New selective phosphodiesterase 4D inhibitors differently acting on long, short, and supershort isoforms

Article abstract of DOI:10.1021/jm900977c

The lack of selective inhibitors toward the long, short, or supershort phosphodiesterases (PDE4s) prevented researchers from carefully defining the connection between different enzyme isoforms, their brain localization, and their role in neurodegenerative diseases such as Alzheimer's disease (AD). In the search for new therapeutic agents for treating memory and learning disorders, we synthesized new rolipram related PDE4 inhibitors, which had some selectivity toward the long form PDE4D3. The first series was synthesized as racemate and then resolved by semipreparative HPLC on chiral supports. Herein we report the synthetic pathways to obtain compounds 1a-c, 2a-c, 3a-c, 4a-f, 5a,b, 6a,b, 7a,b, the chiral analytical study to resolve compounds 1a-c, 2a-c, 3a-c, the molecular docking study for compound 1c, and the biological results and some SAR considerations that provide some insights and hints for the structural requirements for PDE4D subtype selectivity and enzyme inhibition.

Full text of DOI:10.1021/jm900977c

6546 J. Med. Chem. 2009, 52, 6546–6557
DOI: 10.1021/jm900977c
New Selective Phosphodiesterase 4D Inhibitors Differently Acting on Long, Short, and Supershort
Isoforms
Olga Bruno,*,† Alessia Romussi,† Andrea Spallarossa,† Chiara Brullo,† Silvia Schenone,† Francesco Bondavalli,†
Nicolas Vanthuyne,‡ and Christian Roussel ‡
†Dipartimento di Scienze Farmaceutiche, University of Genoa, v.le Benedetto XV, 3-16132 Genoa, Italy, and ‡UMR 6263, Laboratoire de
Steꢀreꢀochimie Dynamique et Chiraliteꢀ Chirosciences, ISM2, University “Paul Ceꢀzanne”, Marseille, France
Received February 25, 2009
The lack of selective inhibitors toward the long, short, or supershort phosphodiesterases (PDE4s)
prevented researchers from carefully defining the connection between different enzyme isoforms, their
brain localization, and their role in neurodegenerative diseases such as Alzheimer’s disease (AD). In the
search for new therapeutic agents for treating memory and learning disorders, we synthesized new
rolipram related PDE4 inhibitors, which had some selectivity toward the long form PDE4D3. The first
series was synthesized as racemate and then resolved by semipreparative HPLC on chiral supports.
Herein we report the synthetic pathways to obtain compounds 1a-c, 2a-c, 3a-c, 4a-f, 5a,b, 6a,b, 7a,b,
the chiral analytical study to resolve compounds 1a-c, 2a-c, 3a-c, the molecular docking study for
compound 1c, and the biological results and some SAR considerations that provide some insights and
hints for the structural requirements for PDE4D subtype selectivity and enzyme inhibition.
Introduction
memory deficits induced by a lowered serotonin neurotrans-
mission.12 Rolipram antagonizes the memory impairment
caused by a deficit in postsynaptic cAMP/protein kinase
A/cAMP-response-element-binding (cAMP/PKA/CREB)
pathway and counteracts the tendency of amyloid beta 42
peptide (Aβ42) to suppress cAMP/PKA/CREB signaling cas-
cade,13 as confirmed by an experimental murine model of AD.14
Recently, subchronic rolipram treatment was also reported to
improve long-term memory performances in normal mice.15
The four PDE4 genes encode for different enzyme isoforms
(namely A, B, C, and D) further classified in 21 subtypes.16-18
On the basis of their primary structures, PDE4 can also be
distinguished in “long”, “short” or “supershort” enzymes.19
In particular, the long isoforms show two regulatory regions,
upstream conserved region 1 and 2 (UCR1 and UCR2),
inserted between the N-terminal portion and the catalytic
domain. The protein kinase A (PKA) mediated phosphoryla-
tion of UCR1 domain activates PDE4 enzyme;20 two phos-
phorylation sites for extracellular signal-regulated kinase
(ERK) and phosphoinositide-3-kinase (PI3K) are also pre-
sent in the catalytic domain (Ser579 and Ser239, respectively),
further modulating enzyme activity.21 UCR1 domain is miss-
ing in the short forms, whereas the supershort forms not only
lack UCR1 but also have a truncated UCR2 domain.
Cyclic adenosine monophosphate (cAMPa) plays a pivotal
role in several physiological events. In the central nervous
system(CNS), cAMP controlsthe signal transductioncascade
activated by neuron stimulation.1 It is implicated in complex
processes such as learning2 and memory.3 Recently, the
involvement of cAMP in the pathological events causing
AD has been clearly demonstrated.4,5
PDE4 specifically catalyze the hydrolysis of cAMP modu-
lating the intracellular concentration of this second messenger.
A number of biological and pharmacological studies have
proven the involvement of PDE4 enzymes in memory processes
and suggested that PDE4 inhibitors may be therapeutic agents
for cognitive disorders. For example, PDE4 enzymes are
involved in N-methyl-D-aspartate (NMDA) receptor-mediated
signal transduction mechanisms. Rolipram is a known non-
selective PDE4 inhibitor, able to reverse the amnesic effects of
the NMDA antagonist MK-801 in rats and to enhance the
NMDA-induced cAMP increase in rat’s cortical neurons.6,7
Moreover, rolipram improves the learning ability of young
rats8,9 and reverses scopolamine-induced cognitive deficits.10,11
The increase of cAMP levels, induced by a PDE4 inhibitor,
enhances 5-hydroxytriptamine (5-HT) release and reverses
It is well-known that PDE4B and PDE4D (particularly
PDE4D1-D5) are highly expressed in hippocampus and cortex
regions, but it is to ascertain what isoforms are overexpressed in
the different brain regions affected in AD patients.22
The lack of selective inhibitors toward the long, short or
supershort forms prevented a clear connection between the
different isoforms, their brain localization, and biological
function. The poor selectivity of rolipram toward the diffe-
rent PDE4 isoforms causes a series of central side effects
*To whom correspondence should be addressed. Phone: þ39 010 353
8367. Fax: þ39 010 353 8358. E-mail: obruno@unige.it.
a Abbreviations: AD, Alzheimer disease; PDE2, PDE3, PDE4,
PDE5, PDE6 phosphodiesterase type 2, 3, 4, 5, 6; cAMP, cyclic
adenosine monophosphate; NMDA, N-methyl-D-aspartate; 5-HT, 5-
hydroxytriptamine; cAMP/PKA/CREB, cAMP/protein kinase A/
cAMP-response-element-binding; Aβ42, amyloide beta 42 peptide;
UCR1 and UCR2, upstream conserved region 1 and 2; PKA, protein
kinase A; ERK, extracellular signal-regulated kinase; PI3K, phosphoi-
nositide-3-kinase; R-TNF, alpha-tumor necrosis factor; HARBS, high
affinity rolipram binding site.
r
pubs.acs.org/jmc
Published on Web 10/14/2009
2009 American Chemical Society

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 21 6547
Scheme 1. Synthesis of Compounds 1a-c, 2a-c, 3a-c, 4a-f, 5a,b, 6a,b, 7a,ba
a Reagents: (a) NH2OH HCl, EtOH, rt, 4 h, then, H2O, NaHCO3; (b) Na, abs EtOH, rt, then epichlorohydrin, DMF, 45 °C, overnight; (c) HNR0R00,
3
45 °C, 12 h; (d) (CH3CO)2O, anhyd CH3COONa, 45 °C, 3 h; (e) benzoyl chloride, Py, 45 °C, 5 h; (f) SOCl2, anhyd TEA, rt, 5 h; (g) BrCH2COBr, anhyd
K2CO3, DCM, 0 °C; (h) HNR0R00, anhyd TEA, DCM, reflux, overnight; (i) BrCH2COOEt, anhyd K2CO3, DCM, rt, then, reflux, overnight; (j) NaOH,
EtOH/H2O, 55 °C, 2 h; (k) SOCl2, DCM, reflux, 4 h; (l) HNR0R00, DCM, 45 °C, 12 h.
(e.g., sedation, emesis), which complicate the understanding
of its physiological action and prevent its use as a drug.
Recently, we reported a series of rolipram analogues
as potential anti-inflammatory agents able to inhibit neutro-
phil activity.23 These 3-cyclopentyloxy-4-methoxybenzalde-
hyde derivatives (e.g., compounds 1a-c, see Scheme 1 and
Table 1) inhibited the superoxide anion production
in neutrophils stimulated with alpha-tumor necrosis factor
(R-TNF), and this effect was related to a dose-dependent
increase of cAMP levels. The most active compound (1c),
screened on several enzymatic assays, resulted in PDE4
competitive antagonist with a high selectivity being inactive
toward PDE3 and PDE5. The binding test on high affinity
rolipram binding site (HARBS) showed a lower ratio PDE4-
inhibition/rolipram binding affinity than rolipram (2.46
versus 152), evidencing a possible lack of side effects in vivo.
These encouraging results prompted us to select 1c as lead
compound for the development of a new series of PDE4
inhibitors which may be useful for treating neurodegenerative
disorders.
hindrance variation on the activity. The replacement of 1c
morpholine moiety with other basic portions led to derivatives
4, while the shortening of the propyl chain led to compounds 6
and 7 (see Table 2).
To assess the influence of the chiral center on the enzymatic
interaction, we developed an analytical method of chiral
HPLC and the racemic mixtures of 1c, 2c, and 3c were
resolved by semipreparative HPLC on chiral supports, as
reported in the Experimental Section.
The racemates 1a-c, 2a-c, 3a-c, 4a-f, and the pure
enantiomers 1c(þ), 1c(-), 2c(þ), 2c(-), 3c(þ), and 3c(-)
were tested on human PDE4 (see Table 1). Compounds 5a,b,
6a,b, and 7a,b were tested on human PDE4D as racemic
mixture (see Table 2).
The most active compounds 4a, 5b, 6b, and 7b were further
tested on a panel of different PDE4 isoforms (namely A4, B2,
C2, D1, D2, D3), considering them as the major representative
(see Table 3). Compound 4a was also tested on PDE1C, PDE2,
PDE3, PDE5, and PDE6 to assess its selectivity profile.
To elucidate the bioactive conformation of compounds
1 inside the PDE4D binding site, a docking calculation was
carried out on 1c. The four possible 1c isomers (namely, E/R,
E/S, Z/R and Z/S) were considered independently. The dock-
ing complexes were calculated using the Autodock 3.05
program according to the procedure described in the Experi-
mental Section.
In this new series of compounds, the 3-cyclopentyloxy-
4-methoxyphenyl moiety was kept constant and modifications
were applied to the linker chain and to the terminal amino
group. Thus, in compounds 2, 3, and 5, the hydroxy group of
the propyl chain was either esterified (2,3) or replaced by a
chlorine atom (5) to evaluate the effect of electronic and steric

