Identification of a Novel 1,2,3,4-Tetrahydrobenzo[b][1,6]naphthyridine Analogue as a Potent Phosphodiesterase 5 Inhibitor with Improved Aqueous Solubility for the Treatment of Alzheimer's Disease

Article abstract of DOI:10.1021/acs.jmedchem.7b00979

Phosphodiesterase 5 (PDE5) hydrolyzes cyclic guanosine monophosphate (cGMP) leading to increased levels of the cAMP response element binding protein (CREB), a transcriptional factor involved with learning and memory processes. We previously reported poten

Full text of DOI:10.1021/acs.jmedchem.7b00979

Subscriber access provided by University of Virginia Libraries & VIVA (Virtual Library of Virginia)
Article
Identification of a Novel 1,2,3,4-tetrahydrobenzo[b][1,6]naphthyridine
Analog as a Potent Phosphodiesterase 5 Inhibitor with Improved
Aqueous Solubility for the Treatment of Alzheimer’s Disease
Jole Fiorito, Jeremie Vendome, Faisal Saeed, Agnieszka Staniszewski, Hong
Zhang, Shijun Yan, Shi-Xian Deng, Ottavio Arancio, and Donald W. Landry
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b00979 • Publication Date (Web): 06 Oct 2017
Downloaded from http://pubs.acs.org on October 7, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Identification of a Novel 1,2,3,4ꢀ
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
tetrahydrobenzo[b][1,6]naphthyridine Analog as a
Potent Phosphodiesterase 5 Inhibitor with Improved
Aqueous Solubility for the Treatment of Alzheimer’s
Disease.
‡
,#
ǁ,†,#
‡,†
‡
‡
Jole Fiorito, Jeremie Vendome, Faisal Saeed, Agnieszka Staniszewski, Hong Zhang,
‡,†
§
‡,
§,
Shijun Yan, Shi-Xian Deng, Ottavio Arancio, * Donald W. Landry *
‡
Taub Institute for Research of Alzheimer’s Disease and the Aging Brain, Columbia University,
630 W 168th Street, New York, NY 10032.
ǁ
Department of Systems Biology, Columbia University Medical Center, New York, NY 10032.
§
Department of Medicine, Columbia University, 650 W 168th Street, New York, NY 10032.
KEYWORDS. PDE5, PDE5 inhibitors, Alzheimer’s disease, Naphthyridines, Longꢀterm
Potentiation, Learning, Memory, In silico Docking.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ABSTRACT
Phosphodiesterase 5 (PDE5) hydrolyzes cyclic guanosine monophosphate (cGMP) leading to
increased levels of cAMP response element binding protein (CREB), a transcriptional factor
involved with learning and memory processes. We previously reported potent quinolineꢀbased
PDE5 inhibitors (PDE5Is) for the treatment of Alzheimer’s disease (AD). However, the low
aqueous solubility rendered them undesirable drug candidates. Here we report a series of novel
PDE5Is with two new scaffolds, 1,2,3,4ꢀtetrahydrobenzo[b][1,6]naphthyridine and 2,3ꢀdihydroꢀ
1Hꢀpyrrolo[3,4ꢀb]quinolinꢀ1ꢀone. Among them, compound 6c, 2ꢀacetylꢀ10ꢀ((3ꢀchloroꢀ4ꢀ
methoxybenzyl)amino)ꢀ1,2,3,4ꢀtetrahydrobenzo[b][1,6]naphthyridineꢀ8ꢀcarbonitrile, the most
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
potent compound, has excellent in vitro IC (0.056 nM) and improved aqueous solubility as well
50
as good efficacy in a mouse model of AD. Furthermore, we are proposing two plausible binding
modes obtained through in silico docking, which provide insights into the structural basis of the
activity of the two series of compounds reported herein.
INTRODUCTION
Alzheimer’s disease (AD) is the most common type of dementia and the sixthꢀleading cause of
death in the United States. AD is a neurodegenerative disorder that compromises the ability of
thinking, memory, and behavior. Senile plaques and neurofibrillary tangles, comprised of
amyloidꢀbeta (Aβ) peptide and microtubule binding protein tau, respectively, are the two main
pathological hallmarks of AD. To date, no effective therapies have been developed for AD and
1
,2
only symptomatic treatments are available. Therefore, major efforts are ongoing to find better
ways to prevent the progression of this disease. Several biological pathways have been
considered for developing valid therapeutic approaches, such as Aβ targeted strategies, tauꢀbased
ACS Paragon Plus Environment

