Discovery of 4-hydroxy-1,6-naphthyridine-3-carbonitrile derivatives as novel PDE10A inhibitors

Article abstract of DOI:10.1016/j.bmcl.2012.01.046

A series of 1,6-naphthyridine-based compounds was synthesized as potent phosphodiesterase 10A (PDE10A) inhibitors. Structure-based chemical modifications of the discovered chemotype served to further improve potency and selectivity over DHODH, laying the foundation for future optimization efforts.

Full text of DOI:10.1016/j.bmcl.2012.01.046

Bioorganic & Medicinal Chemistry Letters 22 (2012) 1944–1948
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of 4-hydroxy-1,6-naphthyridine-3-carbonitrile derivatives as novel
PDE10A inhibitors
Udo Bauer a, , Fabrizio Giordanetto a, , Martin Bauer a, Gavin O’Mahony a, Kjell E. Johansson a, Wolfgang
⇑
⇑
Knecht a, Judith Hartleib-Geschwindner a, Eva Töppner Carlsson a, Cristofer Enroth b
a AstraZeneca, R&D Mölndal, Pepparedsleden 1, S-431 83 Mölndal, Sweden
b Katedralskolan, Department of Natural Sciences, S-53288 Skara, Sweden
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of 1,6-naphthyridine-based compounds was synthesized as potent phosphodiesterase 10A
(PDE10A) inhibitors. Structure-based chemical modifications of the discovered chemotype served to fur-
ther improve potency and selectivity over DHODH, laying the foundation for future optimization efforts.
Ó 2012 Elsevier Ltd. All rights reserved.
Received 8 December 2011
Revised 12 January 2012
Accepted 13 January 2012
Available online 26 January 2012
Keywords:
PDE10A
Phosphodiesterase
Hit to lead
Enzyme inhibitor
Naphthyridine
The cyclic nucleoside monophosphates cAMP and cGMP serve
as critical intracellular signaling agents. Signal termination is reg-
ulated by hydrolytic cleavage of cAMP and cGMP by members of
the phosphodiesterase (PDE) family of enzymes. Among the 11
PDE families, PDE10A has received significant attention as a possi-
ble target for new anti-psychotic or -diabetic drugs.1–3 First gener-
ation, selective PDE10A inhibitors have been tested in animal
models to explore the biochemical and behavioral consequences
of inhibiting this molecular target. Papaverine and TP-10 (Fig. 1)
are active in a number of models that predict anti-psychotic behav-
ior4,5 at doses that overlap the biochemical changes, suggesting a
link between PDE10A inhibition and desired behavioral outcomes.
These results are consistent with earlier studies testing PDE10A
knockout mice in behavioral models.6 Furthermore, papaverine
and TP-10 demonstrated efficacy in rodent models of negative
symptoms and cognition,4 suggesting the potential for broad
efficacy in schizophrenia with PDE10A inhibitors. Recent data
suggest that b-cell PDE10A plays a role in regulating cAMP-induced
glucose-stimulated insulin secretion (GSIS). Inhibition of PDE10A
resulted in an increase in insulin release in vitro and was associ-
ated with improved glucose tolerance in vivo.7 Moreover, inhibi-
tion of PDE10A had beneficial effects on weight loss. PDE10A KO
mice challenged with a high fat diet were resistant to obesity,
and administration of PDE10A selective inhibitors to diet-induced
obese (DIO) mice resulted in weight loss.8 Accordingly, PDE10A
inhibition could provide an alternative therapeutic approach for
the treatment of obesity and diabetes.
Several classes of structurally diverse PDE10A inhibitors have
been described and examples are shown in Figure 1. Starting points
have originated from the natural product papaverine,3 random
screening (TP-10, 1 and 2),9,10 and from fragment screening.11
In this paper, we describe the use of known ligand information
applied to the discovery of PDE10A inhibitors. Starting from 2, a
hit-to-lead program was initiated with the intent to reduce its lipo-
philicity, whilst securing adequate PDE10A potency. Replacement
of the D-alanine moiety of 2 was of particular importance, in order
to minimize acylglucuronide formation and the potential risk for
toxicity. We thus reasoned that a scaffold hop from quinoline to
4-hydroxy-1,6-naphthyridine (Fig. 1) would increase the overall
polarity of the ligands and avoid the amino acid side chain of 2,
whilst enabling chemical variation of the substitution pattern.
