Synthesis and structure - Activity relationship of N-arylrolipram derivatives as inhibitors of PDE4 isozymes

Article abstract of DOI:10.1248/cpb.49.1009

Structure activity studies of N-phenylrolipram derivatives have led to the identification of highly potent PDE4 inhibitors. The potential of these inhibitors for cellular activity was routinely assessed in an assay of fMLP induced oxidative burst in human eosinophils. Since first generation PDE4 inhibitors have been plagued with a number of unwanted side effects, parallel structure activity studies for competition with the [³H]-rolipram binding site in rat brain were performed. In this fashion 5-[4-(3-cyclopentyloxy-4-methoxyphenyl) -2-oxo-pyrrolidin-1-y1]-3-(3-methoxybenzyloxy)benzoic acid N′,N′-dimethylhydrazide (22) was identified as a potent inhibitor of PDE4 which exhibits >1000 fold selectivity versus PDE3, and is a nanomolar inhibitor in all the cellular assays tested. Studies on the stereoselectivity of PDE4 inhibition of this class of rolipram based compounds revealed, that for example (S)-11 is a more potent inhibitor than (R)-11. This effect can also be observed in primary human cells where the (S)-enantiomer is about 10 fold more potent than the corresponding (R)-enantiomer.

Full text of DOI:10.1248/cpb.49.1009

August 2001
Chem. Pharm. Bull. 49(8) 1009—1017 (2001)
1009
Synthesis and Structure–Activity Relationship of N-Arylrolipram
Derivatives as Inhibitors of PDE4 Isozymes
Thomas H. KELLER, Katharine BRAY-FRENCH, F. W. Joachim DEMNITZ, ,1) Thomas MÜLLER,
*
Esteban POMBO-VILLAR, and Christoph WALKER
Respiratory Disease Therapeutic Area, Novartis Horsham Research Center, Wimblehurst Road, Horsham, West Sussex,
RH12 5AB, U.K. Received April 5, 2001; accepted May 14, 2001
Structure activity studies of N-phenylrolipram derivatives have led to the identification of highly potent
PDE4 inhibitors. The potential of these inhibitors for cellular activity was routinely assessed in an assay of fMLP
induced oxidative burst in human eosinophils. Since first generation PDE4 inhibitors have been plagued with a
number of unwanted side effects, parallel structure activity studies for competition with the [3H]-rolipram bind-
ing site in rat brain were performed. In this fashion 5-[4-(3-cyclopentyloxy-4-methoxyphenyl)-2-oxo-pyrrolidin-1-
yl]-3-(3-methoxybenzyloxy)benzoic acid N؅,N؅-dimethylhydrazide (22) was identified as a potent inhibitor of
PDE4 which exhibits Ͼ1000 fold selectivity versus PDE3, and is a nanomolar inhibitor in all the cellular assays
tested. Studies on the stereoselectivity of PDE4 inhibition of this class of rolipram based compounds revealed,
that for example (S)-11 is a more potent inhibitor than (R)-11. This effect can also be observed in primary
human cells where the (S)-enantiomer is about 10 fold more potent than the corresponding (R)-enantiomer.
Key words Rolipram; PDE inhibition; N-arylrolipram derivative; rolipram binding
In past years we have seen a major shift in the perception the most likely side effects are nausea and vomiting and in-
of asthma as a complex disease characterized by a combina- creased gastric acid secretion.8) Accordingly the focus of cur-
tion of reversible obstruction, hyperreactivity and chronic in- rent research effort is aimed at identifying PDE4 inhibitors
flammation of the airways rather than simply in terms of dis- with improved side effect profile compared to first generation
ordered smooth muscle function.2) While all three of these compounds like rolipram, denbufyllin or even theophyllin
pathological features of asthma are important, it is now which has been claimed to owe its clinical efficacy to PDE4
widely accepted that controlling inflammation is the most inhibition.9)
promising approach for the treatment of chronic disease.2)
Rolipram (1), the prototypical, selective PDE4 inhibitor
Accordingly, treatment regimens have been adjusted to in- has been used by many research groups as a starting point
clude inhaled steroids as first line therapy even for mild dis- in their search for potent and selective PDE4 enzyme in-
ease.3) However, despite our improved understanding of the hibitors.10—14) This has led to the identification of a number
underlying cellular events responsible for the progression of of clinical candidates of which RP-73401 (2) and SB-207499
asthma and the widespread use of anti-inflammatory therapy, (3) are the best known representatives. While 3 is currently
morbidity and mortality due to asthma have continued to in- in phase III clinical studies in both asthma and COPD,15,16)
crease.4) As a consequence, alternative strategies to inhibit the development of compound (2) in the asthma indication
the chronic inflammation in asthmatic airways have been in- seems to have been abandoned.
