PDE-10A inhibitors as insulin secretagogues

Article abstract of DOI:10.1016/j.bmcl.2007.02.061

Modulation of cAMP levels has been linked to insulin secretion in preclinical animal models and in humans. The high expression of PDE-10A in pancreatic islets suggested that inhibition of this enzyme may provide the necessary modulation to elicit increase

Full text of DOI:10.1016/j.bmcl.2007.02.061

Bioorganic & Medicinal Chemistry Letters 17 (2007) 2869–2873
PDE-10A inhibitors as insulin secretagogues
Louis-David Cantin,a,* Steven Magnuson,a David Gunn,a Nicole Barucci,b
Marina Breuhaus,d William H. Bullock,a Jennifer Burke,a Thomas H. Claus,b
Michelle Daly,b Lynn DeCarr,b Ann Gore-Willse,b Helana Hoover-Litty,b
Ellalahewage S. Kumarasinghe,a Yaxin Li,b Sidney X. Liang,a James N. Livingston,b
Timothy Lowinger,a Margit MacDougall,b Herbert O. Ogutu,a Alan Olague,a
Ronda Ott-Morgan,c Robert W. Schoenleber,a Adrian Tersteegen,d Philip Wickens,a
Zhonghua Zhang,a Jian Zhu,b Lei Zhua and Laurel J. Sweetb
aDepartment of Chemistry Research, Bayer Pharmaceuticals Corporation, West Haven, CT 06516, USA
bDepartment of Metabolic Disorders Research, Bayer Pharmaceuticals Corporation, West Haven, CT 06516, USA
cDepartment of Research Technologies, Bayer Pharmaceuticals Corporation, West Haven, CT 06516, USA
dDepartment of Cardiovascular Research and Screening, Bayer HealthCare AG, Wuppertal, Germany
Received 29 December 2006; revised 16 February 2007; accepted 21 February 2007
Available online 25 February 2007
Abstract—Modulation of cAMP levels has been linked to insulin secretion in preclinical animal models and in humans. The high
expression of PDE-10A in pancreatic islets suggested that inhibition of this enzyme may provide the necessary modulation to elicit
increased insulin secretion. Using an HTS approach, we have identified quinoline-based PDE-10A inhibitors as insulin secreta-
gogues in vitro. Optimized compounds were evaluated in vivo where improvements in glucose tolerance and increases in insulin
secretion were measured.
Ó 2007 Elsevier Ltd. All rights reserved.
Analogs of the incretin hormone glucagon-like peptide-1
(GLP-1) including Exenatide, Liraglutide, and Albugon
have recently emerged as an important therapeutic ap-
proach for the treatment of type 2 diabetes.1 In pancre-
atic b cells, these peptides activate the GLP-1 receptor,
an adenylate cyclase (AC) coupled GPCR (G-protein
coupled receptor), which elevates the intracellular con-
centration of cyclic adenosine mono-phosphate
(cAMP). Signaling by this second messenger has been
associated with glucose-dependent insulin secretion
(GDIS), stimulation of insulin biosynthesis and increase
in b cell mass. These new GLP-1 peptide agonists are
not orally active but are highly effective when adminis-
tered via subcutaneous injection.
Efforts to identify orally active small molecules that
mimic the agonistic activity of these 30–40 amino acid
peptides have not been successful. Consequently, the
greatest probability to develop an oral therapy lies in
identifying small molecule targets within this signaling
cascade. One such example is the inhibition of DPP4
(dipeptidyl peptidase-4).2 This protease is the key
enzyme responsible for inactivating endogenously
produced GLP-1 through hydrolysis of the penultimate
alanine peptide bond. DPP4 inhibitors have been shown
to enhance GLP-1 signaling and consequently lower
blood glucose when dosed orally in both animal models
of diabetes and in clinical trials.3
Since GLP-1 signals through AC and cAMP, an alterna-
tive approach to enhance GLP-1 signaling would be to
inhibit the phosphodiesterase(s) (PDE) responsible for
hydrolysis of cAMP and termination of the intracellular
signal. Treatment of islets with PDE-3B inhibitors has
been reported to result in glucose-dependent insulin
secretion.4 However, these PDE-3B inhibitors have been
associated with concomitant liver and adipose tissue
Keywords: PDE-10A inhibition; Insulin secretagogue; Quinoline; Type
2 diabetes.