6548 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 21
Bruno et al.
Table 1. Structure and PDE4 Inhibition of Compounds 1a-c, 2a-c, 3a-c, 4a-f, and Rolipram as Reference Compounda
a The PDE4 inhibition is expressed as % inhibition in respect to the control at 10 μM, in duplicate, and as IC50 (μM) for active compounds. Fiducial
intervals at 95% confidence are reported at μM concentration. Cerep’s protocol considers as inactive compounds having % inhibition inferior to 50%,
thus the IC50 values were not calculated for them. Active compounds were further tested at five concentrations in the interval 5 ꢀ 10-8-10-4 M, and the
Chemistry. Derivatives 4a-f were prepared according to
the procedures already reported for the synthesis of deri-
vatives 1.23 Briefly, the reaction of hydroxylamine hydro-
chloride with 3-cyclopentyloxy-4-methoxybenzaldehyde,
prepared by isovanillin’s alkylation,24 led to the 3-cyclopen-
tyloxy-4-methoxybenzaldehyde oxime intermediate (8) as a
mixture of the two isomers (E/Z), one being preponderant
(about 75%) (see Scheme 1). The two isomers were first
separated by column chromatography (Silica gel, gradient of
diethyl ether/petroleum ether 40-60 °C (1:1)) and charac-
1
terized by elemental analysis and H NMR spectral data
(available on Supporting Information). In agreement with

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 21 6549
Table 2. Structure and Recombinant Human PDE4D Inhibition of Compounds 5a,b, 6a,b, 7a,b, and Rolipram as Reference Compounda
a The PDE4D inhibition is expressed as % inhibition in respect to the control at 10 μM, in duplicate, and as IC50 (μM) for active compounds. Fiducial
intervals at 95% confidence are reported at μM concentration. Only compounds showing control inhibition higher than 50% were further tested at five
compounds were obtained as theoretical mixture of four
diastereoisomers (E/R, E/S, derived from E oxime; Z/R and
Z/S derived from Z oxime) in percentages not detectable by
nonchiral analytical methods. Compounds 2a-c and 3a-c
were prepared by acylation of the corresponding alcohols 1,
with acetic anhydride or benzoyl chloride, respectively (see
Scheme 1). The chiral HPLC separation of compounds 1c,
2c, and 3c evidenced the prevalence of the more stable
E isomers with a very small amount of Z isomers (see
Experimental Section). After storage in different conditions
(solvent, temperature), the mixtures were again analyzed by
chiral HPLC and no increase of Z isomers was remarked
(data not reported).
Compounds 5a,b were obtained from the corresponding
alcohols 1c and 4a, respectively, with thionyl chloride in
dichloromethane (see Scheme 1).
The reaction of 8 with bromoacetyl bromide, in the
presence of anhydrous K2CO3, at 0 °C, led to intermediate
10 that was treated with morpholine or 2,6-dimethylmor-
pholine to yield compounds 6a,b (see Scheme 1).
Table 3. Inhibition Activity toward Different PDE4 Isoforms of Com-
pound 4a, 5b, 6b, 7b, and Rolipram as Reference Compounda
compounds
PDE4 isoforms
4a
5b
6b
7b
rolipram
PDE4A4
PDE4B2
PDE4C2
PDE4D1
PDE4D2
PDE4D3
34
39
19
54
53
68
37
24
35
41
62
69
5
14
10
11
15
20
34
23
25
57
76
67
79
65
57
78
80
79
a The results are expressed as % inhibition in respect to the control at
10 μM concentration, in duplicate.
literature data,25 we assigned the E configuration to the
predominant compound which showed the aldehyde proton
signal at 8.01 ppm, shifted 0.41 units to high fields with
respect to the other isomer, owing to the paramagnetic shift
induced by hydroxy group on the aldehyde proton. Because
all the attempts to separate 8 E and Z isomers by HPLC
failed, we carried out a quantitative photodensitometric
analysis, after TLC separation, that evidenced the presence
of 75% of E and 25% of Z isomer, as reported in the
Experimental Section.
Finally, compounds 7a,b were prepared starting from 8
that reacted with 2-bromoacetic acid ethyl ester to give the
intermediate (11) that was then hydrolyzed in NaOH/etha-
nol medium; the so obtained carboxylic acid (12) was
activated in the corresponding acyl chloride and reacted
with the proper amines to yield the desired compounds 7a,
b (see Scheme 1).
Then the oxime (8) was reacted as a E/Z mixture with
epichlorohydrin to give the intermediate epoxide (9) that was
reacted, without purification, with the suitable amines to
give compounds 1a-c and 4a-f (see Scheme 1). All the