vestigated with the aim of supplementing or replacing corti-
costeroids.
We report herein the structure activity of a new class of
rolipram derived PDE4 inhibitors, which were designed to
Modulation of cyclic adenosine-3,5-monophosphate (c- supplement or, ideally, replace corticosteroids as local anti-
AMP) has emerged as one of the most promising approaches inflammatory drugs. As a consequence they were designed
in the search for new anti-asthma drugs.5) Cellular levels of for delivery by inhalation. In theory this should allow for po-
this important second messenger are regulated by phosphodi- tent anti-inflammatory action in the lung, while at the same
esterase enzymes (PDE) which hydrolyze c-AMP to inactive time reducing the chance for side effects.
metabolites. While seven subtypes of the PDE family are
known, recent observations have shown that phosophodi- Chemistry
esterase-4 (PDE4) is the principle regulator of c-AMP in in-
N-Phenyl substitution of rolipram was conveniently
flammatory cells.6) Furthermore, several research groups achieved using copper catalyzed coupling with aryl halides
have reported evidence that selective PDE4 inhibitors sup- (Table 1). This method pioneered by Yamamoto17) proved to
press inflammatory mediator release from activated leuko- be very versatile for the synthesis of a wide range of N-aryl-
cytes and inhibit the recruitment of inflammatory cells into rolipram derivatives in moderate to high yield.18) In general
the airways of animals.5,7)
aryl iodides were the preferred halides in these coupling re-
Due to the presence of the enzymes in immune and in- actions, whereas bromides only gave satisfactory results in
flammatory cells and the biological role of c-AMP in these the case of bromobenzene or when electron withdrawing
cells, PDE4 inhibitors have broad anti-inflammatory effects groups were present (e.g. synthesis of 9).18) The amount of
in vitro and in vivo,5,7) however their wide distribution in the copper powder used to facilitate the reaction was adjusted ac-
body has also raised concerns about dose limiting side ef- cording to the reactivity and the availability of the aryl
fects.8) Clinical studies with rolipram and studies on animals halide. For example the readily available bromobenzene was
with a number of selective PDE4 inhibitors have shown that used in a large excess (6 eq) without any solvent in the pres-
To whom correspondence should be addressed. e-mail: jdemnitz@npd.ufpe.br
© 2001 Pharmaceutical Society of Japan

1010
Vol. 49, No. 8
Table 1. Copper Mediated Coupling of Rolipram with Arylhalides
Aryl halide
Bromobenzene
3,5-Dimethoxy-iodobenzene
3-Nitro-bromobenzene
3-Benzyloxy-iodobenzene
3-(3-Methoxybenzyloxy)iodobenzene
Methyl 5-benzyloxy-3-iodobenzoate (13)
Methoda)/time (h)/temp (°C)
R1
R2
Product (yield %)
A/3.5/150
B/4.5/reflux
B/20/150
B/22/150
C/4/140
H
OMe
NO2
OBn
H
OMe
H
H
H
6 (88)
8 (61)
9 (73)
10 (50)
11 (81)
12 (59)
OCH2C6H4–m-OMe
OBn
C/21/140
CO2Me
a) For synthetic Methods A—C, see general procedure in the experimental section.
ence of 1.85 eq of copper powder (Method A), while the phenyl rolipram derivatives. Methyl 5-hydroxy-3-iodoben-
more precious iodide (13) was reacted with rolipram in zoate19) was benzylated to give 13, which was then coupled
equimolar amounts in the presence of a large excess of cop- in the usual way with rolipram (Table 1). The resulting ester
per powder (Method C).