*
Corresponding author. E-mail: Louis-David.Cantin@astrazeneca.
com
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.02.061

2870
L.-D. Cantin et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2869–2873
Table 1. SAR of substituents of the biphenyl moiety
insulin resistance, counterbalancing the effect on insulin
secretion. In addition, we have recently disclosed PDE-
11A inhibitors useful as insulin secretagogues.5
Compound
R6
PDE-1010
IC50 (nM)
3
4
5
6
7
8
n/a
4-Me
2-CH2-morpholine
400
140
45
Our search of novel targets for insulin secretion led to
the recent discovery of the high expression of PDE-
10A in pancreatic islets. This finding prompted an exten-
sive effort aimed at ascertaining the ability of PDE-10A
inhibitors to produce GDIS.6 Our past research efforts
in the PDE-10A inhibitor field had provided a pyrrolo
dihydroisoquinoline class of compounds (Fig. 1; 1 and
2) that proved to be potent PDE-10A inhibitors (PDE-
10A IC50 1 = 30 nM, 2 = 33 nM) that were also highly
selective over other PDEs.7 Because the variations of
cAMP levels are small and localized, the measurement
of these changes in a whole cell assay proved difficult.
Instead we profiled the compounds in a functional assay
using dispersed islet cells (D.I. assay),8 where com-
pounds 1 and 2 caused significant insulin secretion when
compared to control samples (D.I. fold-over-control:
1 = 1.4, 2 = 1.44).
2-CF3
2-Et
2-OEt
30
35
35
Before starting our formal optimization work, we
decided to incorporate a fluorine atom at C6 and a
nitrile at C3 (see Table 2, compound 14), which were
found to be favorable in a parallel chemical series. These
changes provided some improvements in the potency,
physicochemical properties, and oral exposure of the
compound class. In addition, the presence of at electron
withdrawing group at C-3 provided the synthetic flexi-
bility of introducing the amino-acid head-group as the
last step of the synthesis.
Utilizing a HTS approach, we identified multiple chemi-
cal subclasses of selective PDE-10A inhibitors capable
of causing insulin secretion in vitro, among which quin-
oline scaffold 3 proved to be of high interest (Fig. 2).
Indeed, Kyowa researchers9 had also determined that
similar compounds were PDE-10A inhibitors, targeting
oncology applications. Initial studies allowed us to
determine that a nitrogen atom handle at C4 could be
used to further homologate the carboxylic acid group.
The general synthetic route allowed for the independent
optimization of the different subunits of the analogs;
Scheme 1 illustrates the late-stage introduction of the
amino acid moiety. It should be noted that the Suzuki
and chlorination–displacement steps can be inverted to
allow for an efficient exploration of the SAR of the
biphenyl fragment. The addition of the sodium anion
of acetonitrile to a bromo benzoate, followed by reac-
tion of the incipient enolate with 5-fluoroisatoic anhy-
Previous reports9 indicated that an ortho-substituent on
the distal aryl ring was beneficial. Our synthesis of com-
pounds11 4–8 (Table 1) confirmed the preference for
ortho-substituents with our quinoline amino-acid inhib-
itors, and also revealed a very high tolerance to various
steric and electronic properties. From this initial set of
analogs, compound 8 became the starting point of our
chemical optimization effort aimed at improving the
aqueous solubility and CYP inhibition profile, while
achieving good oral exposure.
Table 2. SAR of the amino acid head-group
CO2H
n(H2C)
R2
R1
R3
N
F
CN
OEt
N
Compound R1
R2
R3
n
PDE-10
IC50 (nM)
MeO
R =
OMe
N
CH3
MeO
N
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
0
1
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
7
H
2
O
OMe
R
OH
H
3
OEt
1
2
Me
Et
Ph
H
H
4
H
7
Figure 1. Pyrrolodihydroisoquinoline PDE-10A inhibitors.
