Discovery of the oxazabicyclo[3.3.1]nonane derivatives as potent and orally active GPR119 agonists

Article abstract of DOI:10.1016/j.bmcl.2015.09.047

The design and synthesis of two conformationally restricted oxazabicyclo octane derivatives as GRP119 agonists is described. Derivatives of scaffold C, with syn configuration, have the best overall profiles with respect to solubility and in vivo efficacy. Compound 25a was found to have extremely potent agonistic activity and was orally active in lowering blood glucose levels in a mouse oral glucose tolerance test at a dose of 0.1 mg/kg.

Full text of DOI:10.1016/j.bmcl.2015.09.047

Bioorganic & Medicinal Chemistry Letters 25 (2015) 5291–5294
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of the oxazabicyclo[3.3.1]nonane derivatives as potent and
orally active GPR119 agonists
a
a
a
a
b
Xing Dai a, , Andrew Stamford , Hong Liu , Bernard Neustadt , Jingsong Hao , Tim Kowalski ,
Brian Hawes b, Xiaoying Xu c, Hana Baker b, Kim O’Neill b, Morgan Woods b, Huadong Tang c
William Greenlee a
⇑
a Department of Medicinal Chemistry, Merck Research Laboratory, 2015 Galloping Hill Road, Kenilworth, NJ 07033, United States
b Department of Biology, Merck Research Laboratory, 2015 Galloping Hill Road, Kenilworth, NJ 07033, United States
c Department of Drug Metabolism, Merck Research Laboratory, 2015 Galloping Hill Road, Kenilworth, NJ 07033, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
The design and synthesis of two conformationally restricted oxazabicyclo octane derivatives as GRP119
agonists is described. Derivatives of scaffold C, with syn configuration, have the best overall profiles with
respect to solubility and in vivo efficacy. Compound 25a was found to have extremely potent agonistic
activity and was orally active in lowering blood glucose levels in a mouse oral glucose tolerance test at
a dose of 0.1 mg/kg.
Received 12 August 2015
Accepted 18 September 2015
Available online 21 September 2015
Keywords:
GPR119 agonist
Type 2 diabetes
Bridged piperidine
Glucose-dependent insulin secretion
Ó 2015 Elsevier Ltd. All rights reserved.
GPR119 is a G-protein-coupled receptor expressed in pancreatic
b-cells and intestinal enteroendocrine cells. Since agonists of
GPR119 stimulate both glucose-dependent insulin secretion and
incretin release, GPR119 agonists have been under evaluation for
their potential to promote anti-diabetic effects and control glucose
homeostasis. Numerous Letters describing GPR119 agonists have
been disclosed, and a number of compounds have been progressed
into clinical trials.1 Most of the reported GPR119 agonists share a
common structural motif incorporating a piperidine with the
piperidine nitrogen capped as a carbamate or by a heterocycle,
and linked at the 4-position via a one atom spacer to an aryl group.
Interestingly, several recent publications underscore the impor-
tance of conformational restriction of the piperidine portion in
controlling human GPR119 pharmacology.2,3
We have recently reported our efforts to identify a series of
GPR119 agonists with a bridged piperidine moiety (Fig. 1).4,5
Bridged piperidines with both anti (1) and syn (2) configurations
functioned as agonists with moderate efficacy at human GPR119
(hcAMP % max) and weak to moderate efficacy at mouse GPR119
(mcAMP % max). Capping of the more potent syn isomer 1 as a sul-
fonamide 3 provided increased mouse GPR119 agonist efficacy.
Further optimization of the pyridinyl ring led to the discovery of
sulfonamide 4, a potent human GPR119 agonist. While 4 is a some-
what less potent agonist of mouse GPR119, its high % max made it
suitable for evaluation in vivo in a mouse oral glucose tolerance
test (oGTT). Compound 4 displayed significant activity in the
mouse oGTT when administrated orally with a minimum effective
dose of 10 mg/kg.4,6 However, the compound suffered from poor
aqueous solubility (0.17 lM, pH 7.4 buffer) which limited the
exposure of the compound in exploratory safety studies and pre-
cluded it from further development. To address this issue we
⇑
Corresponding author.
http://dx.doi.org/10.1016/j.bmcl.2015.09.047
0960-894X/Ó 2015 Elsevier Ltd. All rights reserved.

