Discovery of N-Alkyl Catecholamides as Selective Phosphodiesterase-4 Inhibitors with Anti-neuroinflammation Potential Exhibiting Antidepressant-like Effects at Non-emetic Doses

Article abstract of DOI:10.1021/acschemneuro.6b00271

Depression involving neuroinflammation is one of the most common disabling and life-threatening psychiatric disorders. Phosphodiesterase 4 (PDE4) inhibitors produce potent antidepressant-like and cognition-enhancing effects. However, their clinical utilit

Full text of DOI:10.1021/acschemneuro.6b00271

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Research Article
pubs.acs.org/chemneuro
Discovery of N‑Alkyl Catecholamides as Selective
Phosphodiesterase‑4 Inhibitors with Anti-neuroinflammation
Potential Exhibiting Antidepressant-like Effects at Non-emetic Doses
Zhong-Zhen Zhou,*,† Yu-Fang Cheng,† Zheng-Qiang Zou,† Bing-Chen Ge,† Hui Yu,† Cang Huang,†
Hai-Tao Wang,† Xue-Mei Yang,*,‡ and Jiang-Ping Xu*,†
†Department of Neuropharmacology and Novel Drug Discovery, School of Pharmaceutical Sciences, Southern Medical University,
Guangzhou 510515, China
‡Hygiene Detection Center, School of Public Health, Southern Medical University, Guangzhou 510515, China
S
* Supporting Information
ABSTRACT: Depression involving neuroinflammation is one of the
most common disabling and life-threatening psychiatric disorders.
Phosphodiesterase 4 (PDE4) inhibitors produce potent antidepres-
sant-like and cognition-enhancing effects. However, their clinical
utility is limited by their major side effect of emesis. To obtain more
selective PDE4 inhibitors with antidepressant and anti-neuro-
inflammation potential and less emesis, we designed and synthesized
a series of N-alkyl catecholamides by modifying the 4-methoxybenzyl
group of our hit compound, FCPE07, with an alkyl side chain.
Among these compounds, 10 compounds displayed submicromolar
IC50 values in the mid- to low-nanomolar range. Moreover, 4-
difluoromethoxybenzamides 10g and 10j, bearing isopropyl groups,
exhibited the highest PDE4 inhibitory activities, with IC50 values in
the low-nanomolar range and with higher selectivities for PDE4
(approximately 5000-fold and 2100-fold over other PDEs, respectively). Furthermore, compound 10j displayed anti-
neuroinflammation potential, promising antidepressant-like effects, and a zero incidence rate of emesis at 0.8 mg/kg within 180
min.
KEYWORDS: N-Alkyl catecholamide, selective phosphodiesterase-4 inhibitors, anti-neuroinflammation, antidepressant-like effects,
less side-effect, structure−activity relationships
epression accompanied by cognitive dysfunction is one of
This hypothesis provides a framework in which co-therapy for
Dthe most common psychiatric disorders, affecting up to depression is mediated via cAMP signaling. Also, increasing
21% of the population worldwide.1,2 Furthermore, major
evidence suggest that cAMP signaling plays a pivotal role in
long-term memory-enhancing effects12,13 and anti-inflammatory
depressive disorder is a severe, chronic, and life-threatening
activities.14 Additionally, genetic knockdown of PDE4B15,16 or
PDE4D17,18 enzyme in rats produces an antidepressant
phenotype. Furthermore, the inhibition of PDE4B leads to
reduction of neuroinflammation by dampening microglia
cytokine production,19 while the inhibition of PDE4D has
potent cognitive benefits by augmenting signaling through the
cAMP/protein kinase A/cAMP response element-binding
protein (CREB) pathway for memory consolidation.20 Thus,
PDE4 represents a potential therapeutic target for depression.21
The development of PDE4 inhibitors has primarily focused
on novel anti-inflammatory therapies for asthma, chronic
obstructive pulmonary disease, and atopic dermatitis for several
decades. However, although a large number of PDE4 inhibitors
illness with a lifetime prevalence of more than 10%.3 Mounting
evidence indicates that inflammatory and neurodegenerative
processes play an important role in depression and that
enhanced neurodegeneration in depression may be at least
partly attributed to inflammation.1,4,5 Inhibition of tumor
necrosis factor-α (TNF-α),6 cyclooxygenase-2 (COX-2),7 and
inducible nitric oxide synthase (iNOS)5 alleviates depressive
symptoms. Therefore, antidepressant agents with anti-neuro-
inflammatory effects are proposed to be more effective.
Phosphodiesterase 4 (PDE4), which specifically catalyzes the
hydrolysis of cAMP, is highly expressed in brain regions
regulating memory, anxiety, and depression.8,9 It is composed
of four genes (PDE4A to PDE4D), each with multiple
transcripts that produce three protein isoforms, termed long,
short, and super short.10 Altered cAMP signal transduction is a
promising mechanism underlying reduced synaptic plasticity
and neuronal survival in the pathophysiology of depression.11
Received: August 28, 2016
Accepted: October 3, 2016
© XXXX American Chemical Society
A
DOI: 10.1021/acschemneuro.6b00271
ACS Chem. Neurosci. XXXX, XXX, XXX−XXX