H
120
6
Me
Et
H
5
CO2H
H
-CH2-CF3
-CH2-Ph
-CH2-(3-Pyr)
2
H
3
CO2H
Me
HN
H
0.3
10
1
4
6
Me Me
3
H
H
H
H
H
H
-CH2-CH2-
N
N
-CH2-CH2-CH2ÀCH2-
16
6
R6
R-Me
S-Me
R-Bn
S-Bn
3
4
22
2
Figure 2. Quinoline-based PDE-10A inhibitors from HTS.

L.-D. Cantin et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2869–2873
2871
Table 3. SAR of the quinoline ring substituents
OH
O
F
CN
CO2H
a, b
c, d
OEt
HN
N
Me
CN
N
5
8
Br
6
7
9
10
R4
Br
Cl
OEt
R
N
F
CN
F
CN
e or f, g
N
Compound
R4
PDE-10A
IC50 (nM)
11
Ar
12
R = amino acids
Ar
31
27
32
33
34
35
36
37
38
H
6-F
2
6
Scheme 1. Reagents and conditions: (a) NaH, MeCN, THF; 9, THF;
(b) 5-fluoroisatoic anhydride, THF; 70% yield (two steps); (c) Ar–
B(OH)2, PdCl2(dppb), Na2CO3, DMF, H2O, 120 °C, 80–90% yield; (d)
POCl3, 120 °C, 95% yield; (e) amino acid, Et3N, DMA, 120 °C, 10–
50% yield; (f) amino ester, Et3N, DMSO; (g) LiOH, EtOH, H2O, THF,
30–75% yield (two steps).
6-Cl
18
3
6-MeO
7-Me
7-Cl
6
21
8-Me
5-F,6-F
6-F,7-F
1900
>4000
100
dride, produced the desired quinoline 10 in a one-pot
process (Scheme 1). Subsequent Suzuki-type cross-cou-
pling and chlorination produced advanced intermediate
11. Treatment of 11 with amino acids or to amino esters
followed by hydrolysis, provided target compounds 12.
to the C7 position (34 and 35) and to 6,7-disubstituted
compounds (38). However, introduction of small groups
at either C5 or C8 (36 and37), as well as 6,7-disubstitu-
tion (38), resulted in significant decreases in potency.
The SAR of the middle aromatic ring of the biphenyl
group was also explored, which revealed an interesting
interplay with the amino acid moiety. Indeed, introduc-
tion of small substituents on the proximal aryl ring is
well tolerated when combined with a c-amino acid at
C4 (39–42), whereas the same substituents caused a sig-
nificant decrease in potency when combined with the
alanine at C4 (43–46) (Table 4).
SAR of the amino acid group revealed a general toler-
ance for groups of various sizes and electronic properties
(Table 2). The length of the spacer between the amino
and acid groups has little effect on the potency of the
compounds (13–15). Because of concerns with the
potential formation of reactive metabolites with the b-
amino acid moiety, we focused on a-amino acids. Fur-
ther alkylation or arylation of the nitrogen (16–18)
decreases the potency, and also decreased the stability
of the compounds in aqueous acids. Branching of substit-
uents on the a-carbon is also well tolerated (19–23). How-
ever, we found that the introduction of large lipophilic
groups at the R2/R3 positions generally leads to poor
CYP inhibition profile. Disubstitution is also well toler-
ated, as the gem-dimethyl (24), spiro-cyclopropyl (25),
and spiro-cyclopentyl (26) analogs are potent PDE-10A.
After further profiling of compounds in functional phys-
iochemical and CYP inhibition assays, two compounds
(27 and 40) were progressed toward in vivo evaluation.
