Discovery of novel 1-arylmethyl pyrrolidin-2-yl ethanol amines as calcium-sensing receptor antagonists

Article abstract of DOI:10.1016/j.bmcl.2005.08.095

A 3D quantitative structure-activity relationship study for inhibition of calcium-sensing receptor in the aryloxypropanolamine series predicted that these molecules adopt a U-shaped conformation with pi-stacking between the two aromatic rings. This hypoth

Full text of DOI:10.1016/j.bmcl.2005.08.095

Bioorganic & Medicinal Chemistry Letters 15 (2005) 5478–5482
Discovery of novel 1-arylmethyl pyrrolidin-2-yl ethanol amines
as calcium-sensing receptor antagonists
Ashvinikumar V. Gavai,* Roy J. Vaz, Amarendra B. Mikkilineni, Jacques Y. Roberge,
Yalei Liu, R. Michael Lawrence, James R. Corte, Wu Yang, Mark Bednarz,
John K. Dickson, Jr., Zhengping Ma, Ramakrishna Seethala and Jean H. M. Feyen
Bristol-Myers Squibb Pharmaceutical Research Institute, PO Box 4000, Princeton, NJ 08543-4000, USA
Received 29 July 2005; revised 27 August 2005; accepted 29 August 2005
Available online 10 October 2005
Abstract—A 3D quantitative structure–activity relationship study for inhibition of calcium-sensing receptor in the aryloxypropanol-
amine series predicted that these molecules adopt a U-shaped conformation with pi-stacking between the two aromatic rings. This
hypothesis led to the discovery of novel 1-arylmethyl pyrrolidin-2-yl ethanol amines capable of antagonizing the calcium-sensing
receptor with potency comparable to that of NPS-2143.
ꢀ 2005 Elsevier Ltd. All rights reserved.
Osteoporosis is a serious and prevalent disease associat-
ed with significant morbidity and mortality. In osteopo-
rotic individuals, there is an imbalance between bone
resorption and bone formation resulting in a progressive
loss of bone mass and structure. Current therapies in-
volve anti-resorptive agents that impede further loss of
bone mass once the disease has manifested and can elicit
a modest increase in bone mineral density; however,
bone which is lost is not restored.2 Teriparatide, a syn-
thetic 1–34 amino acid peptide fragment of human para-
thyroid hormone (PTH), is the only FDA-approved
anabolic agent that, upon daily subcutaneous injection,
has been shown to increase bone mineral density and re-
duce fracture rates in humans.2 In contrast to the ana-
bolic effects observed after intermittent administration,
it is well documented that continuous exposure to
PTH results in increases in bone turnover with a subse-
quent loss in bone mass. As a viable alternative to sub-
cutaneous PTH therapy, one can envision stimulating
the release of endogenous PTH from the parathyroid
glands in order to stimulate bone formation. It is well
established that PTH secretion is inversely related to
the plasma calcium concentration, which is mediated
by the calcium-sensing receptor (CaR) expressed on
the surface of parathyroid cells.3 CaR, a member of
the metabotropic glutamate-like G-protein coupled
receptor family, is coupled through changes in phospho-
inositide (PI) turnover to the release of calcium from
intracellular stores. Preclinical proof-of-principle for
the CaR approach includes studies with the CaR antag-
onist (calcilytic), NPS-2143.4 Intravenous, bolus injec-
tion of NPS-2143 to osteopenic ovariectomized rats
resulted in a transient increase in serum PTH compatible
with the anabolic profile of the hormone. Oral adminis-
tration of NPS-2143 resulted in a sustained increase in
plasma PTH and a concurrent increase in bone turnover
without any change in bone mass. The absence of an in-
crease in bone mass is consistent with the observation of
the sustained levels of PTH. Indeed, an increase in bone
mass was observed in the presence of an anti-resorptive
agent due to the uncoupling of bone resorption from the
anabolic component. These results indicate that a calci-
lytic agent with rapid absorption and a short half-life
should evoke the desired transient increase in plasma
PTH levels, translating into a bone anabolic response.