Accordingly, a number of initial compounds were synthesized
(3–4, Scheme 1, and 5–6, Scheme 2) to validate this hypothesis.
The results are summarized in Table 1.
The hypothesized scaffold hop (5, PDE10A IC50: 12 nM) offered
an improvement in ligand efficiency (LE) and ligand lipophilicity
efficiency (LLE) over 2, while a 4-methoxy substituent (6) signifi-
cantly decreased PDE10A potency (Table 1). The 3-cyano substitu-
ent was essential for significant PDE10A affinity (cf. 5 and 6,
Table 1).
⇑
Corresponding authors. Tel.: +46 31 776 2150; fax: +46 31 776 3700 (U.B.); tel.:
+46 31 7065723; fax: +46 31 776 3700 (F.G.).
E-mail address: udo.bauer@astrazeneca.com (U. Bauer).
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2012.01.046

U. Bauer et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1944–1948
1945
O
O
N
O
O
O
N
H
N
N
S
N
O
N
N
F
F
O
F
Papaverine
TP-10
1
IC50 36 nM
IC50 < 1 nM
IC50 35 nM
OH
O
HN
N
OH
N
N
F
R
R
N
O
2
IC50 8 nM
Figure 1. Selected examples of PDE10A inhibitor leads and initial scaffold hopping hypothesis.
O
OH
R
N
NH
R
a or b
d or e
N
3, 4
+
O
O
O
NH2
R = CO2Me (for 3) or Br (for 4)
R = CO2Me
R = Br
R = Ac
c
Scheme 1. Reagents and conditions: Synthesis of 3: R = CO2Me, (a) TBTU, NEt3, DCM, reflux, 3 d; (d) NH3, MeOH, 160 °C, 4 h. Synthesis of 4: R = Br, (b) oxalylchloride, DCM, rt,
0.6 h, then NEt3, DCM, rt, 2 h; (c) n-BuLi, THF, ꢀ78 °C to ꢀ30 °C, 1 h, then CH3CON(CH3)OCH3, ꢀ30 °C to rt, 70 min, then aq HCl; (e) NaOH, dioxane, reflux, 35 min.
Having successfully simplified the ligand, we investigated
O
whether PDE10A potency and LLE could be further improved over
5. Based on the predicted, and subsequently experimentally deter-
mined (vide infra), three-dimensional molecular interaction
between 5 and PDE10A, the distal phenyl ethoxy substituent was
systematically modified, as shown in Table 2. Compounds 7–20
were synthesized following Scheme 2.
OH
N
N
O
N
a
O
N
N
+
O
N
H
Br
H
Cl
Br
30
31
32
In general, ether substituents (5, 8–10) afforded good PDE10A
potency, with the bulky trifluoromethyl showing the best inhibi-
OH
N
tion (9, PDE10A IC50
:
1 nM).12 Small, electron-withdrawing
N
(11–14) and electron-donating (15–16) groups displayed a PDE10A
potency increase dependent on their volume (Table 2). Introduc-
tion of markedly polar substituents such as phenol (7), alcohol
(17) and amines (18, 19) was not favoured, as shown in Table 2.
X-ray crystallography of the complex formed between com-
pound 5 and PDE10A (Fig. 2, PDB code is 4ael) showed an atypical
binding mode, with a papaverine-type hydrogen-bonding interac-
tion with Q726 not observed.13 The only polar interaction is a
hydrogen-bond between Y693 and the N-1 of the naphthyridine
ring system and disrupting this interaction led to a loss of potency.
When R2 (Table 2) is electron-withdrawing (e.g. nitrile), the enol
tautomer of the 4-hydroxynaphthyridine is favoured,14 thus N-1
acts as a hydrogen-bond acceptor and interacts with Y693. Non
electron-withdrawing groups (e.g. R2 = CH3, compound 20 or
R2 = H, compound 4) lead to a preference for the keto tautomer,
thus N1 becomes a hydrogen-bond donor and is no longer avail-
able for interaction with Y693. The R1 substituent on the distal
b or c
N
R
33
Scheme 2. Reagents and conditions: (a) DMA, NaH (2.6 equiv), 0–120 °C, 2 h, 74%.