Both compounds are potent and selective PDE-10A
inhibitors, and show strong insulin secretion in the
Table 4. SAR of the amino acid head-group with modifications to the
biphenyl substituents
Consistent with these findings, we did not measure sig-
nificant eudismic ratios with small substituents (Table
2; 27 and 28), and larger substituents (29 and 30) display
only a slight preference for the (S)-enantiomer. Upon
further profiling of compounds 27 and 28, we measured
a significant difference in the CYP inhibition profile be-
tween the two enantiomers (CYP 2C9 (IC50) 27,
2000 nM; 28, 66 nM). The SAR exploration was contin-
ued using the R-enantiomer of alanine at C4.
X
F
CN
OEt
N
2
R5
3
Compound
X
R5
PDE-10
IC50 (nM)
39
40
41
42
3-F
2-F
18
4
CO2H
Having better defined the SAR at the C4 position, the
corresponding O-linked and CH2-linked analogs of
compound 13 were synthesized.11 In both cases the mod-
ification caused a significant decrease in potency (100-
and 10-fold, respectively). We then turned our attention
to the other positions of the quinoline, keeping the C2–
C4 positions constant (Table 3).11 The C6 position of
the quinoline proved tolerant to both electron-with-
drawing and -donating groups (27, 32, and 33) similarly
2-Me
3-Me
120
34
HN
43
44
45
46
3-F
2-F
160
79
CO2H
2-Me
3-Me
4000
1300
HN

2872
L.-D. Cantin et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2869–2873
Table 5. In vitro profile of compounds 27 and 40
Compound 27
insulin secretagogues in vitro. Further optimization
allowed us to demonstrate that this activity also trans-
lated into glucose lowering effects in vivo, without signs
of hypoglycemia.
Compound 40
PDE-10A (nM)
PDE selectivity12
Disp. Islets (F.O.C.)
6
4
>1000-fold
1.4
>100-fold
1.9
F.O.C., fold over control.
Acknowledgment
dispersed islet cell assay (Table 5).8 The rat oral expo-
sure of both compounds was evaluated at a dose of
1 mpk, and sustained plasma levels were measured well
past 12 h.
´
We thank Jefferson Chin, Laszlo Musza, and Anthony
Paiva for NMR and LC-MS support.
´
References and notes
Having identified compounds with suitable in vitro
pharmacology and in vivo pharmacokinetic profiles,
the in vivo efficacy of these PDE-10A inhibitors was
assessed using a rat intraperitoneal glucose tolerance test
(IPGTT),13 where both compounds significantly re-
duced glucose levels at various doses (e.g., Figs. 3 and
4, 27: 20.4% glucose AUC reduction, 40: 22.8% glucose
AUC reduction) following a glucose challenge. Further,
these compounds were shown not to cause hypoglyce-
mia in rats. In addition, compound 27 provided measur-
able increases in insulin secretion in vivo.14 Taken
together, these results suggest that PDE-10A inhibitors
may prove useful agents in the treatment of type 2 dia-
betes through improvements in the insulin secretion
function of b-cells in a glucose-dependent manner.
1. Drucker, D. J. Cell Metab. 2006, 3, 153.
2. (a) Mest, H. J. Curr. Opin. Investig. Drugs 2006, 7, 338; (b)
Ahren, B. Expert Opin. Investig. Drugs 2006, 15, 431.
3. (a) Weber, A. J. Med. Chem. 2004, 48, 4135; (b) Drucker,
D. J. Exp. Opin. Investig. Drugs 2003, 12, 87; (c)
Wiedeman, P. E.; Trevillyan, J. M. Curr. Opin. Investig.
Drugs 2003, 4, 412; (d) Miller, S. A.; St. Onge, E. L.
Sitagliptin: A Dipeptidyl Peptidase IV Inhibitor for the
Treatment of Type 2 Diabetes. Ann. Pharmacother. 2006,
40, 1336; (e) Ahren, B. Expert Opin. Investig. Drugs 2006,
15, 431.
4. (a) Shafiee-Nick, R.; Pyne, N. J.; Furman, B. L. Br.
J. Pharmacol. 1995, 115, 1486; (b) Han, P.; Werber, J.;
Surana, M.; Fleischer, N.; Michaeli, T. J. Biol. Chem.