(12) was the key intermediate for both the 3-benzyl-
The phenol (7, see Chart 1) was used as a convenient start- oxyphenyl- and the 3-(3-methoxybenzyloxy)phenyl deriva-
ing point for the synthesis of more complex PDE4 inhibitors tives. Hydrolysis of 12 produced the acid (18). Alternatively,
(Table 2). The esters (14) and (15) were prepared in good removal of the benzyl protecting group of 12, and reaction of
yield by reacting 7 with the corresponding acyl halides in 3-methoxybenzyl chloride with the resulting phenol (19)
pyridine. The ether (17) was synthesized by alkylation of the produced 20, which was hydrolyzed to give 3-[4-(3-cy-
sodium salt of 7 with 4-methoxybenzyl chloride in DMF. clopentyloxy-4-methoxyphenyl)-2-oxo-pyrrolidin-1-yl]-5-(3-
One of the key compounds in this series 4-(3-cyclopenty- methoxybenzyloxy)benzoic acid (21). Reaction of 21 with
loxy-4-methoxyphenyl)-1-[3-(3-methoxybenzyloxy)phenyl]- N,N-dimethyl hydrazine finally produced the potent PDE4 in-
pyrrolidin-2-one (11) could be prepared either by direct cop- hibitor (22).
per catalyzed coupling of 3-(3-methoxybenzyloxy)iodoben-
zene with rolipram (Table 1), or alternatively by reaction of 7 Results and Discussion
with 3-methoxybenzyl chloride in the presence of cesium
Reference Compounds In order to compare our PDE4
carbonate (Table 2). Finally, hydrogenation of the nitro sub- isozymes with material used by other research groups, we
stituted derivative (9) with palladium on charcoal, followed prepared the reference inhibitors (2,20) 3,16) 521)) using litera-
by direct acylation of the resulting aniline with 4-methoxy- ture procedures. Table 3 summarizes the results obtained for
benzoyl chloride yielded benzamide (16).
these three compounds and our lead structure rolipram.
Chart 1 describes the synthesis of the 3,5-disubstituted N- While the results for compounds (2) and (5) are basically

August 2001
1011
Table 2. Synthesis of Compounds 11, 14 to 17a)
Halide
X
Reagents
R
Product (yield %)
m-(MeO)C6H4CH2Cl
C6H5COCl
p-(MeO)C6H4COCl
p-(MeO)C6H4COCl
p-(MeO)C6H4CH2Cl
O
O
O
NH
O
Cs2CO3
Pyridine
NaH/DMF
Pyridine
CH2C6H4–m-OMe
COC6H5
COC6H4–p-OMe
COC6H4–p-OMe
CH2C6H4–p-OMe
11 (57)
14 (72)
15 (73)
16 (99)
17 (78)
NaH/DMF
a) For detailed procedures see experimental part.
Reagents and conditions: a: rolipram, Cu, K2CO3, DMF; b: aq. NaOH, methanol; c: H2, Pd/C, methanol; d: Cs2CO3,
m-(MeO)C6H4CH2Cl, acetone; e: aq KOH, ethanol; f: Me2NNH2, EDC, HOBt, DMF.
Chart 1
identical to literature values,15) SB-207499 (3) in our hands is
slightly less potent than reported.15,16) Gratifyingly however,
we were able to confirm the selectivity of 3 for PDE-4D.
Design and Structure–Activity Relationship on the
PDE4 Isozymes In contrast to other research groups which
have mainly focused on modifications at the aromatic cate-
chol ether or the replacement of the pyrrolidone ring of
rolipram,10—16,20) we sought to find potent and selective PDE4
inhibitors through substitution at the pyrrolidone nitrogen.