6550 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 21
Bruno et al.
Table 4. IC50 (μM) Values of Compounds 4a, 5b, 7b, and Rolipram, as
Reference Compound, Determined in Duplicate toward PDE4D3a
compd
IC50 (μM)
95% confidence intervals (μM)
4a
5b
3.46
4.27
2.18-7.32
2.02-8.34
0.65-7.01
nab
7b
rolipram
1.91
0.4-1.8
a Fiducial intervals at 95% confidence are reported at μM concentra-
tion. Compounds were tested at five concentrations in the interval 5 ꢀ
10-8-10-4 M, and the IC50 values were determined by nonlinear
regression analysis of inhibition curve using Hill equation curve fitting
(Graph Pad Prism, San Diego, CA). b na = not available
Table 5. IC50 (μM) and Ki (μM) Valuesa for Compound 4a and
Rolipram, as Reference Compound, against PDE4D1, PDE4D2, and
PDE4D3
substrate
(cAMP)
substrate 4a
(cAMP) IC50
rolipram 4a rolipram
IC50 Ki Ki
(μM) (μM) (μM)
Figure 1. Docking models of enantiomers 1c(E/R) (pale-green
sticks) and 1c(Z/S) (light-blue sticks) inside the PDE4D active site.
The protein is displayed as CR trace. Water molecules are shown as
red spheres, Zn2þ and Mg2þ metal ions are shown as gray and green
spheres, respectively.
enzyme
conc (μM)
Km(μM) (μM)
PDE4D1
PDE4D2
PDE4D3
1
1
1
1.2
4
7.09
9.18
2.86
0.91
1.17
0.55
3.86 0.49
7.34 0.93
1.55 0.30
1.2
a Compound 4a and rolipram, as reference compound, were tested at
six concentrations in the interval 5 ꢀ 10-9-10-4 M, and the IC50 values
were determined by nonlinear regression analysis of inhibition curve
(Graph Pad Prism, San Diego, CA). The Ki values were derived from the
IC50 values using the Cheng-Prusoff equation: IC50 = Ki (1 þ [S]/Km).
of activity in comparison with the previous compounds. All
the others having different amine residues (3-morpholin-
4-yl-propylamino, 1-methyl-ethanolamino, 2,20-dimethyl-
ethanolamino, piperazinyl) were inactive (Table 1).
5a,b, having a chlorine atom instead of a hydroxy group in
the propyl chain, were less active than the corresponding
alcohols 1c and 4a both on the PDE4D test and on the more
specific PDE4D3 test (Tables 2 and 4).
6a,b, bearing the carbonyl group adjacent to the oxime
oxygen, were slightly active at 10 μΜ concentration toward
PDE4D, but this activity was not further confirmed in the
following tests. 7a,b with the carbonyl function near to the
amine nitrogen, showed a good activity toward PDE4D, 7b
being the most active among the whole series of compounds.
In the preliminary screening on different PDEs and PDE4
isoforms, 4a evidenced a good selectivity (Table 3): it in-
hibited the longform PDE4D3 at good level, the short and
supershort PDE4D1 and PDE4D2 with potency around
50%; it was inactive toward the other isoforms PDE4A4,
PDE4B2, PDE4C2, and the other PDEs (PDE1C, PDE2,
PDE3, PDE5, PDE6, data not shown).
The other compounds 5b, 6b, and 7b showed a different
behavior. 5b and 7b inhibited at good level PDE4D3 and
PDE4D2, while the action on PDE4D1 was not relevant
(41% and 57% of inhibition, respectively); they were inactive
on PDE4A4, PDE4B2, and PDE4C2 enzymes; 6b was
completely inactive toward all the tested isoforms
(Table 3). The IC50 determination against PDE4D3 con-
firmed 5b and 7b as less and more active than 4a, respectively
(Table 4).
The radiometric study performed on PDE4D3, PDE4D2,
and PDE4D1 showed that 4a is PDE4D antagonist, with
greater potency toward PDE4D3 in respect to PDE4D1 and
PDE4D2 (see IC50 and Ki values on Table 5) being the ratio
PDE4D1 Ki/PDE4D3 Ki = 2.5 and PDE4D2 Ki/PDE4D3
Ki = 4.7. It is worth mentioning that the same values
calculated for rolipram were 1.6 and 3.1, respectively, con-
firming the higher selectivity of 4a.
Enzymatic Assays. All the synthesized compounds (1a-c,
2a-c, 3a-c, 4a-f, 5a,b, 6a,b, 7a,b) were tested on human
PDE4 (Table 1) or PDE4D enzymes (Table 2) by CEREP
(Celle L’Evescault, France) following standard proce-
dures,26,27 as briefly described in Experimental Section.
The compounds were initially screened at 10 μM concentra-
tion, then only those derivatives showing more than 50%
control inhibition were tested at further concentrations to
determine the IC50 values. Compounds 1c, 2c, and 3c were
tested both as racemic mixtures and as pure E(þ) and E(-)
enantiomers (Table 1).
Compounds 4a, 5b, 6b, and 7b, which showed the best IC50
values in the preliminary PDE4 or PDE4D tests, were then
submitted to a large screening on a panel of different PDE4
isoforms (A4, B2, C2, D1, D2, D3) to assess its ability to
discriminate between them (see Tables 3 and 4). Compound
4a was further tested on other phosphodiesterase types
(PDE1C, PDE2, PDE3, PDE5, and PDE6) to verify its
selectivity (data not reported). The selectivity studies were
performed by Scottish Biomedical (Glasgow, UK) following
the procedures described in Experimental Section.
Results. Among the first synthesized compounds (1, 2, 3),
the morpholine derivatives 1c, 2c, and 3c were PDE4 in-
hibitors at μM concentration, while the pyrrolidinyl and
piperidinyl analogues were completely inactive. The corre-
sponding pure enantiomers 1c(E,þ), 1c(E,-), 2c(E,þ), 2c-
(E,-), 3c(E,þ), and 3c(E,-) also showed a good PDE4
inhibition at 10 μM concentration; however, while no sig-
nificative change in activity was observed between (þ) and
(-) enantiomers of the acetyl (2c) and benzoyl (3c) deriva-
tives and their racemic forms, 1c(þ) was slightly less active
than 1c(-) and they both were less active than their racemic
form (Table 1).
In the second series (compounds 4a-f), only the 2,6-
dimethylmorpholine derivative 4a and the diethanolamine
derivative 4e were active toward PDE4 enzyme. Particularly,
compound 4a showed an increase, whereas 4e had a decrease
Docking Study, SAR, and Discussion. According to the
docking results, out of the four isomers considered, only
isomers E/R and Z/S were predicted to assume a bioactive

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 21 6551
conformation in keeping with the crystallographic data
collected for other PDE4 inhibitors embedding the 3-cyclo-
pentyloxy-4-methoxyphenyl moiety.
The introduction of an electron-withdrawing substituent
in the propyl chain fairly decreases the activity on PDE4D
and also in PDE4D3.
The PDE4D/1c(E/R) complex is stabilized by a number of
hydrophobic contacts involving the 3-cyclopentyloxy-
4-methoxyphenyl moiety and Tyr159, Asn321, Ile336,
Phe340, and Gln369 side chains. The linker is marginally
involved in complex stabilization, whereas the morpholine
ring establishes van der Waals contacts with Asn209,
Asp272, and Met273. Furthermore, the morpholine oxygen
atom is hydrogen bonded to Asn209 side chain amide
The chain lengthening is detrimental for the activity
(3-morpholin-4-yl-propylamino derivative 4b was inactive),
whereas the replacement of the branched propyl chain with
the shorter oxoethyl chain (derivatives 6 and 7) differently
affects activity, according to the position of the carbonyl
group. In fact, compounds 6a,b were completely inactive
toward all the PDE4 tested, whereas compounds 7a,b
inhibited PDE4D at good level.
˚
nitrogen (distance 2.9 A) and to the carboxylic group of
Conclusions. We reported the synthesis and the biological
and analytical studies of new 3-cyclopentyloxy-4-methoxy-
benzaldehyde derivatives which were good PDE4 inhibitors,
compounds 7b and 4a being the most active ones. All the
compounds, except 6a,b, were selective PDE4D inhibitors,
being almost inactive toward PDE4A4, PDE4B2, and
PDE4C2 enzymes. Compound 4a showed the most interest-
ing profile, having a good PDE4 inhibition (IC50=1.46 μM,
see Table 1) and a good selectivity toward the longforms
PDE4D3. Anyway, it is noteworthy the very different beha-
vior of compound 4a in respect to rolipram, which inhibited
quite indifferently all the PDE4 isoforms (Table 3). As far as
we know, no other PDE4D inhibitors with the same biolo-
gical profile have been reported in the literature until now.
In addition, we remark that even little structural modifica-
tions strongly affect the selectivity toward the different
PDE4D isoforms. The SAR informations herein reported
provide some suggestions to have PDE4 inhibitors able to
distinguish among different PDE4D isoforms. Because they
are the most expressed in hyppocampal region, that is the
brain area mainly involved in cognitive and memory dis-
eases, these compounds would represent useful pharmaco-
logical tools to increase the knowledge on the role of PDE4D
in neurodegenerative pathologies.
˚
Asp272 (distance 2.7 A).
Differently from the above-described model, in the
PDE4D/1c(Z/S) complex, the ligand linker is deeply in-
volved in complex stabilization. The hydroxy group would
be hydrogen bonded to two water molecules (namely 1202
and 1105) directly connected to the metal ions in the active
site. Furthermore, the hydroxy group is in contact with
Ser208 side chain. The morpholine ring interacts with
Glu230 and Asn209 side chains, moreover the morpholine
oxygen would be hydrogen bonded to Met273 main chain
NH group and to Asp272 carboxylic functionality. Further
stabilization to this complex is provided by hydrophobic
contacts involving the 3-cyclopentyloxy-4-methoxyphenyl
moiety and Met357, Phe340, Gln369, Phe372, Tyr159, Asn
321, and Phe372. A feature of these interactions was reported
in Figure 1.
This docking study predicted a good interaction with a
catalytic domain of PDE4D for 1c(E,R) and 1c(Z,S) isomers.
Because the Z isomers were not recovered and the absolute
configuration of 1c(E,þ) and 1c(E,-) was not defined, at the
moment it is difficult to confirm the docking prediction on
the basis of enzymatic activity. Moreover, the lack of crystallo-
graphic data on the regulatory UCR1 domain, which is
present only in the PDE4 longforms, limits the understand-
ing of the activity and selectivity profile of our compounds
that can be rationalized only by the SAR considerations.
On the basis of enzymatic results, we can underline some
structural features responsible for PDE4 inhibitory activity
as well as PDE4D3 selectivity. The morpholinyl substructure
at the end of the propyl chain is fundamental for the
inhibitory activity. In fact, other tertiary (pyrrolidine, piperi-
dine, piperazine) or secondary (i.e., methyl-ethanolamine,
2,20-dimethyl-ethanolamine) amino groups gave inactive
compounds, while the diethanolamine (mimiking an open
morpholine ring) 4e was active, although less than all the
morpholine derivatives. The introduction of two methyl
groups in the morpholine always increases the activity (see
compounds 4a, 5b, 7b).
Experimental Section. Chemistry. Chemical Materials
and Methods. All chemicals were purchased from Sigma-
Aldrich srl (Milan, Italy). All the amines were purified by
distillation before the use.
Melting points are not “corrected” and were measured
€
with a Buchi 540 instrument. IR spectra were recorded with a
Perkin-Elmer 398 spectrophotometer. 1H NMR spectra
were recorded on a Varian Gemini 200 (200 MHz) instru-
ment; chemical shifts are reported as δ (ppm) relative to
tetramethylsilane (TMS) as internal standard; signals were
characterized as s (singlet), d (doublet), t (triplet), m
(multiplet), br s (broad signal); J in Hz.
All compounds were tested for purity by TLC (Kieselgel
60F254 DC-Alufolien, E. Merck, Darmstadt, Germany).
Elemental analyses were determined with an elemental
analyzer EA 1110 (Fison-Instruments, Milan, Italy) and the
purity of all synthesized compounds was >95%.
General Procedure for 3-(Cyclopentyloxy)-4-methoxyben-
zaldehyde O-(2-Hydroxy-3-amin/cycloamin-1-ylpropyl)oxi-
mes (1a-c, 4a-f). The procedure already reported for
compounds 1a-c23 has been modified as follows to increase
the reaction yields. To intermediate crude epoxide 9 (1.45 g,
5 mmol), an excess of the suitable amine (5 mL) was added
and the reaction mixture was stirred for 18 h at 40-50 °C,
monitoring the formation of the product by TLC. After
cooling, the mixture was diluted with diethyl ether (10 mL),
washed twice with water, and extracted three times with 1N
HCl solution. The combined aqueous phases were made
alkaline with 1 M NaOH solution and extracted three times
The propyl-chain branching is essential for activity
(compounds previously synthesized with linear chain were
inactive),23 but the role of the different substituents in the
chiral center is very difficult to rationalize. Compounds with
a free hydroxy group (1c, 4a) were highly active but also the
acetyl or benzoyl esters (2c, 3c) showed a good inhibitory
activity, thus the presence of a hydrogen atom responsible
for hydrogen bond is not absolutely necessary.
The pure enantiomers of 2c and 3c showed enzymatic
inhibition quite similar to the corresponding racemate, while
the pure alcohol 1c(E,-) is more active than the pure
1c(E,þ), but the differences observed in respect to the
racemate are not enough to confirm the presence of an
eutomer form.