1999, 274, 22337.
5. (a) Bullock, W.; Chen, L.; Barucci, N.; Breuhaus, M.;
Claus, T.; Daly, M.; Gunn, D.; Lemoine, L.; Li, Y.; Liu,
Q.; Livingston, J.; Sweet, L.; Tersteegen, A.; Town, C.;
Wang, H.; Zhu, J.; Heurich, R.; Wang, Y. Abstracts of
Papers, 232nd ACS National Meeting, San Francisco, CA,
United States, Sept 10–14, 2006; (b) Wang, Y.; Bear, T.;
Bullock, W.; Chen, L.; Guernon, L.; Gunn, D.; Hunyadi,
L.; Kramss, R.; Li, T.; Liang, S.; Liu, Q.; Liu, D.;
Magnuson, S.; Mannelly, G.; Miller, D.; Mull, E.; Natero,
R.; O’Keefe, D.; Qi, N.; Wood, J.; Barucci, N.; Breuhaus,
M.; Claus, T.; Daly, M.; Lemoine, L.; Li, Y.; Livingston,
J.; Sweet, L.; Tersteegen, A.; Wang, H.; Zhu, J.; Heurich,
R. Abstracts of Papers, 232nd ACS National Meeting, San
Francisco, CA, United States, Sept 10–14, 2006.
In summary, we have shown that selective PDE-10A
inhibitors from multiple chemical series are potent
300
250
200
150
Vehicle
100
50
0
1mpk Compound 27
6. Sweet, L. PDE10a inhibitors for treating diabetes and
related disorders. PCT WO 2005012485.
7. Niewoehner, U.; Bauser, M.; Ergueden, J.-K.; Flubacher,
D.; Naab, P.; Repp, T.-O.; Stoltefuss, J.; Burkhardt, N.;
Sewing, A.; Schauer, M.; Schlemmer, K.-H.; Weber, O.;
Boyer, S.J.; Miglarese, M. Preparation of pyrrolo[2,1-
a]isoquinolines as phosphodiesterase 10a inhibitors. WO
2002048144.
0
15
30
45
60
Minutes
Figure 3. Evaluation of compound 27 in the rat IPGTT.
8. Dispersed islet cells assay.Dispersed pancreatic islet isola-
tion. Lean rats (male Sprague–Dawley, 200–250 g) are
anesthetized with CO2 (60 mg/kg, ip) and the abdomen
opened to expose the liver and pancreas. The pancreas is
distended by injection of Hanks’ solution into the bile
duct, and then the pancreas is excised and minced with
scissors while in Hanks’ solution. After rinsing the tissue
with buffer, the pancreas is digested for 10 min with
collagenase, rinsed, and the islets separated from debris on
a Ficoll gradient. The isolated islet fraction is then rinsed
with buffer, incubated with EDTA for 8 min, followed by
incubation with trypsin and DNAase I for an additional
10 min. The dispersed islets are transferred to culture
media containing 8 mM glucose, seeded in ‘V-bottom’ 96-
well plates (2500 cells per well), and cultured over-
night.For assaying insulinotropic compounds, the
280
240
200
160
120
80
40
0
Vehicle
3mpk Compound 40
0
15
30
45
60
Minutes
Figure 4. Evaluation of compound 40 in the rat IPGTT.

L.-D. Cantin et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2869–2873
2873
dispersed islets are pre-incubated in 3 mM glucose for
30 min. The islets are then transferred to media containing
8 mM glucose, test compounds, and 0.3 lM forskolin, and
incubated for an additional 30 min. The media are then
assayed for insulin content using scintillation proximity
assay (SPA). Data are expressed as a fold increase over the
response of a 8-mM glucose and 0.3 lM forskolin solu-
tion. Values reported are means of at least two
experiments.