Indications that this strategy might be successful had been
provided by Christensen22a) and by Bourguignon,22b) who had
shown that N-substitution of rolipram could lead to improved
PDE4 potency, when N-(p-bromobenzyl)rolipram (4a) was
synthesized for structural studies.22a)
Our studies started with the synthesis of compounds (6)
and (4b) (a close analogue of 4a). These initial targets were
chosen based on the working hypothesis that N-phenyl
roliprams would show comparable activity to known N-ben-

1012
Vol. 49, No. 8
Table 3. Inhibition of PDE4 Isoenzymes by Reference Compoundsa)
Compound
PDE4A
PDE4B
PDE4C
PDE4D
RP-73401 (2)
SB-207499 (3)
RS-25344 (5)
3.3 (Ϯ1.1)
410 (Ϯ90)
40 (Ϯ9)
2.6 (Ϯ1.3)
310 (Ϯ110)
34 (Ϯ5)
11 (Ϯ1.4)
840 (Ϯ180)
203 (Ϯ9)
1.6 (Ϯ0.4)
79 (Ϯ28)
1.8 (Ϯ0.2)
a) Data indicated as mean IC50 (ϮS.E.M.) values in nM (nϭ2—5).
Table 4. Inhibition of PDE4 Isoenzymes by N-Arylrolipram Derivativesa)
PDE4A
PDE4B
PDE4C
PDE4D
PDE3
1
4b
6
7
8
1380 (Ϯ370)
520 (Ϯ60)
440 (Ϯ110)
280 (Ϯ8)
100 (Ϯ5)
20 (Ϯ8)
128 (Ϯ41)
115 (Ϯ14)
24 (Ϯ2)
1340 (Ϯ300)
370 (Ϯ60)
430 (Ϯ21)
160 (Ϯ85)
69 (Ϯ9)
8 (Ϯ2)
25 (Ϯ2)
66 (Ϯ16)
44 (Ϯ9)
3 (Ϯ1)
26 (Ϯ10)
32 (Ϯ14)
3 (Ϯ1)
20 (Ϯ7)
5 (Ϯ1)
7940 (Ϯ240)
2480 (Ϯ750)
1650 (Ϯ50)
2430 (Ϯ140)
330 (Ϯ110)
59 (Ϯ12)
440 (Ϯ140)
1270 (Ϯ280)
280 (Ϯ80)
81 (Ϯ27)
143 (Ϯ14)
980 (Ϯ22)
320 (Ϯ22)
128 (Ϯ16)
16 (Ϯ6)
400 (Ϯ80)
320 (Ϯ120)
410 (Ϯ17)
51 (Ϯ23)
72 (Ϯ30)
13 (Ϯ6)
81 (Ϯ31)
81 (Ϯ7)
30 (Ϯ2)
8 (Ϯ1)
40 (Ϯ5)
16 (Ϯ7)
10 (Ϯ2)
13 (Ϯ2)
2 (Ϯ0.5)
Ͼ100000
Ͼ100000
Ͼ100000
Ͼ100000
n.d.
14
Ͼ100000
n.d.
10
12
18
15
16
17
11
21
22
Ͼ10000
Ͼ10000
Ͼ10000
Ͼ10000
Ͼ10000
Ͼ10000
Ͼ10000
Ͼ10000
8 (Ϯ5)
59 (Ϯ17)
40 (Ϯ10)
9 (Ϯ2)
13 (Ϯ1)
2.5 (Ϯ0.5)
a) Data indicated as mean IC50 (ϮS.E.M.) values in nM (nϭ2—6). n.d.ϭnot done.
zyl roliprams, and could therefore serve us as a starting point than the corresponding benzoate (14), but the result was
in our search for potent and selective PDE4 inhibitors. As we thought to be encouraging enough to warrant further investi-
had hoped, the simple N-phenyl substituted rolipram (6) gation. As in the benzoate series the introduction of a p-
showed improved potency compared to the lead compound methoxy group (17) lead to an improvement in the activity
(1) and was similar in activity on the PDE4 isozymes to the on all isozymes, however the most potent compound was
N-benzyl compound (4b) (Table 4). Since the ready availabil- found to be the 3-(3-methoxybenzyloxy)phenyl substituted
ity of a wide variety of phenyl halides promised to give us a rolipram (11).
synthetically very simple and direct access to a new series of
While 11 was an excellent PDE4 inhibitor, it proved to be
proprietary inhibitors, we continued to pursue derivatives of very hydrophobic (clog Pϭ5.2), which caused problems in
6. Simple introduction into the phenyl ring of nitro, amino secondary assays and in vivo models.24) Since our com-
(data not shown) or phenolic groups (e.g. 7) had very little pounds were designed for administration by inhalation, mole-
effect on the potency of the compounds compared with 6. cules with high log P were not thought to be inherently
The first breakthrough occurred with benzoate (14), a com- flawed. Indeed, experience in the field of inhaled b-agonists
pound which is easily accessible in two steps from rolipram, suggests that strongly hydrophobic compounds may be pre-
and showed over 50 times improved potency compared to our ferred in terms of duration of action. However the difficulty
lead compound (1).