(b) RB(OR0)2 (1.1–1.5 equiv), PPh3 (20 mol %), Pd(OAc)2 (10 mol %) or 1,10-bis(di-
tert-butylphosphino)ferrocene palladium dichloride (10 mol %), sodium carbonate
(3 equiv), THF or ACN, water, 30 min, 150 °C microwave. (c) (PinB)2, Pd(dppf)2Cl2,
potassium carbonate (3.3 equiv), DMF, 30 min, 130 °C, then R–I (1.3 equiv),
Pd(dppf)2Cl2, sodium carbonate hydrate, 1 h, 100 °C.
phenyl ring reduces rotational freedom in the biphenyl moiety,
thus stabilizing the bioactive conformation. R1 also occupies a
relatively hydrophobic pocket formed by L635, M713 and F729.
The central phenyl ring is sandwiched between Q726 and F729,

1946
U. Bauer et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1944–1948
Table 1
Table 2
PDE10A potency and ligand efficiency metrics for compounds 2–6
PDE10A potency and ligand efficiency metrics for compounds 5, 7–20
1
2
R1
O
HO
R
R
X
N
N
N
N
Compound R1
R2
PDE10Aa IC50 (nM) LEb
LLEc
R1
X
PDE10A IC50 (nM)
LEb
LLEc
a
Compound
5
7
8
OEt
OH
OMe
OCF3
OCH2CN
F
Cl
CF3
CN
Et
i-Pro
CH2OH
CH2N(Me)2
Morpholinomethyl
OCH2CN
CN
CN
CN
CN
CN
CN
CN
CN
CN
CN
CN
CN
CN
CN
12
65
13
1
2
22
8
3
21
9
0.39
0.38
0.4
0.41
0.41
0.4
5.42
5.53
5.9
2
3
4
5
6
—
—
N
CH
CCN
CCN
8
1720
561
12
0.32
0.3
0.33
0.39
0.31
2.03
2.57
3.28
5.42
1.7
OH
OH
OH
OCH3
9
6.0
10
11
12
13
14
15
16
17
18
19
20
7.66
4.88
5.02
4.99
5.62
4.67
4.69
5.67
4.68
5.1
315
a
0.42
0.4
Results are mean of at least two experiments. Experimental errors within 20% of
value.
b
0.39
0.41
0.41
0.35
0.32
0.31
0.31
Calculated as ꢀRTln(IC50)/HAC.
c
Calculated as pIC50–logD.
3
108
144
67
and modifications (heteroatoms, substituents) to this ring were not
tolerated (data not shown). The hydroxyl group of 5 is exposed to
bulk solvent and not involved in any direct interactions with the
protein, and thus provides a handle to modulate the DMPK and
physical chemical properties of the series. This may also explain
the drop in potency observed between compound 5 (Table 1,
R1 = OH) and compound 6 (Table 1, R1 = OCH3).
CH3 456
4.1
a
Results are mean of at least two experiments. Experimental errors within 20% of
value.
b
Calculated as ꢀRTln(IC50)/HAC.
c
Calculated as pIC50–logD.
Compound 9 displayed remarkable selectivity over other PDEs
(PDE2A IC50
: 2.1 lM; PDE3 IC50: >25 lM and PDE4B IC50:
7
l
M).15 General off-target selectivity was also evaluated in silico,
using a ligand-based similarity method.16 This revealed the struc-
tural similarity between 11 and brequinar, a slow binding and
potent inhibitor of human dihydrorotate dehydrogenase (DHODH)
with reported toxic effects in clinical trial.17 As a result, it was
inferred that the developed series might have a propensity for
inhibiting DHODH, and this was confirmed experimentally
(DHODH IC50: 200 nM).18 While selectivity between DHODH and
PDE10A needs to be properly positioned in the context of pharma-
cokinetic–pharmacodynamic (PK–PD) relationships and risk-bene-
fit balance, we were nevertheless interested to evaluate whether
improvements could have been gained in the present series.
According to a recently published structural complex between
DHODH and a brequinar analog,19 and assuming a similar binding
for 9, modification of its distal phenyl ring could disrupt the inter-
action with DHODH. We thus opted to introduce polarity through
different heterocyclic systems and prepared compounds 21–29,
following Scheme 2. The PDE10A and DHODH inhibition results
for this set are summarized in Table 3.