11. (a) All compounds described gave consistent 1H NMR
and LC/MS data. For more details on their synthetic
preparations, see: (a) Cantin, D.; Magnuson, S.; Gunn,
D.; Bullock, W.; Burke, J.; Fu, W.; Kumarasinghe, E.S.;
Liang, S.X.; Newcom, J.; Ogutu, H.; Wickens, P.; Zhang,
Z.; Bierer, D. Phenyl substituted quinoline and quinazo-
line compounds for the treatment of diabetes.
WO2006034491; (b) Cantin, D.; Magnuson, S.; Gunn,
D.; Bullock, W.; Burke, J.; Fu, W.; Kumarasinghe, E.S.;
Liang, S.X.; Newcom, J.; Ogutu, H.t; Olague, A.; Wang,
M.; Wickens, P.; Zhang, Z.; Bierer, D. Phenyl-substituted
quinoline and quinazoline compounds for the treatment of
diabetes. WO2006034512.
12. PDE selectivity assays were performed with PDE-2, PDE-
3, PDE-4, PDE-5, PDE-7, PDE-8, PDE-9, and PDE-11.
13. In vivo studies. All animals were purchased at 6 weeks of
age and were maintained on standard laboratory rodent
chow ad libitum. Lean rats (male Wistar, 250–300 g) are
fasted overnight and divided into two groups: vehicle and
compound treatment (eight rats per group). Vehicle or
compound is administered via oral gavage (1.5 mL/rat).
Three hours later, a glucose solution (30%, 2 g/kg body
weight) is injected intraperitoneally. Tail blood samples
are collected at 0, 15, 30, and 60-min time points after the
glucose injection to measure blood glucose using Glucom-
eter (Bayer Diagnostics, Mishawaka, IN). FA-GLP-1 was
used as a positive control in this assay, proving ꢀ20%
decrease in glucose AUC.Statistical analysis. All results
are expressed as means SEM for the number of animals
indicated in the figure legends. ANOVA was used to
evaluate the effects of compounds 27 and 40 using InStat
(GraphPad Software Inc., San Diego, CA). The Tukey–
Kramer multiple comparisons test was used for the
parametric ANOVA. Whenever a nonparametric ANOVA
was required, the Kruskal–Wallis Test was used. Results
were considered significant at p < 0.05.
9. Osakada, N.; Haruoka, M.; Ikeda, K.; Toki, S.; Miyaji,
H.; Shimada, J. Preparation of quinoline derivatives as
phosphodiesterase 10A inhibitors. PCT WO 2004002484.
10. PDE inhibition assays.Preparation of PDE isoen-
zymes.Human PDEs 11A, 10A, 9A, 8A, 7B, 4B, 3B, and
2A were recombinantly expressed in Sf9 insect cells. Cell-
free extracts were prepared and stored at À80 °C. PDE5
was purified from human platelets by sonication followed
by centrifugation and ion exchange chromatography of
the supernatant on a Mono Q 10/10 column. PDE6 was
purified from rod outer segments (ROS) of bovine retinae,
activated by mild trypsination, and further purified by ion
exchange chromatography on a Mono Q column by
Dr. Ko¨rschen, Forschungszentrum Julich, Germany, as
¨
described in Tejada et al. Int. J. Impot. Res. 2001, 13, 282–
290. Bovine PDE1 was purchased from Sigma–Aldrich (P
9529, Taufkirchen, Germany).Enzyme inhibition.The com-
3
3
mercially available H-cAMP and H-cGMP Scintillation
Proximity Assay (GE Healthcare, Little Chalfont, Buck-
inghamshire, UK) system was used for PDE enzyme
inhibition studies as detailed in Wunder, F.; Tersteegen,
A.; Rebmann, A.; Erb, C.; Fahrig, T.; Hendrix M. Mol.
Pharmacol. 2005, 68(6), 1775–1781. IC50 values were
determined from sigmoidal curves by plotting the per-
centage of PDE activity versus log compound concentra-
tion. Values reported are means of at least two
experiments. A reference compound was used on every
plate to ensure reproducibility and a typical standard
deviation was 30%.
14. Compound 40 was not tested in the in vivo insulin
secretion assay.