 at the latter receptor.
All variations described so far—at the aryloxy unit, at the spacer,
and fragmentations of the naphthalimide unit—were leading to an
optimization of selectivity and affinity for hb2-AR. Therefore, an at-
tempt was made to gain some affinity for hb3-AR by performing a
Comparative Molecular Similarity Indices Analysis (CoMSIA)27
based on those compounds elaborated around Table 2 combined
with several thienopyrimidine derivatives around hit molecule 6,
which proved to be b3-AR selective.22 A resulting CoMSIA sketch
(Fig. 2) suggested the addition of a hydrogen bond acceptor to
the 5-position of the naphthalimide to enhance ligand affinity,
which was realized by incorporating a hydroxy or methoxy group
(R0 in Table 4). Such derivatizations, however, were not able to gen-
erate compounds with (significantly) increased affinity for hb3-AR.
Instead, hb2-AR selectivity was optimized for compounds 28 and
29, bearing a methoxy group (Table 4), as compared to the parent
molecules 5 and 16.
an EC50 = 2.8 lM (IA = 0.9; data not shown).
With a beneficial effect of hb2-AR antagonists in wound healing
having been described recently,28 compound 4629 displayed an
intriguing activity profile. Consequently, this substrate was sub-
jected to further biological and physicochemical evaluation. It
proved to be completely stable in artificial gastric juice, simulated
intestinal fluids and human plasma within a test period of 6 h. Sol-
ubility was determined to be above 180
fered at pH 4 and 10 M at pH 6, membrane permeability in a
PAMPA was high. Toxicity in a PBMC viability assay was found at
a tolerable level with an ED50 = 9 M, and in a HepG2 assay with
an ED50 = 19 M. A preliminary PK study in male Wistar rats was
lM in aq. medium buf-
l
l
l
performed by po administration of the compound (10 mg/kg),
resulting in a Cmax of 290 nM and a halflife t1/2 of 7.5 h.
The vHTS approach as realized herein enabled the identification
of potent b-adrenoceptor ligands. Subtype selectivity or a well de-
fined cellular functional activity, however, was not necessarily at-
tained into the direction of the ligands used for the pharmacophore
alignment: thienopyrimidines like 6 (Fig. 1) showed a certain
selectivity for hb3-AR based on binding data and a partial agonistic
effect on this receptor in a cellular assay22—as originally envisaged
when template ligands were selected—naphthalimides 4 and 5 dis-
played highest affinity for hb2-AR. For compound 4, this affinity
was accompanied by a certain selectivity combined with an antag-
onistic effect on all three b-AR subtypes. Derivative 5 showed sim-
ilar affinity for hb1- and hb2-AR with a weak agonistic effect on
hb3-AR. As mentioned above, a pharmacophore alignment ap-
proach can not necessarily be expected to yield a more defined re-
sult with regard to selectivity and functional activity, which was
clearly reflected in this study. Taken from recent literature, a dock-
ing approach based on the crystal structure of the inactive state of
the hb2-AR might not result in ligands with a clearly defined func-
tional activity either.30 Thus, the simplicity of the pharmacophore
alignment in combination with the good affinities attained for the
Finally, the carbazolyl derivative 21 was selected for a set of
derivatizations at the naphthalimide site (Table 5), as it was the
only compound identified so far to display significant affinity for
hb3-AR—even though still topped by those for hb1- and hb2-AR.
Again, as already observed for the 2-allyl- and 2-isopropyl-
phenoxy series (Table 4), substitution at C5 of the naphthalimide
in order to add a hydrogen bond acceptor was not providing a shift
towards hb3-AR-selective compounds. Instead, a ligand with an
Table 6
Binding and functional data for selected b-AR ligands
Compd
Binding assay, Ki [nM]
Functional assay, EC50/IC50 [nM] (IA)a
hb1-
hb2-
hb3-
hb1-AR
hb2-AR
hb3-AR
AR
AR
AR
4
38
1000
250
80
37
1630
590
IC50 = 1375
EC50 = 80
(0.4)
IC50 = 377 IC50 = 1474
IC50 = 120 EC50 = 234
(1.0)
46
32
0.3
250
EC50 = 310
(0.2)
IC50 = 32
EC50 = 385
(1.1)
Figure 2. CoMSIA illustration around compound 21: red area indicates a hydrogen
bond acceptor to be desirable within ligand, yellow area indicates a hydrogen bond
donor to be desirable within ligand.
a
EC50 indicates agonistic, IC50 indicates antagonistic effects; IA = intrinsic activ-
ity, relative to isoproterenol (IA = 1.0).