of dissolving 11 in solvents suitable for in vivo studies was
Not surprisingly 14 was equipotent (within the error of the assumed to be responsible for its lack of activity in the
experiment) on the three important isozymes23) PDE4A, Brown Norway rat model of airways inflammation.24)
PDE4B and PDE4D, while PDE4C was inhibited to a slightly
In order to improve the physical properties of these
lesser extent. This pattern of activity is common for rolipram new rolipram derivatives we introduced hydrophilic sub-
based inhibitors,15) and indeed all of the inhibitors synthe- stituents in the 5-position of the N-phenyl ring in 11. Previ-
sized in this study and summarized in Table 4 are between 5 ous studies had indicated that this position may be able to
and 50 times less potent on PDE4C as compared to the other accommodate another substituent without deleterious effects
isozymes.
on PDE4 potency (see for example compound 8). Our hy-
Substitution of the benzoate (14) with a p-methoxy group pothesis proved to be correct and the newly introduced polar
(15) lead to a further improvement in activity, while the cor- groups were well tolerated. The first compound synthesized
responding benzamide (16) was considerably less active. It in this series, methyl 3-benzyloxy-5-[4-(3-cyclopentyloxy-4-
was clear however, that despite their potency, the benzoates methoxyphenyl)-2-oxo-pyrrolidin-1-yl]benzoate (12), was
(14, 15) could not be taken as serious drug candidates, since equipotent with the N-phenyl-monosubstituted analogue
they were expected to be easily hydrolized in vivo. As a con- (10). Hydrolysis of 12 lead to the carboxylic acid (18) which
sequence the corresponding ethers were investigated. The was an even more potent PDE4 inhibitor and showed much
simple 3-benzyloxy ether (10) was about 5 fold less active improved physicochemical characteristics.24,25) Replacement

August 2001
1013
of the benzyl group in 18 with a 3-methoxybenzyl group (21) Table 5. HUEOSa) and High Affinity Rolipram Binding Datab)
led to a further increase in PDE4 potency as had been ob-
Compound
HUEOS
Rolipram binding
served in the previous series. Further extension of this series
led finally to the discovery of 22, a low nanomolar inhibitor
of all PDE4 isozymes, which is more than two orders of
magnitude more potent than our lead compound rolipram.
Rolipram Binding PDE4 enzymes exist in two different
states, distinguishable by either low or high affinity for (R)-
rolipram.26) While the presence of the two forms of PDE4
has been demonstrated both in recombinant enzymes and
cellular systems,5,26) it is still unclear what mechanism regu-
lates the distribution of the two states of PDE4. By studying
the side effect profile of PDE4 inhibitors in animal models
and the resulting correlations of this pharmacological effect
with the affinity of compounds for the high and low affinity
forms of PDE4 enzymes, Torphy and coworkers8,16) have de-
veloped a hypothesis which states that inhibitors with re-
duced ability to compete for the high affinity [3H]-rolipram
binding site (h-PDE4), while maintaining potent activity in
inhibiting PDE4 catalytic activity (l-PDE4), exhibit an im-
proved therapeutic index compared to first generation PDE4
inhibitors. Since SB-207499 (3), a second generation PDE4
inhibitor with a significantly improved side effect profile,15)
was designed on this premise, it was appropriate to test our
PDE4 inhibitors for their h-PDE4 activity.
1
2
3
6
7
14
10
12
18
15
16
17
11
21
22
88 (Ϯ20)
n.d.
n.d.
4.0 (Ϯ1.5)
0.8 (Ϯ0.3)
40 (Ϯ13)
400 (Ϯ55)
97 (Ϯ66)
n.d.
6200 (Ϯ310)
640 (Ϯ140)
370 (Ϯ150)
370 (Ϯ110)
410 (Ϯ82)
170 (Ϯ40)
170 (Ϯ45)
210 (Ϯ86)
1700 (Ϯ300)
61 (Ϯ18)
38 (Ϯ8)
n.d.