5292
X. Dai et al. / Bioorg. Med. Chem. Lett. 25 (2015) 5291–5294
O
HO
OH
N
N
N
N
O
O
NaIO4,
H2O
BnNH2, aq. HCl
HOOC COOH
O
N
N
O
O
O
O
N
NBn
O
O
O
O
Me
Me
Me
Me
5
N
O
O
O
1
2
O
hcAMP EC50 = 22 nM (72% max)
mcAMP EC50 = 55 nM (20% max)
hcAMP EC50 = 190 nM (58% max)
mcAMP EC50 = 690 nM (69% max)
1. H2,Pd/C
2. Boc2O
O
O
HO
Mitsunobu
O
NaBH4
NBoc
HO
NBoc
NBoc
syn
anti
6
7
8
NC
N
N
N
N
N
O
O
Scheme 1. Synthesis of 3-oxa-9-azabicyclo[3.3.1]nonan-7-ol.
O
N
H
O
N
O
O
N
O
S
Me
Me
S
Cl
Me
4
3
O
hcAMP EC50 = 23 nM (57% max)
mcAMP EC50 = 63 nM (76% max)
hcAMP EC50 = 3 nM (96% max)
N
O
N
N
N
NBoc
KOtBu
mcAMP EC50 = 140 nM (111% max)
moGTT: 39% lowering @30 mpk
Equal. Sol. = 0.17 uM (pH 7.4)
HO
ArNH2
NaH
NBoc
N
Cl
Cl
Cl
O
R1
9
R1
10
Figure 1. Initial optimization of bridged piperidine GPR119 agonists.
N
O
O
N
N
Ar
NBoc
Ar
NR2
i) TFA
ii) RCl
N
O
N
O
embarked on efforts to improve the solubility of compounds in this
series which is the subject of this communication. Polar surface
area (PSA), an indicator of hydrogen bonding capacity, plays an
important role in the aqueous solubility.7 Thus we explored intro-
duction of an oxygen atom into the two-carbon-bridge (Fig. 2, scaf-
folds B and C) to increase the PSA of the molecules and provide a H-
bond acceptor site for interaction with water, thereby improving
solubility.
H
H
R1
11
R1
12
Scheme 2. General synthesis of oxa-azabicyclo[3.3.1]nonane-derived GPR119
agonists.
NC
O
NC
anti
O
anti
N
N
N
N
The synthesis of syn and anti 3-oxa-9-azabicyclo[3.3.1]nonan-7-
ols 7 and 8 for scaffold B required development of a synthetic route
to the common precursor 9-benzyl-3-oxa-9-azabicyclo[3.3.1]
nonan-7-one (5). This was achieved by the oxidation of tetrahydro-
furan-3,4-diol and subsequent Robinson–Schopf reaction of the
resulting dialdehyde with benzylamine and acetonedicarboxylic
acid in aqueous medium (Scheme 1).8 Removal of the N-benzyl
group of 5 by catalytic hydrogenation followed by Boc protection
afforded ketone 6. Stereoselective reduction of the oxo function
of 6 with sodium borohydride furnished syn-3-oxa-9-azabicyclo
[3.3.1]nonan-7-ol (7). The anti-3-oxa-9-azabicyclo[3.3.1]nonan-7-
ol (8) was obtained by a simple Mitsunobu inversion of 7.
The starting material for scaffold C, 3-oxa-7-azabicyclo[3.3.1]
nonan-9-ol was synthesized as a mixture of syn and anti isomers
according to published procedures.9 The final targets embodying
scaffolds B and C were readily prepared by analogy to published
procedures as outlined in Scheme 2.10
N
H
O
N
H
O
N
O
O
N
Cl
Me
S
Cl
Me
O
O
13
14
hcAMP EC50 = 268 nM (64% max)
mcAMP EC50 = 1600 nM (100% max)
hcAMP EC50 = 114 nM (54% max)
mcAMP EC50 = 280 nM (34% max)
syn
syn
O
S
O
O
O
NC
N
N
N
N
N
N
O
O
O
O
O
N
H
N
H
O
Cl
Me
Cl
Me
H
H
15
16
hcAMP EC50 = 17 nM (38% max)
mcAMP EC50 = 100 nM (57% max)
moGTT@3 mpk: inactive
hcAMP EC50 = 44 nM (49% max)
mcAMP EC50 = 190 nM (78% max)
moGTT @3 mpk: active
Figure 3. Initial SAR of scaffold B.