3404
S. Tasler et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3399–3404
membrane preparations of HEK-293 cells overexpressing the human b3-AR
(Bmax = 550 fmol/mg protein) using 0.5 nM [125I]-cyanopindolol as radioligand.
19. All aryloxypropanolamine derivatives were tested as racemates.
20. Membrane preparations (CHO-K1 cell line) expressing human b1-, b2- or b3-
ARs (Bmax = 3.78, 1.68 and 47.2 pmol/mg protein, respectively), were purchased
from Euroscreen (now Perkin Elmer). Binding assays were performed according
to the manufacturer’s instructions. The radioligand for all three receptor
subtypes was [125I]-cyanopindolol (125I-CYP) (Amersham) (final concentration
of 0.05, 0.05 and 1.5 nM, respectively). Ki values were calculated using the
Cheng–Prusoff equation on IC50 determinations, which were based on
concentration curves using eight concentrations (half-logarithmic) in
duplicate.
21. Functional response of cells (agonistic or antagonistic) to the test compounds
was tested by measurement of cyclic AMP formation by HTRFÒ (Homogeneous
Time-Resolved Fluorescence) technology (Cisbio International) using a stable
cell line CHO-K1 expressing the human recombinant b3-AR (Euroscreen, now
Perkin Elmer) according to the manufacturer’s instructions. EC50 (agonists) and
IC50 values (antagonists) were determined by dose–response curves based on
eight concentrations (logarithmic) determined in quadruplicate in a 96 half-
original screening hits renders this approach highly valuable. Start-
ing from the naphthalimide-substituted aryloxypropanolamine
scaffold of the original hit molecules 4 and 5, all medicinal chem-
istry endeavours resulted in a stronger emphasis of affinity for
hb2-AR and optimization of selectivity towards the latter AR sub-
type. Such efforts led to the identification of the benzoylpiperidine
derivative 46, displaying excellent affinity for hb2-AR in the picom-
olar range and an antagonistic activity at hb2-AR. Its favourable pi-
lot profile of physicochemical and pharmacokinetic parameters
suggests further optimization towards application of a hb2-AR
antagonist in for example, wound healing.
Acknowledgments
The authors want to thank Oliver Müller and Marcel Kirschstein
for synthetic support.
well plate in a final volume of 100
ll. The antagonistic effect was determined
by preincubation with test compound for 10 min followed by agonist
a
stimulation (0.05 nM isoproterenol) for 30 min.
Functional data on hb1-AR and hb2-AR were determined at Euroscreen (both,
in an agonist and an antagonist assay). EC50 and IC50 values, respectively, were
determined by dose response curves based on eight concentrations
(logarithmic) determined in duplicate.
References and notes
1. (a) Arch, J. R. S.; Ainsworth, A. T.; Cawthorne, M. A.; Piercy, V.; Sennitt, M. V.;
Thody, V. E.; Wilson, C.; Wilson, S. Nature 1984, 309, 163; (b) Emorine, L. J.;
Marullo, S.; Briend-Sutren, M. M.; Patey, G.; Tate, K.; Delavier-Klutchko, C.;
Strosberg, A. D. Science 1989, 245, 1118.
2. Hieble, J. P. Curr. Top. Med. Chem. 2007, 7, 207.
3. Dow, R. L. Exp. Opin. Invest. Drugs 1997, 6, 1811.
22. Tasler, S.; Baumgartner, R.; Ammendola, A.; Wieber, T.; Schachtner, J.; Blisse,
M.; Rath, S.; Michel, S.; Lang, M.; Zaja, M.; Quotschalla, U.; Ney, P. Bioorg. Med.
Chem. Lett., in preparation.
23. In analogy to e.g.: (a) Wagner, S.; Kopka, K.; Law, M. P.; Riemann, B.; Pike, V.
W.; Schober, O.; Schäfers, M. Bioorg. Med. Chem. 2004, 12, 4117; (b) Elzein, E.;
Shenk, K.; Ibrahim, P.; Marquart, T.; Kerwar, S.; Meyer, S.; Ahmed, H.; Zeng, D.;
Chu, N.; Soohoo, D.; Wong, S.; Leung, K.; Zablocki, J. Bioorg. Med. Chem. Lett.