180 (Ϯ20)
17.4 (Ϯ0.8)
45 (Ϯ5)
23 (Ϯ4)
n.d.
8 (Ϯ1)
32 (Ϯ4)
5 (Ϯ2)
20 (Ϯ4)
a) Human eosinophil oxidative burst. b) Data indicated as mean IC50 (ϮS.E.M.)
values in nM (nϭ2—3).
Competition of rolipram derivatives (6—22) for the [3H]-
rolipram binding site in rat brain membranes, (an assay
which has commonly been used to determine h-PDE4 activ-
ity) was measured (Table 5) and compared with the standard
compounds (1) and (3). A closer inspection of these values
suggests that the structure activity of our compounds is very
similar in both the PDE4 enzymatic assays as well as in the
assay which measures h-PDE4 activity. Figure 1 shows a
graph plotting the potency of the compounds at the two bind-
ing sites. To illustrate our findings, we have arbitrarily cho-
sen PDE4D activity for this correlation, however since none
of our compounds shows significant isozyme selectivity,
the results are similar for PDE4A and PDE4B (graph not
shown). The correlation between the potencies of our com-
pounds (6—22) at the two binding sites is surprisingly good,
giving a regression coefficient of 0.88 (S.D. 0.26; pϹ
0.001). As can be seen from Fig. 1, rolipram is the only com-
pound which clearly stands out, while SB-207499 (3) is in-
distiguishable in its behaviour from the rolipram derivatives
(6—22). The data suggest that all of our compounds have a
significantly improved l-PDE4/h-PDE4 ratio compared to
rolipram and should therefore have a side effect profile simi-
lar to SB-207499.
Cellular Assays Since PDE4 inhibitors work at an intra-
cellular level, the panel of compounds was routinely tested
in an assay of fMLP induced oxidative burst in human
eosinophils (HUEOS) (Table 5). This assay served two func-
tions, namely it allowed us to gauge the potential of the
rolipram derivatives to penetrate cellular membranes, and
secondly it allowed us to study the effect of the inhibitors in
one of the most important pro-inflammatory cells in asthma.
Reactive oxygen species and cationic proteins, which are re-
leased upon eosinophil activation, are thought to be prime
contributors to epithelial cell damage and increased hyperre-
activity of asthmatic airways.
Fig. 1. Correlation of pIC50 for PDE4D Activity versus pIC50 for Inhibi-
tion of [3H]-Rolipram Binding to Rat Brain Membranes for Compounds 6—
22
The standard compounds rolipram (n) and SB-207499 (s) are shown for compari-
son. The regression coefficient for compounds 6—22 is Rϭ0.88 (S.D. 0.26; pϹ0.001).
PDE4 inhibitors the inhibition of guinea pig eosinophil ox-
idative burst correlates with PDE4 catalytic activity.27) In our
data set the activity of the inhibitors on PDE4D shows no
significant correlation with their activity to inhibit the pro-
duction of reactive oxygen species from human eosinophils
(Rϭ0.45; S.D.ϭ0.56; pϭ0.09). Nevertheless the HUEOS
assay proved to be very useful as a secondary assay, giving
us a rapid functional readout of PDE4 inhibition, which al-
lowed us to eliminate compounds with poor cellular activity
at an early stage. Gratifyingly, both of our most promising in-
hibitors (21, 22) potently inhibited the activation of this key
effector cell in asthma.
There is convincing evidence that selected T cell popula-
tions play a key role in the orchestration of the inflammatory
response to inhaled allergens and other stimuli in asthma by
the production of a characteristic cytokine pattern.28) IL-4
and IL-5 produced by these allergen specific TH2 cells are
intimately involved in the regulation of IgE production as
well as in the development, activation and selective accumu-
lation of eosinophils, two characteristic features of allergic
asthma.