6552 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 21
Bruno et al.
with CHCl3. The organic layer was dried (MgSO4), filtered,
and evaporated to give yellowish oils. Amines in excess were
distilled under vacuum. The obtained crude oils were puri-
fied as follows: compound 4a was chromatographed on
silica gel (hexane/diethyl ether (5:95)) and then precipitated
as hydrochloride; compounds 4b, 4c, 4d were crystalli-
zed from diethyl ether; compound 4e was precipitated as
hydrochloride.
In the case of piperazine derivative (compound 4f), the
procedure was modified as follows: to a solution of piper-
azine (2.06 g, 3.98 mmol) in absolute ethanol (6 mL), warmed
at 45-50 °C, the epoxide 9 (1.16 g, 3.98 mmol) in absolute
ethanol (2 mL) was added dropwise. The reaction mix-
ture was stirred at 45-50 °C for 2 h. After cooling, the
mixture was treated as above-mentioned to give an oil that
was purified by chromatography on silica gel (methanol/
chloroform (5:90)). Compound 4f was characterized as
dihydrochloride.
Analytical Ddata of 2-[({[3-(Cyclopentyloxy)-4-meth-
oxyphenyl]methylene}amino)oxy]-1-(morpholin-4-ylmethyl)
Ethyl Benzoate (3c). Yield 70% . 1H NMR (CDCl3): δ
1.30-1.95 (m, 8H, 4CH2 cyclopent), 2.30-2.78 (m, 6H, 3
CH2-N), 3.60 (t, J=4.6 Hz, 4H, 2 O-CH2 morph), 3.78 (s,
3H, OCH3), 4.25-4.50 (m, 2H, O-CH2), 4.57-4.75 (m, 1H,
CH cyclopent), 5.39-5.56 (m, 1H, CHOCOPh), 6.70 (d, J=
7.8 Hz, 1H, H-5 Ar), 6.91 (dd, J=1.6, 7.8 Hz, 1H, H-6 Ar),
7.12 (d, J=1.6 Hz, H-2 Ar), 7.20-7.57 and 7.89-8.10 (2m,
6H, 5 H benzoyl þ CHdN). IR (film): cm-1 1714 (CdO).
Anal. Calcd for C27H34N2O6.
General Procedure for 3-(Cyclopentyloxy)-4-methoxyben-
zaldehyde O-[2-Chloro-3-(cycloamin-4-yl)propyl]-oxime (5a,
b). To a solution of the appropriate amino-alcohol (1c, 4a)
(3.12 mmol) in dichloromethane (3 mL) anhydrous triethy-
lamine (0.43 mL, 3.12 mmol) and thionyl chloride (0.27 mL,
3.74 mmol) were added dropwise at room temperature. After
the addition the mixture was stirred at rt for 4 h. The solvent
and excess thionyl chloride were removed under vacuum, the
residue was diluted with dichloromethane (10 mL) and
washed three times with water; the organic layer was dried
(MgSO4), filtered, and evaporated to give an oil which
was purified by chromatography on silica gel (chloroform/
hexane (9:1)).
Analytical Data of 3-(Cyclopentyloxy)-4-methoxybenzal-
dehyde O-[2-Chloro-3-(morpholin-4yl)propyl]-oxime (5a).
Yield 70%. 1H NMR (CDCl3): δ 1.50-2.05 (m, 8H, 4 CH2
cyclopent), 2.60-2.91 (m, 4H, 2 CH2 morph), 2.95-3.13 (m,
1H, CH), 3.48-3.79 (m, 6H, 2CH2 morph þ CH2), 3.84 (s,
3H, OCH3), 4.22-4.45 (m, 2H, CH2-O), 4.70-4.85 (m, 1H,
CH cyclopent), 6.80 (d, J=8.4 Hz, 1H, H-5 Ar), 6.98 (dd, J=
2.2, 8.4 Hz, 1H, H-6 Ar), 7.18 (d, J=2.2 Hz, H-2 Ar), 7.98 (s,
1H, CHdN). Anal. Calcd for C20H29ClN2O4.
Synthesis and Analytical Data of 3-(Cyclopentyloxy)-4-
methoxybenzaldehyde O-(2-Bromoacetyl)oxime (10). To a
mixture of 3-(cyclopentyloxy)-4-methoxybenzaldehyde
oxime 8 (1.45 g, 6.16 mmol) solved in dichloromethane
(6 mL) and anhydrous potassium carbonate (1.70 g, 12.32
mmol) cooled at 0-5 °C, a solution of bromoacetyl bromide
(0.64 mL, 7.39 mmol) in dichloromethane (2 mL) was added
dropwise, over 20 min period, with stirring. After the addi-
tion, the mixture was filtered and the residue washed several
times with dichloromethane; the filtrate was evaporated
under vacuum to yield a yellowish solid that was washed
more times with diethyl ether, giving a white solid.
Analytical Data of 3-(Cyclopentyloxy)-4-methoxybenzal-
dehyde O-[3-(2,6-Dimethyl-morpholin-4-yl)-2-hydroxypropyl]-
oxime (4a). Yield 79%; mp for C22H34N2O5 HCl 110 °C. 1H
3
NMR (CDCl3): δ 1.05 (d, J=6.0 Hz, 3H, CH3 morph), 1.09
(d, J=6.2 Hz, 3H, CH3 morph), 1.49-2.02 (m, 10H, 4 CH2
cyclopent þ N-CH2), 2.29-2.44 and 2.50-2.82 (2m, 4H, 2
N-CH2), 3.50-3.70 (m, 2H, 2 CH morph), 3.79 (s, 3H,
OCH3), 3.91-4.24 (m, 3H, CH2 þ CH-OH), 4.67-4.80 (m,
1H, CH cyclopent), 6.78-7.15 (3m, 3H, Ar), 7.97 (s, 1H,
CHdN). IR (KBr): cm-1 3400-3200 (OH). Anal. Calcd for
C22H34N2O5 HCl.
3
General Procedure for 2-[({[3-(Cyclopentyloxy)-4-meth-
oxyphenyl]methylene}amino)oxy]-1-(cycloamin-ylmethyl)
Ethyl Acetates (2a-c). To a solution of the suitable amino-
alcohols (1a-c) (1.93 mmol) in acetic anhydride (2 mL)
anhydrous sodium acetate (0.2 g, 2.44 mmol) was added
and the mixture was stirred for 3-5 h at 40-50 °C, monitor-
ing the formation of the product by TLC. After cooling, the
mixture was poured into water and extracted three times with
CHCl3. The combined organic layers were washed twice with
1 M NaOH solution and then several times with water, dried
(MgSO4), filtered, and evaporated to give yellow oils which
were chromatographed on Florisil (100-200 mesh) (diethyl
ether).
Analytical Data of 2-[({[3-(Cyclopentyloxy)-4-meth-
oxyphenyl]methylene}amino)oxy]-1-(morpholin-4-ylmethyl)
Ethyl Acetate (2c). Yield 81%. 1H NMR (CDCl3): δ
1.40-1.95 (m, 8H, 4 CH2 cyclopent), 1.99 (s, 3H, CH3),
2.30-2.60 (m, 6H, 3 N-CH2), 3.60 (t, J=4.6, 4H, 2 O-CH2
morph), 3.78 (s, 3H, OCH3), 4.10-4.32 (m, 2H, O-CH2),
4.70-4.80 (m, 1H, CH cyclopent), 5.18-5.31 (m, 1H,
CHOCOCH3), 6.74 (d, J=8.0 Hz, 1H, H-5 Ar), 6.95 (dd,
J=1.8, 8.4 Hz, 1H, H-6 Ar), 7.14 (d, J=1.8 Hz, H-2 Ar), 7.92
(s, 1H, CHdN). IR (film): cm-1 1738 (CdO). Anal. Calcd
for C22H32N2O6.
General Procedure for 2-[({[3-(Cyclopentyloxy)-4-meth-
oxyphenyl]methylene}amino)oxy]-1-(cycloamin-ylmethyl)
Ethyl Benzoates (3a-c). To a solution of the suitable amino-
alcohols (1a-c) (1.6 mmol) in pyridine (2 mL) benzoylchlor-
ide (0.3 mL, 2.24 mmol) was added and the reaction solution
was stirred for 16 h at 40-50 °C. After cooling, the reaction
mixture was diluted with CHCl3 (10 mL), washed three times
with 1N HCl solution, and then several times with water. The
organic layer was dried (MgSO4), filtered, and evaporated to
yield yellow oils which were chromatographed on Florisil
(100-200 mesh) (diethyl ether).
Yield: 84%; mp 126 °C. 1H NMR (CDCl3): δ 1.40-2.12
(m, 8H, 4 CH2 cyclopent), 3.87 (s, 3H, OCH3), 3.97 (s, 2H,
CH2Br), 4.71-4.90 (m, 1H, CH cyclopent), 6.86 (d, J=8.2
Hz, 1H, H-5 Ar), 7.15 (dd, J=2.0, 8.2 Hz, 1H, H-6 Ar), 7.32
(d, J=2.0 Hz, 1H, H-2, Ar), 8.28 (s, 1H, CHdN). IR (KBr):
cm-1 1768 (CdO). Anal. Calcd for C15H18BrNO4.
General Procedure for 3-(Cyclopentyloxy)-4-methoxyben-
zaldehyde O-(2-Cycloamino-acetyl)oximes (6a,b). To a sus-
pension of suitable amine (morpholine or 2,6-
dimethylmorpholine) (5.05 mmol, 1.5 equiv), anhydrous
potassium carbonate (0.86 g, 6.7 mmol, 2 equiv) and potas-
sium iodide (0.02 g, 0.10 mmol, 0.03 equiv) in dichloro-
methane (4 mL), the intermediate 10 (1.2 g, 3.37 mmol,
1equiv), solved in dichloromethane (2 mL), was added
dropwise. The reaction mixture was refluxed for 3 h. After
cooling, the mixture was filtered and the residue was washed
several times with dichloromethane; the solvent was evapo-
rated under reduced pressure and the obtained oil was
diluted with diethyl ether, washed with water, and extracted