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
ACS Chemical Neuroscience
Research Article
Figure 1. PDE4 inhibitors.
have been clinically evaluated, only roflumilast (1)22 and
apremilast (2)23 (Figure 1) are currently approved. In addition,
a small number of PDE4 inhibitors are used in clinical testing to
target neurological diseases. For example, the first-generation
PDE4 inhibitor rolipram (3) exhibits antidepressant activity18,24
and memory enhancement25−27 by elevating the concentration
of intracellular cAMP. However, its side effects resulted in
commercial failure of the product. Nevertheless, many PDE4D
inhibitors with antidepressant activities, memory-enhancing
effects, and less side effects have been identified (Figure 2),
FCPR16 (8)32 bearing a 2-chlorophenyl group was identified
as selective PDE4 inhibitor (about 1111-fold over other PDEs)
with IC50 values in the low-nanomolar range and inhibited LPS-
induced TNF-α production in microglia.
Furthermore, catechol bearing alkyl side chains also displayed
good selectivity for PDE4. For example, compound 9, with an
alkyl side chain, is a selective PDE4 inhibitor with IC50 values in
the low-nanomolar range.33 In this paper, to obtain highly
selective PDE4 inhibitors with antidepressant and anti-neuro-
inflammatory activities and less emesis, we synthesized small
molecules by modifying the 4-methoxybenzyl group in the side
chain of FCPE07 with alkyl side chains (Figure 2). The
pharmacological profile, pharmacokinetics analysis, structure−
activity relationships (SARs), and molecular modeling of these
compounds are also described.
Figure 2. Structures of N-alkyl catecholamides (10a−p).
RESULTS AND DISCUSSION
■
Chemistry. N-Alkyl catecholamides (10a−p) were prepared
from 4-alkoxy-3-alkoxybenzoic acids (11a−d) using our prior
synthetic methodology,31 as shown in Scheme 1. Acids (11a−
d) were converted by thionyl chloride to corresponding
substituted benzoyl chlorides (12a−d), which were subse-
quently reacted with the corresponding amines to prepare
benzamides 10a−p in good yields.
such as DF159687,28 GEBR-7R,29 and ZX-I07 (chlorbipram).30
Our recent study of selective PDE4 inhibitors led to the
discovery of FCPE07 (7),31 which contains a short amide linker
(−CONHCH2−) and a 4-methoxybenzyl group; this com-
pound is a potent and selective PDE4D inhibitor (IC50 = 94
nM) that exhibited a 10-fold selectivity over the PDE4B
subtype and a selectivity of over 1000-fold against other PDE
family members (PDE1−3 and PDE5−11). Compound
Enzymatic Assays. The catalytic domains of all PDE4
subtypes (PDE4A−D) exhibited a high degree of sequence
Scheme 1. Synthetic Route for Compounds 10a−p
B
DOI: 10.1021/acschemneuro.6b00271
ACS Chem. Neurosci. XXXX, XXX, XXX−XXX