The recently disclosed calcilytic compound NPS-2143
(1) had an IC50 of 0.05 lM in a fluorescence-based intra-
cellular calcium mobilization assay in human TT cells
containing the endogenous calcium-sensing receptor.5
The results from a brief SAR study around NPS-2143
are summarized in Table 1.6 Removal of the hydroxyl
group (4) led to greatly reduced activity; however, the
hydroxyl enantiomers 1 and 2 exhibited similar levels
of potency. N-methylation of the secondary amine (3)
Keywords: Osteoporosis; Calcium-sensing receptor antagonist.
*
Corresponding author. Tel.: +1 609 252 5091; fax: +1 609 252
6601; e-mail: ashvinikumar.gavai@bms.com
Ref. 1.
0960-894X/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.08.095

A. V. Gavai et al. / Bioorg. Med. Chem. Lett. 15 (2005) 5478–5482
5479
Table 1. Inhibition of intracellular Ca2+ release by aryloxypropanolamines in human TT cells as measured by FLIPRTM
a
Compound
Structure
CaR IC50 (lM)
CN
Cl
OH
OH
OH
H
N
O
O
O
O
1 (NPS-2143)
0.05
CN
Cl
H
N
2
3
4
5
6
7
8
9
0.05
0.27
0.88
0.03
3.86
4.60
0.06
0.10
0.60
6.18
3.30
9.43
11.36
1.39
CN
Cl
N
CN
Cl
H
N
CN
O
OH
H
N
CN
Cl
OH
OH
H
N
O
O
CN
Cl
H
N
CN
O
OH
H
N
S
CN
O
OH
H
N
HN
CN
Cl
OH
OH
H
N
O
10
F
F
Cl
Cl
H
N
S
11
12
13
14
OH
H
O
N
F3CO
F
SO2Me
O
OH
H
N
OH
H
O
N
MeO2S
MeO
OMe
OH
H
O
N
15
OMe
a Values are means of two experiments.
diminished the potency by 5-fold, while deletion of the
gem-dimethyl (6) resulted in a dramatic loss of potency.
Replacement of naphthalene with other fused bicyclic
aromatic rings (8 and 9) was tolerated with a modest
loss of activity. This information was used to develop
a comparative molecular similarity index analysis
(CoMSiA) model for three-dimensional quantitative
structure–activity relationship in this series. The
CoMSiA model is a well-established method for gener-
ating bioactive conformation hypotheses.7 A conforma-
tional search of NPS-2143 around the bonds labeled 1
and 2 provided three lowest energy conformations as

5480
A. V. Gavai et al. / Bioorg. Med. Chem. Lett. 15 (2005) 5478–5482
by the data shown in Table 1 (5 and 6), the linker con-
taining a gem-dimethyl moiety is also likely to play a
pivotal role in the adoptation of this conformation.
One can envisage a further improvement in potency in
this series by restricting some degrees of freedom to in-
crease the propensity of the molecule to exist in the bio-
active conformation. Out of the various possibilities
considered to constrain the degrees of freedom, the
1-arylmethyl pyrrolodin-2-yl ethanolamines allowed
the resulting compound to generate the desired
U-shaped conformation with pi-stacking between the
two aromatic moieties. As depicted in Figure 3, the
CoMSiA model predicted that R-configuration is essen-
tial at the asymmetric center on the pyrrolidine ring for
the compound to adopt a bioactive U-shaped conforma-
tion similar to NPS-2143. Indeed, all the S-enantiomers
at the pyrrolidine prepared to date have been devoid of
CaR inhibitory activity (data not shown). Further
support for the CoMSiA model is rendered by the lack
of potency of compound 22 (Table 2), which does not
prefer the U-shaped conformation (Fig. 3) due to the
torsion angle requirements imposed by the amide group.