Page 3 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
therapies, oxidative stress reduction, antiꢀinflammatory drugs, and therapies targeting second
1
, 2
messenger systems that are impaired by either Aβ or tau. However, drugs that entered the
clinical trials, e.g. most recently solanezumab, bapineuzumab, and aducanumab, which are
humanized anti–Aβ monoclonal antibodies, have failed to improve clinical outcomes in patients
or shown to induce adverse effects, e.g. vasogenic edema, making the Aβꢀremoving therapy a
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3
ꢀ5
doubtful strategy to pursue for the treatment of AD.
Here we have focused our attention on therapies targeting the nitric oxide (NO) signaling
6
pathway because of its relevance to learning and memory processes. Specifically, NO
stimulates soluble guanylate cyclase (sGC), resulting in increased levels of second messenger
cyclic guanosine monophosphate (cGMP) and activation of cGMPꢀdependent protein kinase
(PKG), which is important for promoting the phosphorylation of cyclic adenosine
7
, 8
monophosphate (cAMP) response element binding protein (CREB) at its Ser133.
Phosphorylation of CREB has been recognized as a crucial event for gene transcription during
synaptic plasticity, which is one of the fundamental processes underlying learning and memory.
9
ꢀ
1
1
Targeting the NO/cGMP/PKG/CREB signaling pathway can be achieved through the
inhibition of phosphodiesterases (PDEs). PDEs are a superfamily of enzymes consisting of
eleven subfamilies (PDE1ꢀ11) that hydrolyze the two nucleotides cGMP and cAMP into their
inactive forms. One fundamental distinction between the subfamilies is based on their different
affinity for the two substrates. PDE5, 6, and 9 are cGMPꢀspecific enzymes, PDE4, 7, and 8 are
cAMPꢀspecific enzymes, while the remaining enzymes are dualꢀsubstrate PDEs, having affinity
1
2
for both cyclic nucleotides. This superfamily of enzymes have been studied as targets for the
treatment of various diseases, including erectile dysfunction, pulmonary arterial hypertension,
1
2ꢀ15
cancer, schizophrenia, and AD.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Over the past years, we have been interested in elucidating the involvement of the
1
6
NO/cGMP/PKG/CREB signaling pathway in AD; we have demonstrated that the inhibition of
PDE5, which specifically catalyzes the hydrolysis of cGMP therefore reducing PKG activity,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
6, 17, 18
leads to an improvement of learning and memory in mouse models of AD.
Previous
research in our laboratories led to a novel series of quinoline inhibitors of PDE5 (PDE5Is)
represented by compound 1 (Figure 1). Compound 1 is a potent and selective PDE5 inhibitor as
determined by an in vitro enzymatic assay [IC (PDE5) = 0.270 nM, IC (PDE6) = 0.339 µM].
5
0
50
When dosed at 3 mg/kg, elevation of cGMP in the hippocampus of mice was observed.
Moreover, compound 1 rescues learning and memory as shown by a pharmacological treatment
1
8
of two different AD mouse models. However, a turbidimetric (kinetic) solubility screening
revealed that compound 1 has low water solubility (Table 2). This poor physicochemical
property precluded the advancement of 1 into additional preclinical studies.
In the course of our drug discovery program we have also discovered other classes of
heterocyclic PDE5Is, deriving from compound 1 and represented by scaffold A and B (Figure 1).
The two scaffolds were designed by locking the rotatable bonds of the hydroxymethyl group of
compound 1 into a ring. By rigidifying the conformation of the parent compound we sought to
improve the potency and selectivity of the new molecules. We then decided to investigate these
new chemical entities with the aim to find a suitable lead compound. Herein, we report the
design and synthesis of 1,2,3,4ꢀtetrahydrobenzo[b][1,6]naphthyridine (scaffold A) and 2,3ꢀ
dihydroꢀ1Hꢀpyrrolo[3,4ꢀb]quinolinꢀ1ꢀone (scaffold B) ligands and their enzymatic activity.
Notably, we could identify a PDE5 inhibitor, compound 6c, that displays enhanced potency,
retained selectivity, and improved water solubility with respect to compound 1, as well as in vivo
19
pharmacological efficacy in the APP/PS1 double transgenic mice that manifest aberrant
ACS Paragon Plus Environment

Page 5 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
K670N/M671L
M146L
accumulation of human Aβ driven by mutations in APP695
and PS1
.
Additionally, we identified two plausible binding modes in the same pocket using in silico
docking, which provides insights into the structural basis of the activity of the new series of
compounds reported here. The validation of one of these binding modes as well as its usage for
further optimization of the herein reported compounds via a structureꢀbased approach are to be
pursued in a following study.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
OMe
OMe
OMe
Cl
Cl
N
Cl
HN
HN
HN
NC
O
OH
NC
R
NC
N R
N
N
N
1
A
B
Figure 1. Structure of compound 1 and new PDE5Is: 1,2,3,4ꢀ
tetrahydrobenzo[b][1,6]naphthyridine (scaffold A) and 2,3ꢀdihydroꢀ1Hꢀpyrrolo[3,4ꢀb]quinolinꢀ1ꢀ
one (scaffold B).
RESULTS AND DISCUSSION
Chemistry. As first step, we targeted the synthesis of a set of tricyclic analogs 6aꢀg (scheme 1),
1
0 (scheme 2), and 13aꢀd (scheme 3). As show in scheme 1, 2ꢀaminoꢀ5ꢀcyanobenzoic acid 4
2
2
was prepared via the cyanation of methyl bromobenzoate 2 with copper cyanide followed by
hydrolysis under basic conditions of ester 3. For the synthesis of the 1,2,3,4ꢀ
tetrahydrobenzo[b][1,6]naphthyridine core, a condensation reaction was used to synthesize
2
3
intermediates 5aꢀd. Specifically, the reaction of 4 with the appropriate Nꢀsubstituted piperidinꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
4
ꢀone in POCl gave 5aꢀd in moderate yields (21ꢀ55%). Then, 3ꢀchloroꢀ4ꢀmethoxibenzylamine
3
was introduced on the tricyclic ring of 5aꢀc via nucleophilic substitution to obtain the final
compound 6aꢀc. Similarly, compounds 6dꢀf were obtained by using the corresponding
commercially available benzylamines (see Experimental Section for details). For the synthesis
of compound 6g, (3ꢀchloroꢀ5ꢀfluoroꢀ4ꢀmethoxyphenyl) methanamine hydrochloride 10 was
prepared starting from the benzaldehyde 7, which was initially reduced into benzyl alcohol 8 via
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a reduction reaction using NaBH . Subsequently, benzylamine 10 was obtained by following the
4
Gabriel synthesis procedure (scheme 1). The synthetic route followed to prepare compound 14
was slightly different than the one used for synthesizing 6aꢀg (scheme 2). Intermediate 5d was
first deꢀalkylated by using αꢀchloroethyl chloroformate (ACEꢀCl); subsequently, after protecting
the aliphatic nitrogen of 11 with a Boc group, the synthesis of 13 was accomplished via the same
nucleophilic substitution reaction conditions used for compounds 6aꢀg. Finally, removal of the
Boc group gave the desired analog 14.
a
Scheme 1. Synthesis of compounds 6aꢀg
ACS Paragon Plus Environment