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
ACS Chemical Neuroscience
Research Article
Table 1. Inhibition of PDE4CAT by Target Compounds 10a−p
a
compd
10a
R
R1
Me
R2
R3
IC50 (nM)
cyclopentyl
H
H
H
H
H
H
H
H
H
H
H
H
H
H
Et
Et
H
i-Pr
170
380
10b
cyclopentyl
Me
n-Pr
n-Bu
i-Pr
10c
cyclopentyl
Me
1000
360
10d
10e
cyclopropylmethyl
cyclopropylmethyl
cyclopropylmethyl
cyclopentyl
Me
Me
n-Pr
n-Bu
i-Pr
740
10f
Me
1500
20
10g
CHF2
CHF2
CHF2
CHF2
CHF2
CHF2
CHF2
CHF2
CHF2
CHF2
CHF2
10h
10i
cyclopentyl
n-Pr
n-Bu
i-Pr
80
cyclopentyl
300
10j
cyclopropylmethyl
cyclopropylmethyl
cyclopropylmethyl
cyclopentyl
60
10k
n-Pr
n-Bu
n-octyl
n-octyl
Et
70
10l
290
10m
10n
10o
10p
FCPE07
rolipram
34000
26000
2000
3900
1100
550
cyclopropylmethyl
cyclopentyl
cyclopropylmethyl
cyclopropylmethyl
Et
b
4-methoxybenzyl
a
b
Data reported represent the mean of two experiments. Data for compound FCPE07 were extracted from ref 31.
conservation, with a single active-site amino acid determining
the nucleotide selectivity.34,35 Therefore, all the synthesized
compounds were initially tested in duplicate on PDE4CAT
(the core catalytic domains of human PDE4) according to
reported protocols,36 using rolipram as a positive control. All
the compounds were tested on PDE4CAT at nine different
concentrations (10−8 to 10−4 M). The IC50 values were
determined by nonlinear regression analysis of their inhibition
curves (see Table 1). Compared with rolipram and FCPE07,
most N-alkyl catecholamides (10a−p) showed excellent
inhibitory activities against PDE4CAT, having IC50 values in
the mid- to low-nanomolar range. N-Alkyl catecholamides
(10a−p) bearing isopropyl and n-propyl groups displayed
higher inhibitory activities then FCPE07 and rolipram.
To obtain more information about their subtype selectivities,
the most active compounds (10g−i and 10j−l) were further
tested in full length human PDE4A4, PDE4B1, PDE4C1, and
PDE4D7, which were purchased form BPS Bioscience Inc. As
shown in Table 2, the results suggest that all the screened
compounds displayed higher PDE4D7 inhibitory activities than
PDE4A4. Compounds 10g and 10j, which bear an isopropyl
group, showed excellent PDE4B1 and PDE4D7 inhibitory
activities, higher than those of rolipram and FCPE07.
PDEs exist as 11 different isozymes involved in various
physiological processes. Therefore, selective inhibition of PDE4
is very important. We determined the selectivity of compounds
10g and 10j toward the other PDE isoforms using human
PDE1A1, PDE1A, PDE2A, PDE3B, PDE5A, PDE6C, PDE7A,
PDE8A1, PDE9A2, PDE10A2, and PDE11A, which were
purchased form BPS Bioscience Inc. (Table 3). As shown in
Table 3, compounds 10g and 10j displayed high PDE4
selectivity, at least 5000- and 2100-fold over other PDEs,
respectively.
Structure−Activity Relationships (SAR) and Docking
Study. Careful analysis of the SAR of N-alkyl catecholamides
(10a−p) reveals that the substituents on the catecholic group
and N-alkyl groups have profound effects on the inhibitory
activities of the compounds. First, N-alkylbenzamides bearing a
difluoromethoxy group (10g−l) always exhibit higher inhib-
Table 3. Inhibition of PDEs (PDE1−3 and PDE5−11)
10g
10j
a
b
c
b
c
PDEs
at 10 μM
IC50 (μM)
at 10 μM
IC50 (μM)
a
PDE1A
PDE2A
PDE3B
PDE5A
PDE6C
PDE7A
PDE8A1
PDE9A2
PDE10A2
PDE11A
3
12
6
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
5
13
6
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
Table 2. Inhibition (IC50, nM) of Four PDE4 Subtypes
compd
10g
PDE4A4
PDE4B1
PDE4C1
PDE4D7
32
110
397
113
160
569
b
10
b
79
b
21
87
23
11
3
22
17
5
10h
10i
b
b
190
47
10j
31
b
164
b
19
5
15
8
10k
122
352
94
10l
b
b
4
8
c
FCPE07
rolipram
976
86
b
12
14
386
118
160
a
b
All PDEs are full enzyme with N-terminal GST-tag. Inhibition (%)
of selected compounds at 10 μM. Data represent the mean of two
experiments. Data represent the mean of two experiments.
a
b
c
Data represent the mean of two experiments. Not tested. Data for
compound FCPE07 were extracted from ref 31.
c
C
DOI: 10.1021/acschemneuro.6b00271
ACS Chem. Neurosci. XXXX, XXX, XXX−XXX