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 21 6553
Table 6. Data for the Chiral HPLC Separations of the Different Racemates at 25 °C, 1 mL/min, UV 254 nm, and Polarimeter
compd
column type (250 mm ꢀ 4.6 mm)
mobile phase
Rt1 (min)
Rt2 (min)
k1
k2
R
1a
Chiralcel OD-H
Chiralpak AS
Chiralpak AD
Chiralcel OB-H
Chiralcel OJ
hexane/isopropyl alcohol 8:2
hexane/isopropyl alcohol 8:2
hexane/isopropyl alcohol 8:2
hexane/ethanol 1:1
29.13
8.95
6.64
3.90
4.19
55.48
8.71
1.98
1.21
0.3
17.49
2.01
hexane/ethanol 1:1
0.39
1b
1c
Chiralcel OD-H
Chiralpak AS
Chiralpak AD
Chiralcel OB-H
Chiralcel OJ
hexane/isopropyl alcohol 8:2
hexane/isopropyl alcohol 8:2
hexane/isopropyl alcohol 8:2
hexane/ethanol 1:1
18.45
8.02
6.33
3.82
4.17
36.09
5.15
1.67
1.11
0.27
0.39
11.03
2.14
hexane/ethanol 1:1
Chiralcel OD-H
Chiralcel OD-H
Chiralpak AS
Whelk-01 (S,S)
Chiralpak AD
hexane/isopropyl alcohol 8:2
hexane/isopropyl alcohol 3:7
hexane/isopropyl alcohol 8:2
hexane/isopropyl alcohol 1:1
hexane/isopropyl alcohol 8:2
23.67 (þ)
8.37 (þ)
10.72
43.63 (-)
13.19 (-)
11.97
6.89
1.79
2.57
4.84
2.13
13.54
3.40
2.99
1.97
1.90
1.16
17.53
9.40
2a
2b
2c
Chiralcel OD-H
Chiralcel OD-H
hexane/isopropyl alcohol 8:2
hexane/isopropyl alcohol 8:2
6.41
5.86
21.15
11.01
1.14
0.95
6.05
2.67
5.32
2.81
Chiralcel OD-H
Chiralcel OD-H
hexane/isopropyl alcohol 8:2
Ethanol
13.06 (þ)
5.81 (þ)
64.41 (-)
14.58 (-)
3.35
0.93
20.47
3.86
6.11
4.15
3a
3b
3c
Chiralcel OD-H
Chiralcel OD-H
hexane/isopropyl alcohol 8:2
hexane/isopropyl alcohol 8:2
6.81 (-)
6.70 (-)
19.80 (þ)
10.93 (þ)
1.27
1.23
5.60
2.64
4.41
2.14
Chiralcel OD-H
Chiralcel OD-H
hexane/isopropyl alcohol 8:2
ethanol
14.63 (-)
6.37 (-)
92.51 (þ)
21.23 (þ)
3.88
1.12
29.84
6.07
7.69
5.41
Table 7. Data for the Best Separation Conditions for 1c, 2c, 3ca
compd
column type (250 mm ꢀ 4.6 mm)
mobile phase
Rt1 (min)
Rt2 (min)
k1
k2
R
1c
2c
3c
Chiralcel OD-H
Chiralcel OD-H
Chiralcel OD-H
hexane/isopropyl alcohol (3/7)
ethanol
ethanol
8.37 (þ)
5.81 (þ)
6.37 (-)
13.19 (-)
14.58 (-)
21.23 (þ)
1.79
0.93
1.12
3.40
3.86
6.07
1.90
4.15
5.41
a Data are for analytical conditions that were scaled up for semipreparative separations.
three times with 1N HCl. The aqueous phase was made
alkaline with NaHCO3 saturated solution and extracted
three times with chloroform; the organic extracts were
washed three times with water, dried (MgSO4), filtered,
and evaporated under reduced pressure to yield yellowish
oils that crystallize by diethyl ether in freezer.
petroleum ether 40-60 °C (2:8)) to yield a pure product as
pale oil.
Yield 57%. 1H NMR (CDCl3): δ 1.27 (t, J=7.2 Hz, 3H,
CH3), 1.46-2.19 (m, 8H, 4 CH2 cyclopent), 3.83 (s, 3H,
OCH3), 4.23 (q, J=7.2 Hz, 2H, CH3-CH2), 4.67 (s, 2H, CH2),
4.70-4.91 (m, 1H, CH cyclopent), 6.80 (d, J=8.4 Hz, 1H,
H-5 Ar), 7.00 (dd, J=1.8, 8.4 Hz, 1H, H-6 Ar), 7.16 (d, J=
1.8 Hz, H-2 Ar), 8.09 (s, 1H, CHdN). IR (KBr): cm-1 1759
(CdO). Anal. Calcd for C17H23NO5.
Analytical Data of 3-(Cyclopentyloxy)-4-methoxybenzal-
dehyde O-(2-Morpholin-4-ylacetyl)oxime (6a). Yield: 28%;
mp 89-90 °C. 1H NMR (CDCl3): δ 1.43-2.13 (m, 8H,
4 CH2 cyclopent), 2.57-2.78 (m, 4H, 2 CH2-N morph),
3.39 (s, 2H, CH2-N), 3.77 (t, J=4.2 Hz, 4H, 2CH2-O morph),
3.87 (s, 3H, OCH3), 4.75-4.92 (m, 1H, CH cyclopent), 6.85
(d, J=8.2 Hz, 1H, H-5 Ar), 7.24 (dd, J=1.8, 8.2 Hz, 1H, H-6
Ar), 7.32 (d, J=1.8 Hz, H-2 Ar), 8.26 (s, 1H, CHdN). IR
(KBr): cm-1 1762 (CdO). Anal. Calcd for C19H26N2O5.
Synthesis and Analytical Data of Ethyl [({[3-(Cyclo-
pentyloxy)-4-methoxyphenyl]methylene}amino)oxy]acetate
(11). To a mixture of oxime 8 (1.15 g, 4.9 mmol) and
anhydrous potassium carbonate (1.35 g, 9.8 mmol), a solu-
tion of ethylbromoacetate (0.65 mL, 5.86 mmol) in dichloro-
methane (1 mL) was added dropwise at rt. The reaction
mixture was heated at 50-60 °C overnight. After cooling, the
residue was filtered and washed several times with dichloro-
methane. The solvent was removed under vacuum, and the
crude oil was chromatographed on silica gel (diethyl ether/
Synthesis and Analytical Data of [({[3-(Cyclopentyloxy)-4-
methoxyphenyl]methylene}amino)oxy]acetic acid (12). To a
solution of 11 (1 g, 3.11 mmol) in ethanol 95% (5 mL), a 2 M
NaOH solution (6.22 mL, 12.44 mmol) was added. The
reaction mixture was heated at 50-60 °C for 2 h. The solvent
was removed under vacuum, and the residue was poured into
water and made acid with 1N HCl. The aqueous phase was
extracted three times with diethyl ether, dried (MgSO4),
filtered, and evaporated. The obtained oil was chromato-
graphed on silica gel (diethyl ether/petroleum ether 40-
60 °C (1:1)) to give a product that was crystallized by diethyl
ether/petroleum ether 40-60 °C mixture (1:1).
Yield 75%; mp 91 °C. 1H NMR (CDCl3): δ 1.40-2.19 (m,
8H, 4 CH2 cyclopent), 3.85 (s, 3H, OCH3), 4.60-4.89 (m, 3H,
CH cyclopent þ CH2),5.70-6.63 (br s, 1H, OH, disappears with
D2O), 6.82 (d, J=8.2 Hz, 1H, H-5 Ar), 7.00 (dd, J=1.8, 8.2 Hz,