The introduction of polar groups in the distal phenyl region was
predicted to have a detrimental effect on DHODH inhibition due to
the electrostatic repulsion with the backbone carbonyl moiety of
L67.19 Accordingly, the pyridine (21–25), pyrimidine (26) and pyra-
zole (27–29) derivatives evaluated here displayed a reduced DHO-
DH inhibition when compared to the phenyl analog 9, with most
of them showing no DHODH inhibition at the highest tested concen-
tration (Table 3). Unfortunately, this normally translated to reduced
PDE10A potency as well (e.g. 22, 26 and 27, Table 3). Increasing the
volume of the ring with lipophilic substituents enhanced both
PDE10A and DHODH inhibition (cf. 24 and 23; 28 and 29, Table 3).
Overall, 24 offered the best compromise, with excellent PDE10A
inhibition (IC50: 1 nM) and improved selectivity over DHODH
(1300ꢁ as judged by data from in vitro enzyme assays).20 Since
compounds are competitive towards PDE10A and based on their
similarity to brequinar likely to be competitive versus ubiquinone,
a substrate of DHODH, as well as intracellular concentration of
Figure 2. X-ray structure of compound 5 with PDE10A.
ubiquinone are unknown, absolute selectivity values in a physiolog-
ical setting cannot be predicted. The selectivity data presented in
Table 3 are thus best approximations. Additionally, 24 maintained
good selectivity (>2000ꢁ) over PDE2A, PDE3 and PDE4B.
The typical synthesis of compounds 5–29 is outlined in
Scheme 2. Common intermediate 32 was synthesized by condensa-
tion of azaisatoic anhydride 30 and bromobenzoyl–acetonitrile 31.
Suzuki coupling of aryl bromide 32 with arylboronic acids or esters
yielded the final 4-hydroxy-1,6-naphthyridine-3-carbonitriles.
When the desired boronic acid was not available, bromide 32
was subjected to palladium-catalyzed boronation with bis(pinaco-
lato)diboron, followed by Suzuki coupling with an aryl iodide in a
one-pot fashion. For the synthesis of compound 21 and 24 please
see Supplementary data. Methyl ether 6 was synthesized from
33, via chlorination with POCl3 and subsequent methanolysis.

U. Bauer et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1944–1948
1947
Table 3
PDE10A potency and selectivity over DHODH for compounds 9, 21–29
N
HO
1
R
N
N
a
a
Compound
R1
PDE10A IC50 (nM)
DHODH IC50 (nM)
DHODH/PDE10A selectivity
F
F
F
O
9
1
200
200ꢁ
*
*
*
21
22
N
N
52
97
>9900
5700
>190ꢁ
59ꢁ
N
*
*
23
24
27
1
>9900
1300
>367ꢁ
1300ꢁ
N
N
Cl
*
*
25
26
31
6300
203ꢁ
>13ꢁ
Cl
N
746
>9900
N
H
N
27
28
1770
608
>9900
>9900
>6ꢁ
*
*
N
N
N
>16ꢁ
N
N
29
13
1100
85ꢁ
*
a
Results are mean of at least two experiments. Experimental errors within 20% of value.
2. Kehler, J.; Kilburn, J. P. Expert Opin. Ther. Pat. 2009, 19, 1715.
3. Siuciak, J. A. CNS Drugs 2008, 22, 983.
4. Grauer, S. M.; Pulito, V. L.; Navarra, R. L.; Kelly, M. P.; Kelley, C.; Graf, R.; Langen,
B.; Logue, S.; Brennan, J.; Jiang, L.; Charych, E.; Egerland, U.; Liu, F.; Marquis, K.
L.; Malamas, M.; Hage, T.; Comery, T. A.; Brandon, N. J. J. Pharmacol. Exp. Ther.
2009, 331, 574.
5. Rodefer, J. S.; Murphy, E. R.; Baxter, M. G. Eur. J. Neurosci. 2005, 21, 1070.
6. Siuciak, J. A.; McCarthy, S. A.; Chapin, D. S.; Fujiwara, R. A.; James, L. C.;
Williams, R. D.; Stock, J. L.; McNeish, J. D.; Strick, C. A.; Menniti, F. S.; Schmidt, C.