Figure 1. Conformations of NPS-2143 used to construct 3 of the 4
models used with CoMSiA studies. The other conformation used was
very similar to that of Model 1 with a H-bond between the hydroxyl
hydrogen and the oxygen atom attached to the phenyl ring.
Standard Error of Prediction for various
conformations
1.15
1.1
1.05
model 1
1
0.95
model 2
The initial route utilized for the synthesis of these com-
pounds is illustrated in Scheme 1. Selective N-alkylation
of D-proline 16 was followed by conversion of the carbox-
ylic acid to the corresponding chloromethyl ketone 17 via
a three-step sequence through a mixed anhydride and
diazoketone intermediates.8 The diastereomeric epoxides
18, obtained by reduction of the ketone and subsequent
intramolecular cyclization, were separated by silica gel
chromatography and coupled with amines 196 to provide
the corresponding ethanol amines 20. The diastereomeric
purity of the final compounds was determined to be
>98% by analytical HPLC.9 A solution-phase parallel
synthesis was developed for the preparation of subse-
quent analogs (Scheme 2). Carbobenzyloxy-D-proline
23 was converted to a mixture of diastereomeric epoxides
24 that were separated by column chromatography, cou-
pled with the amines 19, and the protecting group was re-
moved to give the pyrrolidines 25. Reductive amination
of pyrrolidines 25 with aromatic aldehydes in the pres-
ence of zinc chloride or titanium isopropoxide provided
the desired ethanol amines 20.10
model 3
model 4
0.9
0.85
0.8
0.75
1
2
3
4
component
Figure 2. The standard error of prediction for the four CoMSiA
models plotted versus the number of components. The lowest standard
error of prediction was associated with Model 1.
shown in Figure 1. In addition, another conformation
with an internal H-bond between the hydroxyl hydrogen
and the oxygen atom attached to the phenyl ring (Model
2) was also evaluated. Each model was subjected to a
leave-one-out cross-validation and the corresponding
standard error of prediction is plotted versus the compo-
nent number for the four models (Fig. 2). The CoMSiA
studies predicted Model 1 to be the bioactive conforma-
tion based on the lowest standard error of prediction.
Perhaps, these molecules adopt such a U-shaped
bioactive conformation due to an energetically favorable
pi-stacking between the aromatic groups. As supported
A summary of the SAR around the N-benzylpyrroli-
dine of compound 21 is presented in Table 2. Com-
Figure 3. Compounds with R-configuration at the pyrrolidine ring (e.g., 21d on the right) can exhibit conformations similar to that of NPS-2143
(center structure). Compound 22 on the left is predicted to be inactive.

A. V. Gavai et al. / Bioorg. Med. Chem. Lett. 15 (2005) 5478–5482
5481
Table 2. Inhibition of intracellular Ca2+ release by 1-arylmethyl
pyrrolidin-2-yl ethanolamines in human TT cells as measured by
FLIPRTM
pounds in this series exhibited a modest preference for
the R-configuration at the hydroxyl asymmetric center
(e.g., 21a and 21b). An electron-withdrawing substitu-
tion was preferred at the meta- or ortho-position of
the phenyl ring. For example, a meta-cyano (21c) or ni-
tro (21f) substituent provided an order of magnitude
increase in in potency over the corresponding m-meth-
oxy substituent (21l). Moving the substitution to the
para-position resulted in up to a 5-fold loss of activity
(21f and 21g). Increasing the size of this group resulted
in a concomitant reduction in CaR inhibitory potency
(21j and 21i). Polar functional groups, such as a car-
boxylic acid (21n), proved unfavorable for antagonist
activity. Preliminary exploration of 2,3-disubstitution
failed to provide any clear trends. Whereas the
2-hydroxy-3-fluoro substitution (21p) augmented the
potency 5-fold relative to either substitution alone,
introduction of a 2-hydroxyl in 3-methoxyphenyl or
3-nitrophenyl compounds (21q and 21r) did not
provide any increase in potency.