In order to evaluate both scaffolds B and C, close analogs of
compound 4 were made. As shown in Figure 3, insertion of an oxy-
gen atom into the bridge of compound 4 to afford compound 13
(hEC50 = 268 nM) resulted in potency decrease of almost 90-fold.
However, simply switching the capping group from sulfonamide
to carbamate gave compound 14 of similar human GPR119
potency but significantly weaker functional activity at mouse
GPR119. The corresponding syn isomer 15 was more potent than
the corresponding anti isomer 14. Although this compound was a
NC
anti
NC
N
N
N
N
syn
N
H
O
N
H
O
O
S
N
O
Cl
Me
Cl
Me
S
N
O
O
O
O
18
17
hcAMP EC50 = 51 nM (56% max)
mcAMP EC50 = 844 nM (53% max)
hcAMP EC50 = 1210 nM (52% max)
NC
anti
NC
N
N
syn
N
N
N
H
O
O
N
H
O
N
O
Cl
Me
Cl
Me
N
O
O
O
O
O
B
N
O
19
20
hcAMP EC50 = 7 nM (65% max)
mcAMP EC50 = 28 nM (110% max)
hcAMP EC50 = 20 nM (83% max)
mcAMP EC50 = 140 nM (69% max)
3-oxa-9-azabicyclo[3.3.1]nonan-7-ol
(tPSA = 21.7)
O
A
O
N
O
Figure 4. Initial SAR of scaffold C.11
8-azabicyclo[3.2.1]octan-3-ol
(tPSA = 12.4)
C
N
3-oxa-7-azabicyclo[3.3.1]nonan-9-ol
(tPSA = 21.7)
partial agonist and not active in the mouse oGTT assay at 3 mg/kg
p.o., compound 16, which had a methanesulfonyl group on the
phenyl ring in place of the cyano group of 15, lowered blood
Figure 2. Bridged piperidines.

X. Dai et al. / Bioorg. Med. Chem. Lett. 25 (2015) 5291–5294
5293
anti
bridge in syn isomer 22 was proposed to lock the piperidine ring
in the equatorial conformation, precluding access to the axial
conformation required for agonist activity. Our observation that
the syn isomers 17 and 19 show agonist activity at human GPR119
similar to that of the anti analogs 18 and 20 suggests that analysis
of functional activity based solely on the relative position of the
piperidine rings is insufficient, and that the nature of the
piperidine capping group and/or the distal aryl group also
influences functional GPR119 activity.
syn
N
N
N
N
N
N
O
O
O
O
O
N
O
Me
Me
Me
Me
N
O
O
O
21
22
O
agonist
hcAMP EC50 = 65 nM (78% max)
antagonist
hcAMP EC50 > 10000 nM
Figure 5. Conformation-based hypothesis by McClure and co-workers.2a
Since we had identified promising GPR119 agonist activity with
analogs that incorporated each of the scaffolds B and C, both struc-
tural types were further investigated. To address concerns of sus-
ceptibility to acid cleavage and oxidative metabolism the t-butyl
carbamate of 16 and 19 was replaced with the more stable methyl-
cyclopropyl carbamate affording 23 and 24. As shown in Figure 6,
both 23 and 24 had improved solubilities validating our original
design. In pH 7.4 buffer, the equilibrium solubilities of 23 and 24
glucose levels significantly. Exploration of scaffold C (Fig. 4)
revealed similar trends to those observed for scaffold B, with the
syn isomers (17 and 19) being more potent than the anti isomers
(18 and 20). Moreover, the carbamates (19 and 20) were generally
more potent than the sulfonamides (17 and 18) and also tended to
display greater agonist activity at human and mouse GPR119.