2004, 14, 973.
24. In analogy to, e.g.: (a) Yang, W.; Wang, Y.; Roberge, J. Y.; Ma, Z.; Liu, Y.;
Lawrence, R. M.; Rotella, D. P.; Seethala, R.; Feyen, J. H. M.; Dickson, J. K., Jr.
Bioorg. Med. Chem. Lett. 2005, 15, 1225; (b) Howe, R.; Rao, B. S.; Holloway, B. R.;
Stribling, D. J. Med. Chem. 1992, 35, 1751.
25. (a) Hein, L. Pharm. Unserer Zeit 2004, 33, 434; (b) Griffin, P. P.; Schubert-
Zsilavecz, M.; Stark, H. Pharm. Unserer Zeit 2004, 33, 442.
4. Weber, A. E. Ann. Rep. Med. Chem. 1998, 33, 193.
5. Baker, J. G. Br. J. Pharmacol. 2005, 144, 317.
6. Kolb, P.; Rosenbaum, D. M.; Irwin, J. J.; Fung, J. J.; Kobilka, B. K.; Shoichet, B. K.
Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 6843.
7. Rouget, C.; Breuiller-Fouche, M.; Mercier, F. J.; Leroy, M. J.; Loustalot, C.; Naline,
E.; Frydman, R.; Croci, T.; Morcillo, E. J.; Advenier, C.; Bardou, M. Br. J.
Pharmacol. 2004, 141, 831.
8. (a) Inoue, Y.; Yoshizato, T.; Kawarabayashi, T. J. Obstet. Gynaecol. Res. 2009, 35,
405; (b) Doggrell, S. A. Expert Opin. Pharmacother. 2004, 5, 1917.
9. Sawa, M.; Harada, H. Curr. Med. Chem. 2006, 13, 25.
10. Ursino, M. G.; Vasina, V.; Raschi, E.; Crema, F.; De Ponti, F. Pharmacol. Res. 2009,
59, 221.
26. For a few examples, cf. e.g.: Brana, M. F.; Gradillas, A.; Gomez, A.; Acero, N.;
Llinares, F.; Munoz-Mingarro, D.; Abradelo, C.; Rey-Stolle, F.; Yuste, M.;
Campos, J.; Gallo, M. A.; Espinosa, A. J. Med. Chem. 2004, 47, 2236.
27. For a CoMSIA (Klebe, G.; Abraham, U.; Mietzner, T. J. Med. Chem. 1994, 37,
11. www.clinicaltrials.gov, Feb 2010.
12. (a) Palczewski, K.; Kumasaka, T.; Hori, T.; Behnke, C. A.; Motoshima, H.; Fox, B.
A.; Le Trong, I.; Teller, D. C.; Okada, T.; Stenkamp, R. E.; Yamamoto, M.; Miyano,
M. Science 2000, 289, 739; (b) Okada, T.; Sugihara, M.; Bondar, A.-N.; Elstner,
M.; Entel, P.; Buss, V. J. Mol. Biol. 2004, 342, 571.
4130),
two
enantiomeric
data
sets
of
41
3D-structures
of
aryloxypropanolamine compounds with affinity for hb3-AR were aligned on
a template thienopyrimidine derivative22 using 4SC’s proprietary software
4SCanÒ 17
3D-QSAR models using the five CoMSIA descriptors (steric,
.
13. (a) Cherezov, V.; Rosenbaum, D. M.; Hanson, M. A.; Rasmussen, S. G. F.; Thian, F.
S.; Kobilka, T. S.; Choi, H.-J.; Kuhn, P.; Weis, W. I.; Kobilka, B. K.; Stevens, R. C.
Science 2007, 318, 1258; (b) Rosenbaum, D. M.; Cherezov, V.; Hanson, M. A.;
Rasmussen, S. G. F.; Thian, F. S.; Kobilka, T. S.; Choi, H.-J.; Yao, X.-J.; Weis, W. I.;
Stevens, R. C.; Kobilka, B. K. Science 2007, 318, 1266; (c) Rasmussen, S. G. F.;
Choi, H.-J.; Rosenbaum, D. M.; Kobilka, T. S.; Thian, F. S.; Edwards, P. C.;
Burghammer, M.; Ratnala, V. R. P.; Sanishvili, R.; Fischetti, R. F.; Schertler, G. F.