6554 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 21
Bruno et al.
Table 8. Time of the Collected Fractions and Quantities of the Pure Enantiomer Obtained for 1c, 2c, 3c
impuritiesa
compd
1st enantiomer
intermediate fractionb
2nd enantiomer
total (mg)
399.1
1c
time (min)
quantities (mg)
4.40-7.70
40
7.70-11.19
164.2
11.19-12.28
28.2
12.28-22.00
166.7
2c
3c
time (min)
quantities (mg)
3.29-7.00
80
7.00-12.00
39.8
12.00-13.49
70
13.49-30.00
94.3
284.1
134
time (min)
quantities (mg)
3.51-6.29
23
6.29-6.97
77.2
6.97-10.00
56.2
10.00-20.00
8.8
a Corresponding to the mixture of Z(þ,-)-isomers. b Corresponding to the superimposed E enantiomers peaks.
Table 9. Enantiomeric Excess and Specific Rotation of the E Enantio-
mers Separated by Semipreparative Chiral HPLC
The values of the retention time (Rt in minutes), the
retention factor (k = (Rt - Rt0)/Rt0, where Rt0 = 3 was
determined by injection of tri-tert-butyl benzene) and the
enantioselectivity (R=k2/k1) are reported in Table 6.
All the columns used in the chiral HPLC experiments were
polysaccharide based chiral stationary phases (CSPs) except
for Whelk-01 (S,S). Table 6 collects the experimental proce-
dure used in the screening for each compound.
25a
compounds
e.e. (%)
RD
þ13
-13
þ11
-11
-18
þ18
1c
2c
3c
1st enantiomer E(þ)
2nd enantiomer E(-)
1st enantiomer E(þ)
2nd enantiomer E(-)
1st enantiomer E(-)
2nd enantiomer E(þ)
98
96
>99
>99
99
99
During the analytical work, we received the biological
results for the in vitro pharmacology assays on PDE4
enzyme for the series 1a-b, 2a-c, and 3a-c.
a Determined in CHCl3, c = 0.1 g/100 mL.
1H, H-6 Ar), 7.17 (d, J=1.8 Hz, H-2 Ar), 8.11 (s, 1H, CHdN).
IR (KBr): cm-1 2500-3500 (OH). Anal. Calcd for C14H17NO5.
General Procedure for 3-(Cyclopentyloxy)-4-methoxyben-
zaldehyde O-(2-Cycloamin-4-yl-2-oxoethyl) oximes (7a,b).
To a solution of the intermediate carboxylic acid 12 (2 mmol)
in dichloromethane (4 mL), a solution of thionyl chloride
(0.17 mL, 2.4 mmol) in dichloromethane (3 mL) was added
dropwise at rt and the reaction mixture was refluxed for
4-5 h. The solvent and excess thionyl chloride were removed
under reduced pressure. The obtained oil was dissolved in
dichloromethane and added to a solution of the suitable
amine (morpholine or 2,6-dimethylmorpholine) (2.12 mmol)
and anhydrous triethylamine (0.39 mL, 2.8 mmol) in di-
chloromethane. The reaction mixture was refluxed over-
night. After cooling, the mixture was washed twice with
brine, dried (MgSO4), filtered, and evaporated to yield crude
oils which were purified by crystallization with diethyl ether
(compound 7a) or by chromatography (Florisil 100-
200 mesh, diethyl ether) (compound 7b).
As it is shown in the biological section, interesting results
were given by compounds 2c and 3c (for compound 1c the
activity was already proved), while all the other compounds
of the series 1-3 resulted inactive; so we decided to optimize
the analytical conditions of separation only for compounds
1c, 2c, and 3c in order to develop the semipreparative
separation.
In Table 7, we report the conditions that allowed a very
good separation of the E(þ) and E(-) enantiomers: the
chromatograms show good values of retention time, excel-
lent baseline separation, and good resolution. In any case,
one observed the presence of one or two minor peaks
between 3 and 6 min, which were assigned to the Z isomers
(HPLC chromatograms are available on Supporting In-
formations).
Enantiomers Recovery by Semipreparative Separation.
All the experiments were performed on a Chiralcel OD
(250 mm ꢀ 10 mm) column by successive injections on a
Knauer unit composed of a Smartline 1000 pump, a Smart-
line 3900 autosampler, a Smartline 2500 UV detector, and a
valve to collect separately the different fractions. The collec-
tion procedures used for each compound are reported below.
Compound 1c (400 mg) was dissolved in hexane/isopropyl
alcohol mixture (3:7) (10 mg/mL), and the solution was
injected 80 times in portions of 500 μL each (5 mg) every
22 min; the separation was performed by hexane/isopropyl
alcohol mixture (3:7), at 30 °C with a flow rate of 5 mL/min;
it took about 30 h and needed 9 L of solvent.
Compound 2c (285 mg) was dissolved in ethanol (19 mg/
mL), and the solution was injected 15 times in portions of
1 mL each (19 mg) every 30 min; the separation was
performed by ethanol, at 30 °C with a flow rate of 5 mL/
min; it took about 7.5 h and needed 2.2 L of solvent.
Compound 3c (300 mg) was dissolved in ethanol (30 mg/
mL), and the solution was injected 20 times in portions of
500 μL each (15 mg) every 40 min; the separation was
performed by ethanol, at 30 °C with a flow rate of 5 mL/
min; it took about 13 h and needed 4 L of solvent.
Analytical Data of 3-(Cyclopentyloxy)-4-methoxybenzal-
dehyde O-(2-Morpholin-4-yl-2-oxoethyl) oxime (7a). Yield
1
75%; mp 99-100 °C. H NMR (CDCl3): δ 1.43-2.10 (m,
8H, 4CH2 cyclopent), 3.39-3.72 (m, 8H, 4 CH2 morph), 3.83
(s, 3H, OCH3), 4.62-4.90 (m, 3H, CH cyclopent þ O-CH2);
6.80 (d, J=8.3 Hz, 1H, H-5 Ar), 7.00 (dd, J=1.6, 8.3 Hz, 1H,
H-6 Ar), 7.16 (d, J=1.6 Hz, H-2 Ar), 8.07 (s, 1H, CHdN). IR
(KBr): cm-1 1660 (CdO). Anal. Calcd for C19H26N2O5.
Analytical data of compounds 4b-f, 2a,b, 3a,b, 5b, 6b, and
7b are available in Supporting Informations.
Chiral HPLC Analysis of Compounds 1a-c, 2a-c, 3a-c.
An analytical study for the separation of the enantiomers
was developed for the compounds 1a-c, 2a-c, and 3a-c
obtained as racemate in the synthesis.
All the compounds, dissolved in ethanol, were screened on
different chiral columns using different mobile phases. The
chiral HPLC experiments were performed on a screening
unit composed of a Merck D-7000 system manager, Merck-
Lachrom L-7100 pump, Merck Lachrom L-7360 oven,
which can accommodate 12 chiral columns alimented by a
Valco 12 positions valve, Merck-Lachrom 7400 UV detector,
and Jasco OR-1590 polarimeter.
In Table 8, we reported the time of the collected fractions
and the quantities of the pure enantiomer obtained after
solvent evaporation.