J. Neuropharmacology 2006, 51, 374.
7. Cantin, L.-D.; Magnuson, S.; Gunn, D.; Barucci, N.; Breuhaus, M.; Bullock, W. H.;
Burke, J.; Claus, T. H.; Daly, M.; DeCarr, L.; Gore-Willse, A.; Hoover-Litty, H.;
Kumarasinghe, E. S.; Li, Y.; Liang, S. X.; Livingston, J. N.; Lowinger, T.;
MacDougall, M.; Ogutu, H. O.; Olague, A.; Ott-Morgan, R.; Schoenleber, R. W.;
Tersteegen, A.; Wickens, P.; Zhang, Z.; Zhu, J.; Zhu, L.; Sweet, L. Bioorg. Med.
Chem. Lett. 2007, 17, 2869.
In summary, initial scaffold hopping from 2 with the aim to re-
duce lipophilicity and potential risk for reactive metabolites affor-
ded 4-hydroxy-1,6-naphthyridine-3-carbonitriles as novel and
potent PDE10A inhibitors. Structure-based chemical modification
of the discovered chemotype served to further improve selectivity
over DHODH, laying the foundation for future optimization efforts.
Acknowledgments
We thank Anna Tigerström for the protein purification and crys-
tallisation of the enzyme-inhibitor complex, and Kenth Hallberg
for assistance in solving the X-ray structure of PDE10A.
8. Black, S.C. WO2005/120514, 2005.
9. Verhoest, P. R.; Chapin, D. S.; Corman, M.; Fonseca, K.; Harms, J. F.; Hou, X.;
Marr, E. S.; Menniti, F. S.; Nelson, F.; O’Connor, R.; Pandit, J.; Proulx-LaFrance,
C.; Schmidt, A. W.; Schmidt, C. J.; Suiciak, J. A.; Liras, S. J. Med. Chem. 2009, 52,
5188.
10. Hofgen, N.; Stange, H.; Schindler, R.; Lankau, H. J.; Grunwald, C.; Langen, B.;
Egerland, U.; Tremmel, P.; Pangalos, M. N.; Marquis, K. L.; Hage, T.; Harrison, B.
L.; Malamas, M. S.; Brandon, N. J.; Kronbach, T. J. Med. Chem. 2010, 53, 4399.
11. Law, R.; Barker, O.; Barker, J. J.; Hesterkamp, T.; Godemann, R.; Andersen, O.;
Fryatt, T.; Courtney, S.; Hallett, D.; Whittaker, M. J. Comput. Aided Mol. Des.
2009, 23, 459–473.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2012.01.046.
References and notes
1. Chappie, T.; Humphrey, J.; Menniti, F.; Schmidt, C. Curr. Opin. Drug Discov Dev.
2009, 12, 458.
12. PDE10 enzymatic assay: Details can be found in the Supplementary data.

1948
U. Bauer et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1944–1948
13. Andersen, O. A.; Schonfeld, D. L.; Toogood-Johnson, I.; Felicetti, B.; Albrecht, C.;
Fryatt, T.; Whittaker, M.; Hallett, D. Barker, Acta Crystallogr., Sect D 2009, 65,
872.
17. Burris, H. A., III; Raymond, E.; Awada, A.; Kuhn, J. G.; O’Rourke, T. J.; Brentzel, J.;
Lynch, W.; King, S. Y.; Brown, T. D.; von Hoff, D. D. InVest. New Drugs 1998, 16,
19.
14. Katritzky, A. R.; Lagowski, J. M. Adv. Heterocycl. Chem. 1963, 1, 311.
15. PDE2A, PDE3 and PDE4B enzymatic assay. Details can be found in the
Supplementary data.
18. DHODH assay: Details can be found in the supporting information and Lit 20.
19. Walse, B.; Dufe, V. T.; Svensson, B.; Fritzson, I.; Dahlberg, L.; Khairoullina, A.;
Wellmar, U.; Al-Karadaghi, S. Biochemistry 2008, 47, 8929.
16. Mestres, J.; Martin-Couce, L.; Gregori-Puigjane, E. J. Chem. Inf. Model 2006, 46,
2725.
20. Knecht, W.; Henseling, J.; Löffler, M. Chem. Biol. Interact. 2000, 124, 61.