X.; Weis, W. I.; Kobilka, B. K. Nature 2007, 450, 383.
14. Warne, T.; Serrano-Vega, M. J.; Baker, J. G.; Moukhametzianov, R.; Edwards, P.
C.; Henderson, R.; Leslie, A. G. W.; Tate, C. G.; Schertler, G. F. X. Nature 2008,
454, 486.
15. (a) Tanaka, N.; Tamai, T.; Mukaiyama, H.; Hirabayashi, A.; Muranaka, H.;
Ishikawa, T.; Kobayashi, J.; Akahane, S.; Akahane, M. J. Med. Chem. 2003, 46,
105; (b) Yanagisawa, T.; Sato, T.; Yamada, H.; Sukegawa, J.; Nunoki, K. Tohoku J.
Exp. Med. 2000, 192, 181; (c) Hu, B.; Jennings, L. L. Prog. Med. Chem. 2003, 41,
167.
electrostatic, hydrophobic, H-donor, H-acceptor) were calculated with SYBYL
(version 7.0; Tripos Associates: St. Louis, MO, USA, 2005; http://
www.tripos.com/). Only the data set with S-configuration at the
propanolamine moiety resulted in correlation between predicted and
measured pKi values (q2 = 0.55, cross-validated RMSE = 0.55, 6 components),
predicting by itself the biologically active stereoisomer, and was used for
further predictions.
28. (a) Pullar, C. E.; Rizzo, A.; Isseroff, R. R. J. Biol. Chem. 2006, 281, 21225; (b) Pullar,
C. E.; Grahn, J. C.; Liu, W.; Isseroff, R. R. FASEB J. 2006, 20, 76; (c) Sivamani, R. K.;
Pullar, C. E.; Manabat-Hidalgo, C. G.; Rocke, D. M.; Carlsen, R. C.; Greenhalgh, D.
G.; Isseroff, R. R. PLoS Med. 2009, 6, 105.
29. NMR data (ppm) for compound 46: 1H NMR (300 MHz, DMSO-d6): d = 1.60 (m,
2H), 1.74 (m, 2H), 2.24 (m, 2H), 2.54 (dd, J = 12.5, 5.7 Hz, 1H), 2.67 (dd, J = 12.5,
4.6 Hz, 1H), 3.00 (mC, 2H), 3.38 (mC, 1H), 4.17 (mC, 3H), 4.92 (d, J = 2.6 Hz, 1H,
OH), 6.69 (d, J = 8.0 Hz, 1H), 7.06 (d, J = 8.0 Hz, 1H), 7.14 (dd, J = 7.8, 6.9 Hz, 1H),
7.29 (t, J = 8.0 Hz, 1H), 7.33 (dd, J = 8.0, 6.9 Hz, 1H), 7.34 (t, J = 8.9 Hz, 2H), 7.44
(d, J = 8.0 Hz, 1H), 8.05 (dd, J = 8.9, 5.6 Hz, 2H), 8.25 (d, J = 7.8 Hz, 1H), 11.20 (s,
1H, NH); 13C NMR (75.5 MHz, DMSO-d6): d = 28.45, 42.56, 53.29, 61.51, 66.83,
70.79, 100.4, 103.7, 110.3, 111.6 (Cq), 115.7 (d, 2JCF = 21.9 Hz), 118.4, 121.7 (Cq),
16. Zheng, W.; Nikulin, V. I.; Konkar, A. A.; Vansal, S. S.; Shams, G.; Feller, D. R.;
Miller, D. D. J. Med. Chem. 1999, 42, 2287.
17. Seifert, M. H. J.; Wolf, K.; Vitt, D. Biosilico 2003, 1, 143.
18. Initial screening was performed at MDS, Taiwan, in a radioligand receptor
3
4
122.4, 124.4, 126.4, 131.1 (d, JCF = 9.4 Hz), 132.3 (d, JCF = 3.0 Hz, Cq), 138.9
1
binding assay at
a compound concentration of 10 lM (in duplicate). For
(Cq), 141.1 (Cq), 155.0 (Cq), 164.9 (d, JCF = 252 Hz, Cq), 201.1 (C@O).
potential hits (radioligand replacement >20%), a semi-quantitative IC50 was
determined (six concentrations in duplicate). The assay was based on
30. de Graaf, C.; Rognan, D. J. Med. Chem. 2008, 51, 4978.