Article
The purity control was done on the same HPLC apparatus
used in the analytical separation with the same experimental
conditions (injected volume: 20 μL).
The optical rotatory powers of each enantiomer were
measured on a 241 MC Perkin-Elmer polarimeter with
a sodium lamp (589 nm) and a double-jacketed 10 cm cell
at 25 °C.
In Table 9, we showed both the analytical data for the
purity control of each enantiomer expressed as enantiomeric
excess (e.e.) and the specific rotation (RD25).
In conclusion, we succeed in obtaining each E enantiomers
with a good enantiomeric excess and in quantity suitable for
the in vitro enzymatic assays, while Z enantiomers were not
recovered.
Photodensitometric Analysis of Compound 8. To calculate
the percentage of E and Z isomers of compound 8, we carried
out a quantitative photodensitometric analysis after TLC
separation by TLSee instruments (GioPharma, srl, Genoa,
Italy).
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 21 6555
inserted in the protein model and used for docking calcula-
tion; (iii) the 1XOM coordinates were selected (being the
highest resolution model) and Cilomilast was removed from
the binding site; (iv) hydrogen atoms were added and partial
charges were assigned according to CVFF forcefield (Insight
II, biopolymer module).
The ligand “root” was defined automatically and all bonds
were allowed to freely rotate. A 60 ꢀ 60 ꢀ 60 grid (grid
˚
spacing 0.375 A) was centered on the ligand binding site and
electrostatic and affinity maps for each atom type of the
ligand were calculated. The docking search was performed
using the genetic algorithm local search protocol as imple-
mented in AutoDock (number of runs: 10; number of
individuals in population: 150; rate of gene mutation: 0.02;
rate of crossover: 0.8). The docking poses were clustered
˚
(rmsd 2.0 A), and the best conformation of the low energy,
highest populated cluster was selected as the binding con-
formation.
All the calculations were carried out on Silicon Graphic
Indigo2 and Origin 200 workstations. Model analyses were
performed using the CCP4 program suite.31 The programs
Molscript32 and Raster3D33 were used to draw the figures.
Enzyme Assays. PDE4 and PDE4D. PDE4 and PDE4D
enzyme assays were performed by CEREP (Celle L’Eves-
cault, France), using standard procedures as reported in the
literature. Particularly, PDE4 assays were performed using
enzyme purified from human monocyte (U-937);26 PDE4D
tests were carried out on human recombinant PDE4D
obtained from Sf9 cells.27In all the experiments, rolipram
as reference compound was tested at nine concentrations, in
duplicate, to obtain an inhibition curve in order to validate
this experiment. Title compounds were solved in DMSO at
10-2 M concentration and then diluted with water to the
final suitable concentrations. All the compounds were tested
preliminary at 10-5 M concentration, in duplicate. Results
are expressed as a percent inhibition of control activity.
Results showing an inhibition of the control higher than
50% are considered to represent significant effects of the test
compounds. 50% is the most common cutoff value for
further investigation (determination of IC50 value from
concentration-response curves).
Compounds showing inhibition control higher than 50%
were further tested at five concentrations in the interval 5 ꢀ
10-8-10-4 M. IC50 values for rolipram and tested com-
pounds were determined by nonlinear regression analysis of
its inhibition curve, using Hill equation curve fitting (Graph
Pad Prism software). The IC50 values obtained for the
reference compounds are within accepted limits of historic
averages obtained (0.5 log unit. The IC50 values and the
relative Fiducial intervals at 95% confidence for each tested
compounds were reported at μM concentration (Tables 1
and 2) .
Method. The 3-cyclopentyloxy-4-methoxybenzalde-
hyde oxime (8) (250.0 mg) was solved in absolute ethanol
(25.0 mL) obtaining a 10.0 mg/mL solution. Then, 1.0 μL of
analyte solution was spotted on a TLC plate (Silica gel 60
F254, Merck, 10 cm ꢀ 10 cm, 1 cm bottom, six spots, each of
10 μg); the plate was then developed (solvent front=9 cm)
with diethyl ether /petroleum ether 40-60 °C mixture (1:1)
as eluent and dried. The plate was finally analyzed by the
TLComp apparatus as follows.
After elution, the TLC plate was lighted in a suitable dark-
room (TLCam) by a lamp at 254 nm (UV absorption) or
366 nm (fluorescent emission) wavelength and then was
photographed by a digital camera. Some TLC image para-
meters (brightness, contrast, gray scale, multiple releases)
can be optimized after and before the acquisition by the
TLImage software. In this TLC UV absorption analysis, the
spot brightness depends on the energetic radiation of each
pixel and corresponds to the sample concentration, the dark
being the maximal concentration detectable by the system,
while the white corresponds to the analyte’s lack. The
contrast is a measure of sensibility and the multiple releases
(from 1 to 999) make possible an increase of S/N (signal/
noise) ratio and LDR (limit of rilevability) optimization. The
optimized plate image was then elaborated by a second
software, namely TLSee, and converted into a graph like
to a typical HPLC chromatogram, where each spot corre-
sponds to a peak and the peak’s area gives a quantitative
analyte measure.
The plate image after TLImage optimization, the chro-
matogram obtained by TLSee, a histogram relative to the
peak’s area of each spots, a table with absolute and percent
area values, were reported in Supporting Informations.
Docking Study. The molecular structures of the four 1c
isomers were generated by InsightII (builder module), para-
metrized according to the CVFF forcefield,29 and their
PDE4D complexes were calculated using Autodock 3.05.30
The model of the PDE4D enzyme was derived by the
crystal structure of the PDE4D/Cilomilast complex (PDB
code 1XOM) according to the following protocol: (i) the
crystal structures of the available PDE4B and D/ligand
complexes (PDB codes: 1XOM, 1XLZ, 1XON, 1Y2D,
1XOT, 1Q9M, 1OYN) were superimposed (lsqkab program,
CCP4 suite) using the CR atoms; (ii) the PDE4 binding site
was visually inspected and the water molecules conserved in
all the analyzed structures (as well as the metal ions) were
PDE4 Subtype Assays by IMAP Methods. PDE4A4,
PDE4B2, PDE4C2, and PDE4D1-D3 enzyme assays were
performed by Scottish Biomedical (Glasgow, Scotland, UK)
using recombinant human PDE enzymes expressed in a
baculoviral system. The preliminary screening assays were
performed by the IMAP technology (Molecular Devices),
which is based on the high affinity binding of phosphate by
immobilized metal coordination complexes on nanoparti-
cles. The binding reagent complexes with phosphate groups
on nucleotide monophosphate generated from cyclic nucleo-
tides (cAMP) through phosphodiesterases. With fluore-
scence polarization detection, binding causes a change in