R
R1
R2
R3
N
X
N
H
HO
21: X = CH
2
22: X = CO
a
Compound
R1
R2
R3
CaR IC50 (lM)
21a(S)
21b(R)
21c(R)
21d(R)
21e(R)
21f(R)
21g(R)
21h(R)
21i(R)
21j(R)
21k(R)
21l(R)
21m(R)
21n(R)
21o(R)
21p(R)
21q(R)
21r(R)
21s(R)
22(R)
CN
CN
H
H
H
1.27
0.34
0.19
0.25
0.16
0.14
0.79
0.97
0.75
1.57
2.47
1.19
1.16
6.84
0.36
0.14
1.56
0.14
0.24
8.43
H
H
CN
H
H
H
CN
H
NO2
H
H
NO2
H
H
H
NO2
H
OH
H
H
F
Br
H
H
H
H
Ph
H
H
OMe
OH
CO2H
F
H
H
H
A brief SAR study was carried out focusing on the naph-
thalene ring replacements of compound 21c. Table 3 sum-
marizes the antagonist activities of exemplary analogues
prepared in this series. Similar to the observations made
in the aryloxypropanolamine series (Table 1), replace-
ment of the naphthalene moiety with a 2-substituted ben-
zothiophene or indole (26a and 26b) resulted in a modest
loss of activity. The decrease in CaR inhibitory potency
was more dramatic for 3-substituted benzothiophene
(26c) and 4-methoxyphenyl (26f) replacements of the
naphthalene functionality. The lower potency of com-
pound 26e suggests that a H-bond accepting nitrogen is
not tolerated in the potentially hydrophobic pocket occu-
H
H
F
H
OH
OH
OH
CN
H
F
H
OMe
NO2
Cl
H
H
H
CN
H
a Values are means of two experiments.
a, b, c
d, e
CH2Cl
*
CO2H
N
N
H
O
Ar1
Ar2
17
16
Table 3. Inhibition of intracellular Ca2+ release by 1-arylmethyl
pyrrolidin-2-yl ethanolamines in human TT cells as measured by
FLIPRTM
H2N
Ar2
19
*
*
N
H
N
N
O
OH
f
Ar1
Ar1
Ar2
R
18
20
NC
N
N
H
Scheme 1. Reagents and conditions: (a) Ar1CH2Br, KOH, iPrOH,
40 ꢁC; (b) i—iBuCOCl, Et3N, THF, À15 ꢁC to rt; ii—CH2N2, Et2O,
0 ꢁC; (c) HCl, dioxane, 0 ꢁC; (d) NaBH4, THF, MeOH, 0 ꢁC; (e) KOH,
EtOH, rt; (f) EtOH, 110 ꢁC.
OH
a
Compound
Ar2
CaR IC50 (lM)
26a
0.27
0.44
1.00
1.35
2.84
2.67
S
Ar2
H2N
26b
26c
26d
26e
26f
a, b, c
19
*
N
H
CO2H
N
N
d, e
O
Cbz
Cbz
24
23
S
N
S
N
Ar2
Ar2
*
*
f
N
H
N
N
N
H
H
OH
OH
Ar1
25
20
N
H
Scheme 2. Reagents and conditions: (a) i—iBuCOCl, Et3N, THF,
À15 ꢁC to rt; ii—CH2N2, Et2O, 0 ꢁC; iii—HCl, dioxane, 0 ꢁC; (b)
NaBH4, THF, MeOH, 0 ꢁC; (c) KOH, EtOH, rt; (d) EtOH, 110 ꢁC; (e)
H2, Pd/C, EtOH; (f) i—Ar1CHO, ZnCl2 or Ti(OiPr)4, MeOH/THF;
ii—NaBH3CN, THF.
OMe
a Values are means of two experiments.