The agonist activity we have observed with scaffold C syn iso-
mers 17 and 19 stands in contrast to results for similar analogs
reported by McClure et al. (Fig. 5).2a They reported anti isomer 21
to be an agonist whereas its syn counterpart 22 was reported to
be an antagonist devoid of intrinsic agonist activity at human
GPR119. To rationalize this observation, McClure et al. invoked
that the oxygen-containing bridge of anti isomer 21 locks the
piperidine ring in the axial conformation that is required for
agonist activity. On the other hand, the presence of the ether
were 3.4 and 6 lM, respectively, compared to 0.17 lM for com-
pound 4. In vitro activity trends with 23 and 24 were consistent
with those observed for the corresponding t-butyl carbamates 16
and 19. Scaffold C as in 24 conferred greater potency at human
and mouse GPR119 relative to 23, although agonist efficacy was
similar between the two compounds. Both 23 and 24 showed
robust glucose lowering activity in a mouse oGTT test at the oral
dose of 3 mg/kg, with the significantly higher unbound plasma
O
scaffold B
NC
O
O
scaffold C
N
N
S
N
N
N
O
O
N
H
O
N
O
N
H
O
Cl
Me
24
hEC50 = 4 nM (74% max)
F
Me
O
H
O
a
23
hEC50 = 70 nM (87% max)
mEC50 = 511 nM (79% max)
mEC50 = 33 nM (102% max)
moGTT: 84% @3 mpk (p<0.001)a
moGTT: 79% @3 mpk
(p<0.001)
(unbound plasma conc. (1h) 100 nM)
Equil. Sol. = 3.4 uM (pH 7.4)
(unbound plasma conc. (1h) 3 nM)
Equil. Sol. = 6 uM (pH 7.4)
Rat AUC0-6 h = 2396 nM.h (10 mpk)b
Rat AUC0-6 h = 4981 nM.h (10 mpk)b
Figure 6. Profile of compounds 23 and 24. a % glucose lowering relative to vehicle treated group. b plasma AUC0–6h determined following oral administration of compound in
20% hydroxypropyl b-cyclodextrin vehicle. Plasma levels averaged from 2 rats following serial bleeding in Sprague–Dawley rats.
Table 1
SAR of scaffold C
Compound
hEC50 (nm) (% max)
mEC50 (nm) (% max)
moGTTa@ 3 mpk (%)
Rat AUC0–6h (nM h)b@10 mpk
NC
N
N
N
N
N
OMe
N
N
H
O
O
O
O
25a
25b
7 (88)
43 (99)
78
8235
N
N
N
N
O
O
O
O
Cl
O
O
O
O
O
O
O
O
NC
N
H
25(88)
5 (99)
16 (78)
254 (105)
47 (116)
128 (143)
78
78
75
1189
1425
1490
Cl
H
NC
N
N
H
25c
25d
F
OMe
N
NC
O
Cl
Me
a
% glucose lowering relative to vehicle treated group.6
Plasma AUC0-6h determined following oral administration of compound in 20% hydroxypropyl b-cyclodextrin vehicle. Plasma levels averaged from 2 rats following serial
b
bleeding in Sprague–Dawley rats.

5294
X. Dai et al. / Bioorg. Med. Chem. Lett. 25 (2015) 5291–5294
significantly in the mouse oGTT assay at low oral doses and plasma
oGTT
A
levels. Our subsequent efforts that led to the identification of a
highly effective molecule in the 3-oxa-7-azabicyclo[3.3.1]nonan-
9-ol series will be detailed elsewhere.
300
250
200
150
100
Vehicle
0.1 mpk
0.3 mpk
1 mpk
Acknowledgment
3 mpk
Water
The authors would like to thank Tze-Ming Chan and Rebecca
Osterman for their NMR spectroscopy support.
Supplementary data
-30 -20 -10
0
10 20 30 40 50 60
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2015.09.
047.
Time (min)
B
AUC0-60min
15000
14000
13000
12000
11000
10000
9000
References and notes
1. For reviews: (a) Shah, U.; Edmondson, S.; Szewczyk, J. W. In New Therapeutic
Strategies for Type 2 Diabetes; Jones, Rob M., Eds.; RSC Drug Discovery Series;
Royal Society of Chemistry: Thomas Graham House, Science Park, Milton Road,
Cambridge CB4 0WF, 2012; Vol. 27, pp 177.; (b) Yoshida, S. Expert Opin. Invest.