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
ACS Chemical Neuroscience
Research Article
itory activities than their corresponding analogues bearing a
methoxy group (10a−f). Second, the inhibitory activities of
compounds 10a−p against PDE4CAT, PDE4A, and PDE4D
were consistently improved by shortening the chain length.
Among these compounds, 4-difluoromethoxybenzamides bear-
ing a long alkyl chain (e.g., 10m and 10n) displayed
substantially weaker inhibitory activities. However, 4-difluor-
omethoxybenzamides bearing propyl groups (such as n-Pr and
i-Pr) exhibited the highest PDE4 inhibitory activities, with IC50
values in the low-nanomolar range. Third, the configuration of
the alkyl groups on the catechol-based amides affects their
inhibitory activities. For example, all benzamides bearing an
isopropyl group (such as 10a, 10d, 10g, and 10j) showed
higher inhibitory activities than their corresponding analogues
with an n-propyl group (such as 10b, 10e, 10h, and 10k).
Finally, N,N-dialkyl catecholamides displayed very weak
inhibitory activities (e.g., 10o and 10p). The SAR results for
the N-alkyl catecholamides (10a−p) reveal that the activities
decrease in the following order: isopropyl > n-propyl > n-butyl
> diethyl > n-octyl.
example, FCPE07 forms two water bridges in contact with the
Ser374 and Ser521 residues.31
Anti-neuroinflammatory Activity in Microglial Cells.
Previous research41−43 suggests that inhibition of the
inflammatory cytokine tumor necrosis factor (TNF) in
depressed patients with increased inflammation can be
associated with decreased depressive symptoms. Also, the
pro-inflammatory toll receptor TNF-α pathway in monocytes,
macrophages, and microglial cells (BV-2) is regulated by
PDE4.4,44 Here lipopolysaccharide (LPS) was used to induce
an inflammatory effect in microglia BV-2 cells. To determine
the effect of 10j on BV-2 activation, Western blot analysis was
performed to assess the protein levels of TNF-α, iNOS, and
COX-2 in response to LPS.
First, microglial cells were treated with different concen-
trations of 10j (5, 20, and 80 μM, dimethyl sulfoxide (DMSO)
final concentration = 0.1%) for 1 h and then incubated with
LPS (1 μg/mL) for 24 h. Prior to studying the impact of 10j on
LPS-induced neuroinflammation, we examined the potential
toxic effects of 10j on BV-2 microglial cells. Our data showed
that individual treatments with 10j and LPS did not produce
any signs of toxicity at the selected concentrations. We also
found that 10j alone at a dose of 80 μM had no toxic effects on
BV-2 microglial cells (Figure 4).
To gain insight into the SARs and the observed selectivity,
we studied the binding of the synthetic inhibitors in PDE4D
(PDB code 3G4K28). The catechol motif in catecholic PDE4
inhibitors is known to accept a double hydrogen bond from the
carboxamide group of a glutamine residue at the back of the
PDE4 catalytic site.34,37,38 As shown in Figure 3, docking of
Figure 3. Potential binding configuration of 10g and 10j in the active
site of PDE4D. The C atoms in 10g and 10j are colored in magenta
and yellow, respectively; the total docking scores of 10g and 10j are
9.3383 and 9.1722, respectively. The native ligand, rolipram (total
docking score = 9.018), is shown in the thin wire model, and C atoms
are colored in cyan. The H-bonds between each ligand and the active-
site amino acids are shown in yellow.
Figure 4. Cell viability of BV-2 cells treated with 10j. Cells were
incubated with indicated concentrations of 10j for 1 h before
incubation with LPS (1 μg/mL) for 24 h. Cell viability was evaluated
using the MTT assay. Results are shown as percentages of the control
sample, which was set as 100%. Values are means standard error of
mean (SEM) from three independent experiments.
compounds 10g and 10j into PDE4D showed very similar
binding configurations and reasonably good overlap with the
native ligand. In the active site of PDE4D, the catechol moieties
in 10g and 10j were predicted to occupy the pocket of the
catechol moiety in rolipram, with subtle variations, and interact
with the binding site by hydrogen bonding with the NH2 group
of the Gln535 residue and by π-stacking with Phe538.
Meanwhile, the 4-difluoromethyl group is in position to form
a hydrogen bond with Asn487. Thus, the improvement in
activity of the N-alkylbenzamides (10g−l) bearing a difluor-
omethoxy group could be attributed to the extra H-bond of the
4-difluoromethyl group with the Asn487 residue. In addition,
the linker region (CONHCH2) developed an additional
hydrogen bond with a conserved water molecule, which
extends to Tyr325. Similar water bridges were also observed
in other docking studies of selective PDE4 inhibitors.31,39,40 For
We then investigated the effect of compound 10j on the
expression of pro-inflammatory factors in BV-2 cells. As shown
in Figure 5, compound 10j alone had no effect on the
production of TNF-α, iNOS, and COX-2 in microglia at a
concentration of 80 μM. Stimulation of microglia with LPS
alone led to significant increases in TNF-α, iNOS, and COX-2
levels 24 h after treatment. Interestingly, compound 10j was
able to suppress LPS-induced TNF-α, iNOS, and COX-2
production in a concentration-dependent manner (Figure
5B,C,D). Compound 10j inhibited LPS-induced TNF-α,
iNOS, and COX-2 production in microglia (p < 0.01) at a
concentration of 80 μM. Furthermore, microglia cells treated
with 10j at a concentration of 5 μM also inhibited LPS-induced
TNF-α production (Figure 5B,C,D). These data demonstrate
that PDE4 inhibitor compound 10j can effectively inhibit LPS-
D
DOI: 10.1021/acschemneuro.6b00271
ACS Chem. Neurosci. XXXX, XXX, XXX−XXX