Page 7 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
Reagents and conditions: (i) CuCN, NMP, 200 °C, 4h; (ii) EtOH/KOH 1M, 50 °C, 2h; (iii) 1ꢀ
substituted piperidinꢀ4ꢀone, POCl , 60 °C, 6h; (iv) Appropriate benzylamine, TEA, NaI, NMP,
30 °C, overnight; (v) NaBH , THF, r.t., overnight; (vi) phthalimide, DIAD, PPh , THF, r.t.,
4h; (vii) a) N H , MeOH, reflux, overnight; b) HCl/Et O
3
1
2
4
3
2
4
2
a
Scheme 2. Synthesis of compound 14
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
Reagents and conditions: (i) a) ACEꢀCl, ClCH CH Cl, reflux; b) MeOH, reflux, 2h; (ii) Boc O,
2
2
2
CH Cl; (iii) 3ꢀchloroꢀ4ꢀmethoxybenzylamine, TEA, NaI, NMP, 130 °C, overnight; (iv)
Et O/HCl 2.0M, CH Cl ,1,4ꢀdioxane, r.t., overnight.
2
2
2
2
We also sought to introduce other functionalities at the naphthyridine 8ꢀposition. In the course
of this effort, analogs 17aꢀd were synthesized using the same synthetic route of 6aꢀg but staring
form commercially available materials 15aꢀc, as shown in scheme 3.
a
Scheme 3. Synthesis of compounds 17aꢀd
a
Reagents and conditions: (i) 1ꢀsubstituted piperidinꢀ4ꢀone, POCl , 60 °C, 6h; (ii) benzylamine,
3
TEA, NaI, NMP, 130 °C, overnight.
1
Of note, during the chemical characterization of the naphthyridineꢀbased compounds by H
1
3
NMR and C NMR spectroscopy, the appearance of extra peaks was observed for derivatives
bearing the acetyl group at the naphthyridine 2ꢀposition. This phenomenon was attributed to the
ACS Paragon Plus Environment

Page 9 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
presence of equilibrating rotamers, which were confirmed by using additional spectroscopic
2
4
techniques. 1D NOESY experiments allowed us to identify the protons undergoing chemical
exchange and to give a clear interpretation of the NMR spectra. A detailed description of the
spectroscopic profile of compounds 5c (Fig. S1 and S2) and 6c (Fig. S3 and S4) as representative
spectra is reported in Supporting Information. In addition, as already reported for our previous
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
8
series of quinolineꢀbased compounds, a spinꢀspin coupling between NH protons and their
vicinal aliphatic CH protons was observed for most of the naphthyridineꢀbased compounds and
2
confirmed by D O exchange. The presence of clear split signals (doublets and triplets) was also
2
depending on the solvent used to record the NMR experiment.
The general synthesis of 1Hꢀpyrroloquinolinone analogs 24aꢀf is outlined in scheme 4.
Starting with 5ꢀsubstituted 2ꢀaminobenzoic acids 18aꢀc, 2Hꢀbenzo[d][1,3]oxazineꢀ2,4(1H)ꢀ
diones 19aꢀc were prepared from the reaction with triphosgene. Readily accessible isatoic
anhydride 19d and 5ꢀchloroisatoic anhydride 19e were utilized for the synthesis of intermediates
20d and 20e. Treatment of 19aꢀe with methyl acetoacetate yielded the desired methyl 4ꢀoxoꢀ1,4ꢀ
dihydroquinolineꢀ3ꢀcarboxylate derivatives 20aꢀe, which were subjected to a chlorination
reaction to obtain the 2ꢀmethylquinoline derivatives 21aꢀe. Next, bromination of the methyl
group followed by cyclization with ethylamine in EtOH provided 2,3ꢀdihydroꢀ1Hꢀpyrrolo[3,4ꢀ
b]quinolinꢀ1ꢀone derivatives 22aꢀf. Introduction of 3ꢀchloroꢀ4ꢀmethoxibenzylamine on the
tricyclic template via nucleophilic substitution gave the final compound 24aꢀf.
a
Scheme 4. Synthesis of compounds 24aꢀf
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
Reagents and conditions: (i) triphosgene, 1,4ꢀdioxane, 90°C, 2h; (ii) methyl acetoacetate, NaH,
DMA, 120°C, 20min; (iii) POCl , 110°C, 30min; (iv) NBS, AIBN, CCl , reflux, 4h; (v)
ethylamine or tertꢀbutylamine, EtOH, reflux, 2h; (vi) 3ꢀchloroꢀ4ꢀmethoxybenzylamine HCl,
DIPEA, nꢀpropanol, 90°C, 2h.
3
4
Inhibition of PDE5 and selectivity versus PDE6. Both naphthyridine (6aꢀg, 14, and 17aꢀd)
and 1Hꢀpyrroloquinolinone (24aꢀf) analogs were tested on human recombinant PDE5A1 in a
fluorescent enzymatic assay performed by BPS Bioscience (Table 1). With respect to 1, both
scaffold A and B retained potent PDE5 activities. While compounds 6b, 6g, and 24c showed
potency comparable to that one of compound 1, compounds 6c and 24a exhibited the highest
inhibitory activity in the picomolar range. On the contrary, compounds 6d, 17a and 24f
displayed the lowest potency, which was still in the nanomolar range. The most potent PDE5Is
6
c and 24a were also evaluated for selectivity versus PDE6C isozyme, which shares a high
similarity in the catalytic site with PDE5 with respect to the molecular binding and therefore
2
5
represents a major obstacle in developing highly selective PDE5Is. Although both scaffolds
were well tolerated, it appears that a larger aliphatic ring fused to the quinoline moiety
contributes to increasing the PDE5 selectivity versus PDE6, as compound 6c bearing a 6ꢀ
membered aliphatic ring showed to be ~500ꢀfold more potent against PDE5, while the 5ꢀ
membered fused ring in compound 24a leads to a selectivity for PDE5 of ~100ꢀfold compared to
PDE6.
ACS Paragon Plus Environment