6556 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 21
Bruno et al.
the rate of the molecular motion of the phosphate bearing
molecule and results in an increase in the fluorescence
polarization value observed for the fluorescent label atta-
ched to the substrate.
As for PDE4 and PDE4D tests, compounds 4a, 5b, 6b, and
7b were preliminary tested at 10 μM concentration in dupli-
cate (Table 3), then compounds 4a, 5b, and 7b were tested by
the same IMAP method on PDE4D3, at further five con-
centrations from 10-4 to 10-8 M, and IC50 values were
determined by nonlinear regression analysis of its inhibition
curve (GraphPad Prism software) (Table 4).
PDE4D1, PDE4D2, PDE4D3 Assays by Radiometric
Method. An additional study on compound 4a was per-
formed to verify its antagonist activity on PDE4D1,
PDE4D2, and PDE4D3 and to determine its IC50 and Ki
values.
The phosphodiesterase assays were performed using re-
combinant human PDE4D1, PDE4D2, and PDE4D3 en-
zymes expressed in a baculoviral system (0.5 U/μL in 20 mM
Supporting Information Available: Analytical data of com-
pounds 4b-f, 2a,b, 3a,b, 5b, 6b, and 7b. 1H NMR spectra of each
isomer of compound 8. HPLC chromatograms of compounds
1c, 2c, and 3c as racemate and as pure enantiomers. Figures and
table relative to the photodensitometric analysis of compound 8.
This material is available free of charge via the Internet at http://
pubs.acs.org.
References
(1) Siegelbaum, S. A.; Schwartz, J. H. Modulation of synaptic trans-
mission: second messangers. In Principles of Neural Science;
Kandel, E. R, Schwartz, J. H, Jessel, T. M., Ed.; McGraw-Hill Press:
New York, 2000; pp 229-252.
(2) Dudai, Y. Cyclic AMP and learning in Drosophila. Adv. Cycl.
Nucleic Protein Phosphoryl. Res. 1986, 20, 343–361.
(3) Morimoto, B. H.; Koshland, D. E., Jr. Identification of cyclic AMP
as the response regulator for neurosecretory potentiation: a mem-
ory model system. Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 10835–
10839.
(4) Bonkale, W. L.; Cowburn, R. F.; Ohm, T. G.; Bogdanovic, N.;
Fastbom, J. A quantitative autoradiographic study of [3H]cAMP
binding to cytosolic and particulate protein kinase A in post-
mortem brain staged for Alzheimer’s disease neurofibrillary
changes and amyloid deposits. Brain Res. 1999, 818, 383–396.
(5) Cowborn, R. F.; Fowler, C. J.; O’Neill, C. Neurotransmitters,
signal transduction and second-messengers in Alzheimer’s disease.
Acta Neurol. Scand. Suppl. 1996, 165, 25–32.
(6) Zhang, H. T.; Crismass, A. M.; Dorairaj, N. R.; Chandler, L. J.;
O’Donnel, J. M. Inhibition of cyclic AMP phosphodiesterase
(PDE4) reverses memory deficits associated with NMDA receptor
antagonism. Neuropsycopharmacology 2000, 23, 198–204.
(7) Survana, N. U.; O’Donnel, J. M. Hydrolisis of N-methyl-D-aspar-
tate receptor-stimulated cAMP and cGMP by PDE4 and PDE2
phosphodiesterases in primary neuronal cultures of rat cerebral
cortex and hyppocampus. J. Pharmacol. Exp. Ther. 2002, 302, 249–
256.
(8) Blokland, A.; Schreiber, R.; Prickaerts, J. Improving memory: a
role for phosphodiesterases. Curr. Pharm. Des. 2006, 12, 2511–
2523.
(9) Zhang, H. T.; Zhao, Y.; Huang, Y.; Dorairaj, N. R.; Chandler,
L. J.; O’Donnel, J. M. Inhibition of phosphodiesterase 4 (PDE4)
enzyme reverses memory deficits produced by infusion of MEK
inhibitor U0126 into the CA1 subregion of the rat hyppocampus.
Neuropsychopharmacology 2004, 29, 1432–1439.
(10) Zhang, H. T.; O’Donnel, J. M. Effect of rolipram on scopolamine-
induced impairment of working and reference memory in the
radial-arm maze test in rats. Psychopharmacology (Berlin) 2000,
150, 311–316.
(11) Rutten, K.; Prickaerts, J.; Blokland, A. Rolipram reverse scopola-
mine-induced and time-dependent memory deficits in object re-
cognition by different mechanism of action. Neurobiol. Learn.
Mem. 2006, 85, 132–138.
(12) Schoffelmeer, A. N.; Wardeh, G.; Mulder, A. H. Cyclic AMP
facilitates the electrically evoked release of radiolabelled noradre-
naline, dopamine and 5-hydroxytryptamine from rat brain slices.
Naunyn-Schmiedeberg’s Arch. Pharmacol. 1985, 330, 74–76.
(13) Masliah, E. Mechanisms of synaptic dysfunction in Alzheimer’s
disease. Histol. Histopathol. 1995, 10, 509–519.
Tris HCl pH 7.4) (1 unit converts 1 pmol of cAMP to AMP
3
per min at 30 °C, pH 7.4). The enzyme concentration used for
the inhibition assay is one that has been previously identified
as producing activity at the high end of the linear range.
These enzymes were tested for their similarity to PDE
enzymes taken from human tissue using known inhibitor
standard (rolipram). The radiometric assay is a modification
of the two-step method of Thompson and Appleman,28
adapted for 96-well plate format. 4a and rolipram, as re-
ference compound, were dissolved in 100% DMSO at a
concentration of 10 mM. All subsequent assays were per-
formed in 5% DMSO (final). The compounds were tested by
performing a serial dilution at a starting concentration of
100 μM to a final concentration of 1 nM, in duplicate.
Enzyme solution was incubated in each plate well with
cAMP solution (1 μM) as substrate, [3H]-cAMP as tracer,
and the test compounds for 20 min at 30 °C. The enzyme was
denaturated by incubation for 2 min at 70 °C, and the [3H]-50-
AMP was converted in [3H]-adenosine by addition of snake
venom nucleotidase. Uncleaved cyclic nucleotide substrate
was removed by Dowex resin solution. After centrifugation,
the supernatant was added to a scintillant plate containing
200 μL of Microscint-20 for well. The level of reaction
product (labeled phosphate) was finally detected with a
scintillation counter and plotted as CPM values against
the logarithm of inhibitor’s concentration to generate the
dose-response curves (Prism software, GraphPad Inc.)
from which the IC50 values were calculated.
(14) Gong, B.; Vitolo, O. V.; Trinchese, F.; Liu, S.; Shelanski, M.;
Arancio, O. Persistent improvement in synaptic and cognitive
functions in an Alzheimer mouse model after rolipram treatment.
J. Clin. Invest. 2004, 114, 1624–1634.
(15) Rutten, K.; Prickaerts, J.; Schaenzle, G.; Rosenbrock, H.;
Blokland, A. Subchronic rolipram treatment leads to a persistent
improvement in long-term object memory in rats. Neurobiol.
Learn. Mem. 2008, 90, 569–575.
A control solution containing 11 μL of DMSO, 50 μL of
3H-cAMP, and 50 μL of enzyme per well, and blank solu-
tions being the same as the control without enzyme, were
used in each assay. Basing on the close structural relation
with our lead compound 1c, we assume that compound 4a is
a
competitive inhibitor of PDE4D1, PDE4D2, and
(16) Houslay, M. D. PDE4 cAMP-specific phosphodiesterases. Prog.
PDE4D3, and we calculated the Ki values from the IC50
values using the Cheng-Prusoff equation: IC50=Ki (1 þ [S]/
Km) (Table 5).
Nucleic Acid Res. Mol. Biol. 2001, 69, 249–315.
(17) Chandrasekaran, A.; Toh, K. Y.; Low, S. H.; Hong Tay, S. K.;
Brenner, S.; Li Meng Goh, D. Identification and Characterization
of novel mouse PDE4D isoforms: Molecular cloning, subcellular
distribution and detection of isoform-specific intracellular locali-
zation signals. Cell. Signalling 2008, 20, 22139–22153.
(18) Lynex, C. N.; Chen, M. L.; Toh, K. Y.; Low, R. W. C.; Goh, D. L.
M.; Hong Tay, S. K. Identification and molecular characterization
of a novel PDE4D11 cAMP-specific phosphodiesterase isoform.
Cell. Signal. 2008, 20, 2247–2255.
Acknowledgment. We thank Dr. R. Raggio, O. Gagliardo,
and F. Tuberoni for spectral recording and Prof. M. Bradley
for English revision of the text. We give a particular aknow-
ledgement to Prof. G. Amandola, who provided us the TLSee
Instrument for photodensitometric analysis. Financial sup-
port from MIUR (Minister of University and Research,
PRIN grant 2005032713-005) is gratefully acknowledged.
(19) Houslay, M. D.; Adams, D. R. PDE4 cAMP phosphodiesterases:
modular enzymes that orchestrate signalling cross-talk, desensiti-
zation and compartmentalization. Biochem. J. 2003, 370, 1–18.

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 21 6557
(20) MacKenzie, S. J.; Baillie, G. S.; McPhee, I.; MacKenzie, C.;
Seamons, R.; McSorley, T.; Millen, J.; Beard, M. B.; van Heeke,
G.; Housley, M. D. Long PDE4 cAMP specific phosphodiesterases
are activated by protein kinase A-mediated phosphorylation of a
single serine residue in upstream conserved region 1 (UCR1). Br. J.
Pharmacol. 2002, 136, 421–433.
(21) Hill, E. V.; Sheppard, C. L.; Cheung, Y. F.; Gall, I.; Krause, E.;
Houslay, M. D. Oxidative stress employs phosphatidyl inositol
3-kinase and ERK signalling pathways to activate cAMP phos-
phodiesterase-4D3 (PDE4D3) through multisite phosphorylation
at Ser239 and Ser579. Cell. Signalling 2006, 18, 2056–2069.
(22) Zhang, H. T.; Huang, Y.; Survana, N. U.; Deng, C.; Crissman,
A. M.; Hopper, A. T.; De Vivo, M.; Rose, G. M.; O’Donnel, J. M.
Effects of the novel PDE4 inhibitors MEM1018 and MEM1091 on
memory in the radial-arm maze and inhibitory avoidance tests in
rats. Psychopharmacology 2005, 179, 613–619.
(26) Torphy, T. J.; Zhou, H. L.; Cieslinski, L. B. Stimulation of beta
adrenoceptors in a human monocyte cell line (U937) up-regulates
cyclic AMP-specific phosphodiesterase activity. J. Pharmacol.
Exp. Ther. 1992, 263, 1195–1205.
(27) Saldou, N.; Obernolte, R.; Huber, A.; Badcker, P. A.; Wilhelm, R.;
Alvarez, R.; Li, B.; Xia, L.; Callan, O.; Su, C.; Jargin, K.; Shelton,
E. R. Comparison of recombinant human PDE4 isoforms: inter-
action with substrate and inhibitors. Cell. Signalling 1998, 10, 427–
440.
(28) Thompson, W. J.; Appleman, M. M. Multiple cyclic nucleotide
phosphodiesterase activities from rat brain. Biochemistry 1971, 10,
311–316.
(29) Dauber-Osguthorpe, P.; Roberts, V. A.; Osguthorpe, D. J.; Wolff,
J.; Genest, M.; Hagler, A. T. Structure and energetics of ligand
binding to proteins: Escherichia coli dihydrofolate reductase-tri-
methoprim, a drug-receptor system. Proteins 1988, 4, 31–47.
(30) Morris, G. M.; Goodsell, D. S.; Halliday, R. S.; Huey, R.; Hart,
W. E.; Belew, R. K; Olson, A. J. Automated docking unsing a
Lamarckiangenetic algorithm and an empirical binding free energy
function. J. Comput. Chem. 1998, 19, 1639–1662.
(23) Bruno, O.; Brullo, C.; Arduino, N.; Schenone, S.; Ranise, A.;
Bondavalli, F.; Ottonello, L.; Dapino, P.; Dallegri, F. Synthesis
and biological evaluation of neutrophilic inflammation inhibitors.
Farmaco 2004, 59, 223–235.
(24) Ashton, M. J.; Cook, D. C.; Fenton, G; Karlsson, J. A.; Palfreyman,
M. N.; Raeburn, D; Ratcliffe, A. J.; Souness, J. E.; Thurairatnam, S.;
Vicker, N. Selective type IV phosphodiesterase inhibitors as
antiasthmatic agents. The syntheses and biological activities of
3-(cyclopentyloxy)-4-methoxybenzamides and analogues. J. Med.
Chem. 1994, 37, 1696–1703.
(25) McCarty, C. G. Syn-anti isomerization and rearrangements. In:
The Chemistry of the Carbon-Nitrogen Double Bond. Patai, S., Ed.;
Wiley and Sons: New York, 1970; pp 363-464.
(31) The CCP4 suite: programs for protein crystallography. Collabora-
tive Computational Project, Number 4. Acta Crystallogr., Sect. D:
Biol. Crystallogr. 1994, 50, 760-763.
(32) Kraulis, P. J. MOLSCRIPT: a program to produce both detailed
and schematic plots of protein structures. J. Appl. Crystallogr.
1991, 24, 946–950.
(33) Esnouf, R. M. An extensively modified version of MolScript that
includes greatly enhanced coloring capabilities. J. Mol. Graphics
Modell. 1997, 15, 132–134.