5482
A. V. Gavai et al. / Bioorg. Med. Chem. Lett. 15 (2005) 5478–5482
pied by the bicyclic aromatic ring on the surface of the cal-
cium-sensing receptor.
References and notes
1. Present address: Sanofi-Aventis Pharmaceuticals, 1041
Route 202/206 North, Bridgewater, NJ 08807.
2. For recent reviews, see: (a) Doggrell, S. A. Drugs Today
2003, 39, 633; (b) Rotella, D. P. Curr. Opin. Drug Discov.
Dev. 2002, 5, 477; (c) Gowen, M.; Emery, J. G.; Kumar, S.
Emerg. Drugs 2000, 5, 1.
3. (a) Nemeth, E. F. Curr. Pharm. Des. 2002, 8, 2077; (b)
Nemeth, E. F. Mol. Endocrinol. 2002, 29, 15.
4. Gowen, M.; Stroup, G. B.; Dodds, R. A.; James, I. E.;
Votta, B. J.; Smith, B. R.; Bhatnagar, P. K.; Lago, A. M.;
Callahan, J. F.; DelMar, E. G.; Miller, M. A.; Nemeth, E.
F.; Fox, J. J. Clin. Invest. 2000, 105, 1595.
Compounds 21c and 21p have emerged as the lead candi-
dates from this series with CaR antagonist potency that is
3- to 4-fold less than that of NPS-2143. These compounds
were evaluated in a functional assay for inhibition of
extracellular Ca2+-induced inositol triphosphate (IP) gen-
eration in rat MTC cells that constitutively express the
calcium-sensing receptor.11 Compounds 21c and 21p
inhibited Ca2+-induced IP accumulation with respective
IC50 values of 2.20 and 2.56 lM, relative to NPS-2143
which exhibited an IC50 of 0.53 lM in this assay.
5. Inhibition by test compounds of intracellular calcium
release stimulated by extracellular calcium was measured
using FLIPR technology as previously described: Arey, B.
J.; Seethala, R.; Ma, Z.; Fura, A.; Morin, J.; Swartz, J.;
Vyas, V.; Yang, W.; Dickson, J. K.; Feyen, J. H. M.
Endocrinology 2005, 146, 2015.
6. These compounds were synthesized by procedures report-
ed in the literature, e.g., Doggrell, S. A.; del Fresno, M.;
Castaner, J. Drugs Future 2002, 27, 140.
In conclusion, a three-dimensional quantitative struc-
ture–activity relationship study for inhibition of calci-
um-sensing receptor in the aryloxypropanolamine series
predicted that these molecules adopt a U-shaped confor-
mation with pi-stacking between the two aromatic rings
and this hypothesis led to the discovery of novel 1-arylm-
ethyl pyrrolidin-2-yl ethanol amines capable of antago-
nizing the human and rat calcium-sensing receptors with
potencycomparable to that of NPS-2143. These data offer
a new avenue for the development of novel anabolic
agents for the treatment of osteoporosis.
7. Klebe, G.; Abraham, U.; Mietzner, T. J. Med. Chem.
1994, 37, 4130.
8. Albeck, A.; Persky, R. Tetrahedron 1994, 50, 6333.
9. Structural assignments were supported by mass spectral
data and H NMR obtained on purified samples. HPLC:
1
YMC S5 ODS 4.6 · 50 mm; flow rate 4 mL/min; 220 nm;
10–90% aqueous methanol with 0.2% H3PO4.
10. Kim, S.; Oh, C. H.; Ko, J. S.; Ahn, K. H.; Kim, Y. J.
J. Org. Chem. 1985, 50, 1927.
Acknowledgments
We acknowledge the members of the Bristol-Myers
Squibb Analytical Chemistry Department for analytical
and spectral determinations.
11. An assay for the inhibition by test compounds of
extracellular Ca2+-induced inositol phosphate generation
in cells expressing the CaR has been described (see Ref. 5).