Page 11 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 1. PDE5 and PDE6 activity of naphthyridine and 1Hꢀpyrroloquinolinone analogs.
PDE5A1
PDE6C
PDE5A1
PDE6C
Cmpd
1
a
a
Cmpd
a
a
IC (nM)
IC (nM)
IC (nM)
IC (nM)
5
0
50
50
50
b
b
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
0.27
0.59 ± 0.3
0.19 ± 0.01
339
17b
17c
17d
24a
24b
24c
24d
24e
24f
45.2 ± 2.3
3.1 ± 0.1
21.6 ± 1.0
n.d.
n.d.
n.d.
6
6
6
6
6
6
6
1
1
a
b
c
n.d.
n.d.
0.056 ± 0.01 30.1 ± 1.1
0.059 ± 0.005 6.6 ± 0.3
d
e
337 ± 30.5
57 ± 6.8
n.d.
n.d.
n.d.
3.8 ± 0.3
n.d.
n.d.
n.d.
n.d.
n.d.
0.29 ± 0.018
64 ± 5.4
f
1.5 ± 0.2
g
4
7a
0.32 ± 0.016 n.d.
1.7 ± 0.1
1.55 ± 0.14
110 ± 3.3
n.d.
n.d.
333 ± 26.3
a
Enzymatic activity was performed by BPS Bioscience using recombinant human enzymes;
b
As reported in reference 18. Values are reported as average ± S.E. (standard error).
Computational study. In order to better understand the structural basis of PDE5 inhibition by
the new inhibitors reported here, we have conducted a computational study to identify a binding
mode allowing to account for the experimental structureꢀactivity relationship observed within the
naphthyridine and the 1Hꢀpyrroloquinolinone series (Fig. 1, scaffold A and B respectively).
First, we performed unconstrained in silico docking of compounds 6c and 24a, the most potent
compounds of the naphthyridine and the 1Hꢀpyrroloquinolinone series respectively, into the
putative cGMP pocket of PDE5A1 (see methods in Experimental Section). We targeted this
pocket based on the fact that multiple complex crystal structures show that diverse PDE5 small
molecule inhibitors consistently bind in it (e.g., pdb codes 1UHO, 3BJC, 3B2R, 3HDZ, 3SIE,
4G2W, 4G2Y, 4I9Z, 4OEW and 4OEX), showing that it has highly druggable characteristics
(Supporting Information, Fig. S5A). The docking pose with the best docking score obtained for
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
compound 6c (Fig. 2A and 2B) and compound 24a (Supporting Information, Fig. S7A) were
very consistent (see Supporting Information, Fig. S6A). Interestingly, a second binding mode
where the tricyclic ring is flipped compared to the best docking pose described above, was also
consistently found for both compounds (Fig. 2C and Supporting Information, Fig. S7B). Of
note, the two alternative binding poses occupy approximately the same space within the cGMP
pocket for compounds 6c and 24a, as the tricyclic ring is rotated by ~180° around its central axis
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(Supporting Information, Fig. S6C and S6D). Similarly to most PDE5 inhibitors, both
2
+
2+
compounds lack interaction with the structural Mg and Zn ions present in the enzymatic
pocket.
ACS Paragon Plus Environment

Page 13 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
A Compound 6c
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
8
0°
Mg2+
Zn2+
F787
B
F787
I813
Q817
M816
Q817
I813
L804
L804
V782
V782
Y612
8
0°
M816
F786
F786
Q775
Y612
Zn2+
L765
Compound 6c
Mg2+
F820
F820
F787
I768
L725
L765
L725
F787
C
L804
Q817
V782
I813
I813
Q817
L804
M816
8
0°
V782
F786
M816
F786
Y612
Y612
I768
Zn2+
L765
Compound 6c
Mg2+
F820
F820
L725
L725
L765
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Figure 2. Docked binding modes of compound 6c into PDE5 cGMP pocket. (A) Overview of
PDE5 with compound 6c docked in the cGMP pocket in its best docking pose. (B) Closeꢀup
view of the best docking pose for compound 6c. This pose is very similar to the best docking
pose found for compound 24a (Supporting Information, Fig S6A). Note that for both
compounds, postꢀdocking minimization led to a significant change in Q775 side chain
orientation compared to the PDB structure used for docking (3TGG). The refined side chain
orientation, which optimizes the geometry of the hydrogen bond with the nitrile group of the two
compounds, is consistent with an alternative side chain orientations observed for residue Q775 in
other PDE5 crystal structures (Supporting Information, Fig. S5B and C). (C) Closeꢀup view of
the alternative docking pose found for compound 6c. A very similar alternative docking pose
was found for compound 24a (Supporting Information, Fig S7B). Dashed lines in B and C
represent intermolecular hydrogen bonds.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Best docking pose
Alternative docking pose
(
Proposed binding mode)
GLN
8
17
GLN
75
7
GLN
17
8
PHE
20
8
PHE
8
20
ACS Paragon Plus Environment