Drugs 2012, 21, 321; (c) Jones, R. M.; Leonard, J. N. Annu. Rep. Med. Chem. 2009,
44, 149; For recent publications: (d) Wacker, D. A.; Wang, Y.; Broekema, M.;
Rossi, K.; O’Connor, S.; Hong, Z.; Wu, G.; Malmstrom, S. E.; Hung, C.-P.; LaMarre,
L.; Chimalakonda, A.; Zhang, L.; Xin, L.; Cai, H.; Chu, C.; Boehm, S.; Zalaznick, J.;
Ponticiello, R.; Sereda, L.; Han, S.-P.; Zebo, R.; Zinker, B.; Luk, C. E.; Wong, R.;
Everlof, G.; Li, Y.-X.; Wu, C. K.; Lee, M.; Griffen, S.; Miller, K. J.; Krupinski, J.;
Robl, J. A. J. Med. Chem. 2014, 57, 7499; (e) Scott, J. S.; Bowker, S. S.;
Brocklehurst, K. J.; Brown, H. S.; Clarke, D. S.; Easter, A.; Ertan, A.; Goldberg, K.;
Hudson, J. A.; Kavanagh, S.; Laber, D.; Leach, A. G.; MacFaul, P. A.; Martin, E. A.;
McKerrecher, D.; Schofield, P.; Svensson, P. H.; Teague, J. J. Med. Chem. 2014, 57,
8984.
2. (a) McClure, K. F.; Darout, E.; Guimarães, C. R. W.; DeNinno, M. P.; Mascitti, V.;
Munchhof, M. J.; Robinson, R. P.; Kohrt, J.; Harris, A. R.; Moore, D. E.; Li, B.;
Samp, L.; Lefker, B. A.; Futatsugi, K.; Kung, D.; Bonin, P. D.; Cornelius, P.; Wang,
R.; Salter, E.; Hornby, S.; Kalgutkar, A. S.; Chen, Y. J. Med. Chem. 2011, 54, 1948;
(b) Darout, E.; Robinson, R. P.; McClure, K. F.; Corbett, M.; Li, B.; Shavnya, A.;
Andrews, M. P.; Jones, C. S.; Li, Q.; Minich, M. L.; Mascitti, V.; Guimarães, C. R.
W.; Munchhof, M. J.; Bahnck, K. B.; Cai, C.; Price, D. A.; Liras, S.; Bonin, P. D.;
Cornelius, P.; Wang, R.; Bagdasarian, V.; Sobota, C. P.; Hornby, S.; Masterson, V.
M.; Joseph, R. M.; Kalgutkar, A. S.; Chen, Y. J. Med. Chem. 2013, 56, 301; (c)
Deninno, M. P.; Futatsugi, K.; Lefker, B. A.; Mascitti, V.; McClure, K. F.;
Munchhof, M. J., WO 2,010,106,457, 2010.
3. (a) Scott, J. S.; Brocklehurst, K. J.; Brown, H. S.; Clarke, D. S.; Coe, H.;
Groombridge, S. D.; Laber, D.; MacFaul, P. A.; McKerrecher, D.; Schofield, P.
Bioorg. Med. Chem. Lett. 2013, 23, 3175; (b) Pham, T.-A. N.; Yang, Z.; Fang, Y.;
Luo, J.; Lee, J.; Park, H. Bioorg. Med. Chem. Lett. 2013, 23, 1349; (c) Jones, R. M.;
Buzard, D. J.; Han, S.; Kim, S. H.; Lehmann, J.; Zhu, X., WO 2,013,055,910, 2013.
4. Xia, Y.; Chackalamannil, S.; Greenlee, W. J.; Jayne, C.; Neustadt, B.; Stamford, A.;
Vaccaro, H.; Xu, X.; Baker, H.; O’Neill, K.; Woods, M.; Hawes, B.; Kowalski, T.
Bioorg. Med. Chem. Lett. 2011, 21, 3290.