Page 15 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 3. Twoꢀdimensional interaction diagrams for compound 6c. The best docking pose
(left) as well as the alternative docking pose (right) are represented. The legend for the residue
types and the types of interactions is indicated at the bottom of the figure. The twoꢀdimensional
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
interaction diagrams for compound 24a are represented in Supporting Information, Fig. S7.
In order to compare these two binding modes and potentially provide further validation for one
of them, we explored whether they could account for the structureꢀactivity relationship observed
within the naphthyridine and the 1Hꢀpyrroloquinolinone series. The interactions of compound
6c with PDE5 for both binding modes are summarized in Figure 3. Expectedly, the interactions
of compound 24a with PDE5 are very similar for both binding modes (Supporting Information,
Fig. S7). Noticeably, these interactions can provide a qualitative explanation for some of the
structureꢀactivity relationships observed for the two series of compounds. For the best docking
pose, in particular, the hydrogen bond between the nitrile group of compounds 6c or 24a and the
amide group of residue Q775, located at the deep end of the binding pocket, can explain the
significant impact of chemical changes at this position on potency (compare compounds 6c and
17aꢀc for the naphthyridine series and compounds 24aꢀd and 24e for the 1Hꢀpyrroloquinolinone
series). Among the groups that can be compared for this position, the best potency is obtained
with a nitrile group (IC ~ 0.6 nM for both compound 6c and 24a) while the weakest potency,
5
0
within both series, is obtained in the absence of any acceptor (IC of 110 nM and 64 nM for
5
0
compounds 17a and 24d, respectively). The presence of a weaker acceptor (e.g., F, Cl or Br in
compounds 24c, 24e and 17c respectively) and/or of a shorter linker between the acceptor and
the tricyclic ring core (e.g., ꢀOMe in compounds 17b and 24b respectively), which places the
acceptor further away from the Q775 donor, result in intermediate potencies (Table 1).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
On the other side of the tricyclic ring, for the naphthyridine series, the water bridge formed
between the acetyl group and M816 backbone in the best docking pose can also explain the
reduced potency with groups lacking good acceptor or donor (e.g., IC of 0.19 nM, 0.59 nM and
5
0
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1.55 nM with an ethyl group (compound 6b), methyl group (compound 6a) and hydrogen
(compound 14) respectively) (Table 1). This role of the acetyl group in the naphthyridine series
could however also be explained in the alternative docking pose, by the hydrogen bond formed
with residue Q817. In the alternative binding mode, however, the importance of the nitrile group
is more difficult to explain as it is mostly solvent exposed.
The two binding modes can also provide an explanation for the importance of the methoxy
group in paraꢀ position of the benzyl ring of compound 6c (IC ~ 0.056 nM for a methoxy group
50
in compound 6c compared to IC of 337, 57 and 1.5 nM for a trifluoromethyl, unsubstituted
5
0
pyridine and a single fluoride in compounds 6d, 6e and 6f, respectively). The benzyl ring
carrying the group occupies the same small pocket lined by F786, F787, L804, I813 and M816 in
both binding modes (Fig. 2). Performing molecular dynamics simulations, water clustering and
26, 27
statistical thermodynamic analysis using WaterMap
shows that in the absence of ligand, this
small pocket is likely to be occupied by several “highꢀenergy” water molecules (i.e., predicted by
WaterMap to be significantly more energetically stable in the bulk solvent than in the binding
pocket, where it however binds in order to avoid vacuum) (Supporting Information. Fig. S8).
Overlapping of a ligand with one of these highꢀenergy hydration sites by either a polar or nonꢀ
polar group allows at the same time to displaces the unstable water into the bulk solvent while
avoiding vacuum, and is energetically quite favorable. One of these highꢀenergy hydration sites
precisely overlaps with the position of the methoxy group for both docked binding modes of
compound 6c (Supporting Information, Fig. S8), while smaller groups like an unsubstituted
ACS Paragon Plus Environment

Page 17 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
pyridine or a fluoride (compounds 6e and 6f) would not be bulky enough to displace this water
molecule to the solvent. For the case of the trifluoromethyl substituent (IC of 337 nM for
50
compound 6d), the significantly reduced potency can be explained by the fact that while it is
certainly bulky enough to desolvate the pocket as well as the methoxy group, its significantly
larger size does not sterically fit as well in the pocket.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Overall, these qualitative explanations globally argue rather in favor of the best docking pose
relative to the alternative pose, although more work would be needed to further validate one of
these poses.
Physicochemical Properties. To determine the potential drugꢀlikeness of compounds 6c and
24a, their in silico physicochemical properties were evaluated and compared to those of
compound 1. As shown in Table 2, both compounds 6c and 24a obey the Lipinski’s rule of five
and fulfill the restrictions of topological polar surface area (tPSA) for drugs targeting the central
2
8
nervous system. Additionally, water solubility of compound 6c and 24a was assessed to establish
whether the new scaffolds showed increased solubility than compound 1. Kinetic solubility was
measured in phosphate buffered saline following an 18ꢀhour incubation period according to the procedure
described in the Experimental Section. Compound 24a was found to have a low kinetic solubility,
while compound 6c showed improved water solubility with respect to compounds 1 (Table 2).
Melting point of compound 6c, 24a, and 1 has been reported in table 2 as well. These findings
suggest that compound 6c represent a good candidate for further analyses. We then engaged
compound 6c in a set of experiments aimed at evaluating its pharmacological properties and
potential use in AD.
Table 2. Water solubility and physicochemical properties of 1, 6c, and 24a.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Water
Melting
point (°C)
Cmpds
1
MW
HBA
HBD
cLogP
3.42
3.41
3.66
<5
tPSA solubility
b
a
(µM)
393.87
420.90
406.87
<450
5
2
78.17
78.25
78.25
<90
<3.9
>500
3.1
171.0ꢀ173.0
165.5ꢀ167.0
207.0ꢀ209.0
6c
6
1
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
24a
6
1
RO5
<10
<5
MW: molecular weight, HBA: hydrogenꢀbond acceptor atoms, HBD: hydrogenꢀbond donor
atoms, clogP: calculated logarithm of the octanolꢀwater partition coefficient, tPSA: topological
polar surface area are calculated using http://www.molinspiration.com/cgiꢀbin/properties; RO5:
Lipinski’s rule of five; Calculated as turbidimetric (kinetic) solubility by Cyprotex US, LLC;
As an average of 2 independent measurements determined on MelꢀTemp II, Laboratory devices,
a
b
USA.
Effects of Compound 6c on Hippocampal cGMP Levels and Synaptic Plasticity. Based on
its potency and selectivity profile, as well as its favorable aqueous solubility, compound 6c was
moved into a functional assay aimed at evaluating its ability to increase cGMP levels in the
hippocampus of mice. Wildꢀtype (WT) mice were treated with compound 6c at two different
concentrations (3 and 10 mg/kg). Mice were subjected to foot shock and sacrificed 10, 60, and
1
80 sec afterward (Figure 4). These time points were chosen based on previous experiments in
1
8
which we observed a change in cGMP levels immediately after foot shock. Increased levels of
cGMP were observed at both concentrations and all of the 3 different time points. A doseꢀ
dependent increase of cGMP levels was detected in hippocampi collected at 10 and 60 sec after
foot shock, with a 1.5ꢀ and 1.7ꢀfold increase at 3 and 10 mg/kg, respectively, compared to
untreated mice for both time points. At 180 sec, a slight increase in cGMP levels was detected in
mice treated with 3mg/kg of 6c but it failed to reach statistical significance (p=0.097), while the
higher dose produced a 1.3ꢀfold increase of cGMP levels with respect to vehicleꢀtreated mice. A
compensatory cellular mechanism leading to the hydrolysis of cGMP by other PDEs could
ACS Paragon Plus Environment

Page 19 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
explain the reduced increase of cGMP levels at 180 sec with respect to 60 sec after foot shock for
mice treated with 10 mg/kg of compound 6c. However, longer time points or the employment of
inhibitors of other cGMPꢀhydrolyzing PDEs were not explored to verify this hypothesis. In any
case, the NO signal transduction cascade, involving cGMP, is sufficient to phosphorylate CREB
that leads to memory formation through transcription of memoryꢀrelated genes. Such a cellular
response is triggered by memoryꢀstimulating events, such as foot shock. Indeed, vehicleꢀtreated
mice subjected to foot shock (Fig. 4, veh) showed an increase of cGMP levels compared to
vehicleꢀtreated mice that did not receive foot shock (Fig. 4, Basal). Notably, compound 6c was
effective in elevating hippocampal cGMP levels, indirectly proving its ability to cross the bloodꢀ
brain barrier (BBB).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. cGMP levels in hippocampus of mice treated with compound 6c at 3 and 10
mg/kg via intraperitoneally. Values are the mean of duplicate determinations. Error bars show
S.E.M. (n=3 per group); *p<0.01.
To determine whether compound 6c was able to strengthen synaptic transmissions, we carried
out electrophysiological experiments to measure the compound’s effects on hippocampal longꢀ
term potentiation (LTP), a type of synaptic plasticity thought to underlie learning and memory
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
16, 29
and to be impaired by the elevation of Aβ.
Slices of hippocampus were collected from 3ꢀ4
month old APP/PS1 transgenic mice and WT littermate controls and were treated either with
vehicle or 6c (50 nM through the bath perfusion for 10 min prior to the tetanus) (Fig. 5A and B).
According to our previous results, vehicleꢀtreated APP/PS1 mice showed a reduction in LTP
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
0
compared to WT mice. Compound 6c rescued the defect in LTP; residual potentiation
recorded in slices from APP/PS1 mice treated with compound 6c (201.76% ±14.09) was
increased with respect to slices from APP/PS1 mice treated with vehicle (137.30% ±10.83) (Fig.
5
A and B). Compound 6c did not have any effect on WT slices, consisting with the possibility
that the inhibitor is acting on the LTP mechanisms underlying memory loss instead of a more
generalized effect (Fig. 5A and B). Thus, by acting on the NO/cGMP signaling pathway,
compound 6c reꢀestablished the normal increase of synaptic strength in the hippocampus of
APP/PS1 mice (Fig. 5A and B). A similar beneficial effect was observed when compound 6c
was injected to APP/PS1 mice from the age of 3ꢀ4 months for 3 weeks and hippocampal slices
were harvested at 6ꢀ7 months of age (Fig. 5C and D). LTP was restored in compoundꢀtreated
APP/PS1 mice (226.22% ±6.71) compared to nonꢀtreated transgenic mice (164.11% ±10.36)
(Fig. 5C and D), suggesting an enduring effect of 6c on potentiating synaptic plasticity. These
results corroborate our previous findings on the prolonged beneficial effect induced by the PDE5
inhibition on synaptic and cognitive abnormalities in APP/PS1 mice that persists beyond the
1
8, 17
administration of the inhibitor.
ACS Paragon Plus Environment

Page 21 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. Beneficial effect of compound 6c on synaptic deficits in APP/PS1 mice. Two hours
recording following tetanic stimulation of the Schaffer collateral fibers at the CA3ꢀCA1
connection. A) Slices were prepared from 3ꢀmonthꢀold WT and APP/PS1 mice treated with
vehicle or compound 6c at 50nM (acute perfusion, *p<0.01 and F(1,13)=15.57). B) Residual
potentiation recorded during the last 5 minutes of a 2 hr recording shown in panel A; C) Daily
treatment with compound 6c (3mg/kg, i.p.) for 3 weeks at the age of 3ꢀ4 months, slices were
recorded at 6ꢀ7 months of age; *p<0.01 and F(1,15)=27.0; D) Residual potentiation recorded
during the last 5 minutes of a 2 hr recording shown in panel C.
Plasma and Brain Levels and in vitro Microsomal Stability of Compound 6c. To further
validate compound 6c as a promising candidate for treating a central nervous disease, a
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
pharmacokinetic study was performed. We evaluated brain and plasma levels in mice after
intraperitoneal (i.p.) administration of compound 6c at 25 mg/kg at different time points (0, 5,
15, 30, 60, and 120 min). As shown in Table 3, peak plasma concentration of 6c occurred 0.25 h
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
after dosing and its plasma halfꢀlife was 1.07 h. At 0.25 h, concentration of 6c in the brain and
plasma were 492.25 ng/g and 1571.44 ng/mL, respectively. The logBB value was used to
evaluate BBB permeability; 6c showed the logBB values of −0.50, which was within the
3
1
established threshold that defines a compound as crossing the BBB (logBB ≥ −1). Therefore,
compound 6c demonstrated to cross the BBB.
Table 3. Plasma and brain concentrations, logBB, halfꢀlife, and in vitro microsomal stability of
compound 6c.
a
a
a
Microsomal
stability
Plasma
ng/mL)
Brain
b
Halfꢀlife
(h)
c
Sample
n=3
LogBB
(
(ng/g)
(
min)
.0 (HLM)
3.2 (MLM)
6
1571.44
492.25
−0.50
1.07
a
b
Values determined 0.25 h after administration; BB: ratio of the brain to plasma concentration;
c
Microsomal stability was performed by Absorption Systems.
Microsomal stability of compound 6c was measured in both mouse liver microsomes (MLM)
and human liver microsomes (HLM) (Table 3). Such short halfꢀlives were indicative of a rapid
metabolism of the compound by both mouse and human microsomes. Optimizing the metabolic
stability while maintaining potency is an effort to be pursued in a following study. The structural
insights gained through the computational study presented above will be leveraged in this
endeavor. A first possible metabolic liability to be further explored is the deꢀmethylation of the
methoxy group on the benzyl ring of compound 6c. If this metabolic liability is confirmed (for
example through microsomal stability profiling of compounds 6d, 6e and 6f), the binding mode
ACS Paragon Plus Environment