5. Detail information of cAMP assay is available in the Supporting information.
6. Compounds were initially screened in vivo using an acute oral glucose
tolerance test (oGTT) in male C57B1/6NCrl mice (6–8 week old). Test
compound or vehicle was orally administered to overnight-fasted animals at
3 mg/kg. Glucose was administered to the animals 30 min post-dosing (3 g/kg
p.o.) and blood glucose levels were measured using a hand-held glucometer
(Bayer Breeze 2) prior to compound or vehicle dosing, prior to glucose
administration, and 20 min after glucose administration. The glucose level of
mice dosed with test compound was expressed as a percentage of the glucose
level in vehicle treated mice (defined as 100%).
0.02 µM 0.09 µM
*
*
0.22 µM 0.35µM
*
*
40% 44% 64% 62%
8000
Water Vehicle 0.1
0.3
1
3
Dose (mg/kg)
Figure 7. oGTT PK/PD response of orally administered compound 25a in lean mice
(10% NMP, 5% cremophor, and 20% HPbCD vehicle).
exposure of 23 likely compensating for it weaker potency at mouse
GPR119 relative to 24. Compounds 23 and 24 were also evaluated
in an oral rat pharmacokinetic assay and both compounds showed
high, sustained plasma exposures.
Having identified highly potent GPR119 agonistic activity hav-
ing scaffold C of 24, we shifted our focus to modifications else-
where in these molecules. Several examples are highlighted in
Table 1 and encompassed the substituent at the 5-position of the
pyrimidine core (25a and 25b), the halogen substituent on the phe-
nyl ring (25c), and the linker between the phenyl and pyrimidine
rings (25d). These compounds retained potent human GPR119
EC50 values while maintaining a consistent trend of being an order
of magnitude less potent at mouse GPR119. In general, these com-
pounds also had strong agonist properties as reflected by the high
% max values for human and mouse GPR119. The compounds uni-
formly displayed excellent in vivo activity in the mouse oGTT assay
when orally administered at a dose of 3 mg/kg. In a pharmacoki-
netic and pharmacodynamic dose–response study in lean mice
(Fig. 7),12 compound 25a orally administered across the dose range
0.1–3 mg/kg elicited a dose dependent lowering of blood glucose
excursion (as represented by the glucose AUC excursion from 0
to 60 min) with statistically significant lowering observed at all
doses. The total plasma concentrations of 25a increased in a less
than dose proportional across the dose range. Additionally, 25a
7. Beaumont, K.; Cole, S. M.; Gibson, K.; Gosset, J. R. From RSC Drug Discovery Series
1 Metabolism, Pharmacokinetics and Toxicity of Functional Groups; Royal
Society of Chemistry: Thomas Graham House, Science Park, Milton Road,
Cambridge CB4 0WF, 2010, pp 61–98.
8. Attila, A.; Pal, H.; Szilvia, J.; Laszlo, K.; Gyula, B. Tetrahedron 2001, 57, 235.
9. Huttenloch, O.; Laxman, E.; Waldmann, H. Chem. Eur. J. 2002, 8, 4767.
10. Xia, Y.; Boyle, C. D.; Greenlee, W. J.; Chackalamannil, S.; Jayne, C. L.; Stamford,
A. W.; Dai, X.; Harris, J. M.; Neustadt, B. R.; Neelamkavil, S. F.; Shah, U. G.;
Lankin, C. M.; Liu, H. WO 2,009,055,331, 2009.
maintained improved solubility with measured solubility 30
in pH 7.4 buffer.
lM
In summary, we prepared and profiled a series of GPR119 ago-
nists that incorporated two distinct conformationally restricted
oxazabicyclic scaffolds: 3-oxa-9-azabicyclo[3.3.1]nonan-7-ol (scaf-
fold B) and 3-oxa-7-azabicyclo[3.3.1]nonan-9-ol (scaffold C), with
a goal of improving the solubility over an initial lead 4. From this
work, we have discovered a potent and orally active GPR119 ago-
nist 25a, incorporating scaffold C, which lowered blood glucose
11. NMR studies of 19 and 20 are available in the Supporting information.
12. oGTT dose response studies were conducted in a similar manner as acute oGTT
with the following modifications. A vehicle-treated group challenged with
water rather than glucose was included in the study. Blood glucose was
measured prior to compound or vehicle dosing, prior to glucose or water
administration, and 20, 40, and 60 min after glucose or water administration.
The glucose area under the curve (AUC) from 0 to 60 min was calculated and
compared to the vehicle treated group.