Page 23 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
proposed for compound 6c will be used to guide the choice of an isostere to replace the methoxy
group (i.e. bulky enough to desolvate the pocket while not too bulky to sterically fit well in it).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Behavioral Studies of Compound 6c. Despite analysis of microsomal stability suggested that
further improvement is needed, we wanted to verify if the new scaffold (compound 6c) retained
the capability of rescuing the memory loss in the APP/PS1 mouse model of AD. To this end,
two memory tasks fear conditioning (FC) and 2ꢀday radial arm water maze (RAWM) were
performed to assess associative and spatial memory, respectively (Fig. 6). APP/PS1 mice were
subjected to both tasks at the age of 3ꢀ4 month (Fig. 6, A and B), when they clearly show
3
0, 32
cognitive impairment.
In these experiments, both APP/PS1 and WT mice were treated with
6c (3 mg/kg, i.p.) daily for 3 weeks and compared to vehicleꢀtreated APP/PS1 and WT mice.
Percentage of freezing in the FC task (Fig. 6A) defines the animal response to an electric foot
shock (see Experimental Section for detailed methods). Figure 6A shows that vehicleꢀtreated
transgenic mice froze less (~15% freezing) compared to vehicleꢀ and compoundꢀtreated WT
mice as well as compoundꢀtreated APP/PS1 mice, indicating the inability of these animals to
associate the noxious stimulus with the environment where the stimulus occurs. Treatment with
6c, on the other hand, rescued contextual learning in APP/PS1 mice (~30% freezing) without
affecting memory in WT mice (~30% freezing). In the 2ꢀday RAMW (Fig. 6B), the number of
errors refers to the ability of the animal to reach the hidden platform, which is placed in one of
the 6 arms of a maze (see Experimental Section for detailed methods). Compared to vehicleꢀ
treated APP/PS1 (~3 errors), compoundꢀtreated APP/PS1 showed a fewer number of errors (~1
error) suggesting that compound 6c is able to rescue spatial working memory. As expected,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
vehicleꢀ and compoundꢀtreated WT mice performed well in the 2ꢀday RAWM as they show no
symptoms of memory impairment.
To investigate whether the beneficial effects of compound 6c on synaptic and cognitive
abnormalities in APP/PS1 mice was persisting beyond its administration, both WT and APP/PS1
mice that were treated for 3 weeks with 6c at the age of 3 months were assessed for contextual
and spatial memory when they were 6ꢀ7 month old (Fig. 6, C and D). In the contextual memory
experiments, transgenic mice treated with the compound froze ~40% of the time compared to
~15% in vehicleꢀtreated transgenic mice (Fig. 6C) and made ~1 error in the 2ꢀday RAWM test
compared to ~2.5 in vehicleꢀtreated transgenic mice (Fig. 6D). Thus, synaptic and cognitive
improvements persist beyond the administration of compound 6c. For all experiments, we
performed cued conditioning and visible platform tasks to verify that contextual memoryꢀ
independent mechanisms interfere with the interpretation of the results. None of these controls
showed differences among various groups of mice (see Supporting Information, Figures S9 and
S10).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment

Page 25 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. Beneficial effect of compound 6c on cognitive deficits in APP/PS1 mice. A) Daily
treatment with compound 6c (3mg/kg, i.p.) for 3 weeks at the age of 3ꢀ4 month reꢀestablished
normal freezing in a test for contextual fear memory. B) Daily treatment with compound 6c
(3mg/kg, i.p.) for 3 weeks at the age of 3ꢀ4 months reduced the number of errors with the 2ꢀday
radial arm water maze. C) Daily treatment with compound 6c (3mg/kg, i.p.) for 3 weeks at the
age of 3ꢀ4 months reꢀestablished normal freezing in a test for contextual fear memory when mice
were examined at 6ꢀ7 months of age. D) Daily treatment with compound 6c (3mg/kg, i.p.) for 3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
weeks at the age of 3ꢀ4 months reduced the number of errors with the 2ꢀday radial arm water
maze when mice were examined at 6ꢀ7 months of age. * p < 0.05
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
CONCLUSION
We reported an extensive study of new chemical entities with PDE5 inhibitory activity. In
particular, we reported the naphthyridine derivative 6c, which was found to have excellent PDE5
1
8
potency and selectivity, with improved water solubility with respect to the quinoline analog 1.
Despite its lack of in vitro metabolic stability, which suggests the need of further effort toward
the optimization of this drug property, physicochemical and pharmacological properties of
compound 6c were found to be optimal for a potential drug candidate.
Using in silico docking, we were furthermore able to identify two plausible binding modes and
provide structural insights into the structureꢀactivity relationship for the two series of compounds
reported here. Further validation of one of these binding modes is work for a following study
and should open the way to the structureꢀbased optimization of the compounds reported here.
We further showed that treatment with compound 6c increased cGMP levels in the hippocampus
of mice and improved learning and memory deficits in a mouse model of AD. These findings
strengthen the hypothesis that the NO/cGMP/PKG/CREB molecular pathway plays an important
role in learning and memory and that the inhibition of PDE5 represents a sound AD therapeutic
approach for either slowing down or preventing the progress of the disease.
EXPERIMENTAL SECTION
ACS Paragon Plus Environment

Page 27 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Chemistry. General Method. All solvents and chemicals were reagent grade. Unless otherwise
mentioned, all reagents and solvents were purchased from commercial vendors and used as
received. Silica gel chromatography was performed using glass columns packed with silica gel
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
13
(
200−400 mesh, Aldrich Chemical). H and C NMR spectra were recorded on an Agilentꢀ
NMRꢀvnmrs400 (400 MHz and 100 MHz, respectively) spectrometer and were determined in
chloroformꢀd, DMSOꢀd6, acetoneꢀd6 or methanolꢀd4 with tetramethylsilane (TMS) (0.00 ppm)
or solvent peaks as the internal reference. Chemical shifts (δ) are reported in ppm relative to the
reference signal, and coupling constant (J) values are reported in Hertz (Hz), multiplicity (s =
singlet, d = doublet, dd = double doublet, t = triplet, q = quartet, m = multiplet, br = broad). The
purity of all final compounds was ≥ 95 % and was determined by LC/MS at wavelengths 220
and 254 nm. Liquid chromatographyꢀmass spectrometry (LC/MS) was performed on two
different instruments, a Shimadzu 2010A and a Shimadzu 2020 UFLC mass spectrometer, using
a Waters Sunfire column (C18, 5ꢁm, 2.1 mm x 50 mm, a linear gradient from 5 % to 100 % B
over 15 min, then 100 % B for 2 min (A = 0.1 % formic acid + H O, B = 0.1 % formic acid +
2
CH CN), flow rate 0.2000 mL/min). HRMS (ESI) was performed by Furong Sun and associates
3
at the University of Illinois (UrbanaꢀChampaign). Thin layer chromatography (TLC) was
performed on EMD precoated silica gel 60 F254 plates, and spots were visualized with UV light
(
254 nm). All reagents and solvents were used as received from major commercial suppliers,
such as SigmaꢀAldrich, Fisher Scientific, and Alfa Aesar without further purification. All airꢀ or
moistureꢀsensitive reactions were run under an atmosphere of argon in ovenꢀdried glassware
unless otherwise noted.
General procedure A for the synthesis of 10ꢀchloroꢀ1,2,3,4ꢀ
tetrahydrobenzo[b][1,6]naphthyridine. Nꢀsubtituted piperidinꢀ4ꢀone (1 equiv) and 5ꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 64
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
substituted 2ꢀaminobenzoic acid (1 equiv) were heated to 60 C in POCl (which was used as a
3
reagent and a solvent) for 6h. The excess of POCl was evaporated off and the residue was
3
treated with iced H O and NaOH 10% (till pH alkaline) and extracted with CH Cl . The organic
2
2
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
layers were combined and dried over Na SO , filtered and evaporated under reduced pressure.
2
4
Residue was purified by column chromatography on silica gel.
General procedure B for nucleophilic substitution on 10ꢀchloroꢀ1,2,3,4ꢀ
tetrahydrobenzo[b][1,6]naphthyridine. A mixture of 10ꢀchloroꢀ1,2,3,4ꢀ
tetrahydrobenzo[b][1,6]naphthyridine (1 equiv), benzylamine (3 equiv), TEA (3 equiv) and NaI
(
0.1 equiv) in NMP was heated to 130 °C overnight. The mixture was diluted with CH Cl and
2 2
washed with H O twice and brine once. The organic layers were dried over Na SO , filtered and
2
2
4
evaporated under reduced pressure. The residues were purified by column chromatography on
silica gel.
General procedure C for the synthesis of 2Hꢀbenzo[d][1,3]oxazineꢀ2,4(1H)ꢀdione.
Triphosgene (1 equiv) was added to a suspension of 5ꢀsubstituted 2ꢀaminoꢀbenzoic acid (3
equiv) in 1,4ꢀdioxane at 0 °C. The homogeneous reaction mixture was heated to 90 °C for 2h
and then cooled down. The resulting precipitate was isolated by filtration.
General procedure D for formation of 4ꢀoxoꢀ1,4ꢀdihydroquinolineꢀ3ꢀcarboxylate. NaH
(
1.2 equiv) was added portion wise to a solution of methyl acetoacetate (1.2 equiv) in DMA. 2Hꢀ
benzo[d][1,3]oxazineꢀ2,4(1H)ꢀdione (1 equiv) was added and the reaction mixture was stirred to
20 ⁰C for 30 min. The solvent was reduced and water was added. The resulting precipitate was
collected by filtration.
1
ACS Paragon Plus Environment

Page 29 of 64
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
General procedure E for chlorination of methyl 4ꢀoxoꢀ1,4ꢀdihydroquinolineꢀ3ꢀ
carboxylate. Methyl 4ꢀoxoꢀ1,4ꢀdihydroquinolineꢀ3ꢀcarboxylate was suspended in POCl and
3
heated to 110 °C for 30 min. The homogeneous reaction mixture was slowly poured into iced
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
NH aq. and the aqueous phase was extracted with AcOEt (3x 50 mL). The organic layers were
3
dried over Na SO , filtered and evaporated under reduce pressure to give the desire compound.
2
4
General procedure F for bromination of methyl 2ꢀmethylquinolineꢀ3ꢀcarboxylate. A
solution of methyl 2ꢀmethylquinolineꢀ3ꢀcarboxylate (1.0 equiv), Nꢀbromosuccinimide (1.5
equiv) and 2,2’ꢀAzobis(2ꢀmethylpropionitrile) (0.2 equiv) in CCl was heated to reflux for 4h.
4
The reaction mixture was cooled down and the solid was filtered off. The filtrate was
concentrated and the residue was purified by flash chromatography.
General procedure G for formation of 2,3ꢀdihydroꢀ1Hꢀpyrrolo[3,4ꢀb]quinolinꢀ1ꢀone.
EtNH (2M in THF, 3.0 equiv) or tert-butylamine (2M in THF, 3.0 equiv) was added to a
2
solution of methyl 2ꢀ(bromomethyl)ꢀ4ꢀchloroquinolineꢀ3ꢀcarboxylate (1.0 equiv) in ethanol. The
reaction mixture was heated to reflux for 2h. The solvent was evaporated off and the residue was
purified by flash chromatography to give the desired compound.
General procedure H for nucleophilic substitution on 2,3ꢀdihydroꢀ1Hꢀpyrrolo[3,4ꢀ
b]quinolinꢀ1ꢀone. A solution of 2,3ꢀdihydroꢀ1Hꢀpyrrolo[3,4ꢀb]quinolinꢀ1ꢀone (1.0 equiv), 3ꢀ
chloroꢀ4ꢀmethoxybenzylamine hydrochloride (1.5 equiv) and DIPEA (6.0 equiv) in nꢀpropanol
was heated to 90 °C for 2h. After cooling down, the resulting precipitate was collected by
filtration.
Methyl 2ꢀaminoꢀ5ꢀcyanobenzoate (3). A mixture of 2 (21.72 mmol) and CuCN (26.07
mmol) in NMP was stirred for 5h at 210 °C. After cooling the reaction down, the mixture was
ACS Paragon Plus Environment