In an attempt to mimic allergen induced T cell activation
There have been suggestions in the literature that for

1014
Vol. 49, No. 8
Fig. 3. (A) Stereoselectivity of Inhibition of PDE4D Enzymatic Activity
and High Affinity [3H]-Rolipram Binding (h-PDE4) of the (R)- and (S)-
Enantiomers of Rolipram (1) and Compound 11
(B) Stereoselectivity of Anti-CD3 Induced IFN-g, IL-4 and IL-5 Production
by (R)- and (S)-11
Fig. 2. Effect of Compounds 1, 3, 21 and 22 on LPS Induced TNF-a Pro-
duction as Well as Anti-CD3 Induced IL-4, IFN-g Production and Anti-CD3
Stimulated T-Cell Proliferation from Human Peripheral Blood Mononuclear
Cells
The assays were performed with human peripheral mononuclear cells. Data shown as
mean IC50 (ϮS.E.M.) values (nϭ2—4).
For details see experimental section. Data shown as IC50s (nϭ2—6).
rality of the two enantiomers of rolipram.
events, human peripheral blood mononuclear cells (HPBMC)
were stimulated with anti-CD3 antibodies, and the effect of
selected inhibitors on cytokine release measured. Figure 2
shows the results of compounds (21) and (22) on IL-4 and
IFN-g release. While both rolipram derivatives are equipo-
tent in inhibiting the TH1 cytokine IFN-g, 22 is clearly supe-
rior in suppressing the TH2 cytokine IL-4, which has been
implicated as one of the key mediators in the induction and
maintenance of asthma. Compared to our lead compound
rolipram but also the second generation PDE4 inhibitor SB-
207499 (3), 22 shows a clearly improved profile in this key
assay.
Finally we tested the activity of our most promising com-
pounds on the production of TNF-a, a pro-inflammatory cy-
tokine primarily produced by macrophages and monocytes,
which has been shown to play an important role in the late
response to allergen challenge.29) In the event, stimulation of
HPBMC with lipopolysaccharide leads to production of
TNF-a, which was nicely suppressed by our new rolipram
analogues (Fig. 2). Again compound 22 proved to be about
50 times more potent than either 1 or 3.
Stereoselective Inhibition of PDE4 For convenience, all
of our initial work was performed with racemic inhibitors.
We have however recently developed a convenient, large
scale synthesis of (R)- and (S)-rolipram30) which provided ac-
cess to enatiomerically pure inhibitors. While 22 is one of
our most interesting compounds, we currently cannot dis-
close data on the enantiomers of this compound. Due to the
interesting findings about the inhibition stereoselectivity, we
have included data on (R)- and (S)-11, two compounds which
have been released for publication, in contrast to other PDE4
inhibitors of this series. Both enantiomers of rolipram could
be coupled with iodo-3-(3-methoxybenzyloxy)benzene under
the usual conditions to give (R)-11 and (S)-11 in greater than
99% ee as determined by chiral HPLC. The stereochemistry
of (R)-11 and (S)-11 was assigned based on the known chi-
In our hands,30) both enantiomers of rolipram are micro-
molar inhibitors of the PDE4 isozymes, the (R)-enantiomer
being the more potent inhibitor on all four isozymes, exhibit-
ing eudismic ratios31) between 3 and 5. This finding is in
overall agreement with the existing literature. While it has
been shown that the eudismic ratio can vary depending on
the source of the PDE enzyme, there is general agreement
that (R)-(Ϫ)-rolipram is the more potent of the two enan-
tiomers.22,33)
As representative examples, Fig. 3 shows a comparison of
the IC50’s for PDE4D inhibition of the (R)- and (S)-enan-
tiomers of compounds (1) (rolipram) and (11). Surprisingly
(S)-11 is approximately 30 fold more potent as an inhibitor
of the PDE4D isozyme than its corresponding (R)-enan-
tiomer. The results for the other isozymes are not shown,
since again there is no significant difference between the po-
tency and the potency ratio of the two enantiomers on the
different isozymes. In agreement with our findings with the
racemic inhibitors, h-PDE4 activity closely paralelled the
catalytic activity in (R)-11 and (S)-11 (Fig. 3). As a compari-
son Fig. 3 also shows the previously published30) h-PDE4
values for the enantimers of rolipram.
As expected this stereoselectivity of PDE4 inhibition can
also be observed in cellular assays (Fig. 3). Both of the T-cell
derived cytokines IFN-g and IL-4 are preferentially inhibit-
ed by (S)-11 (Fig. 3), even though the stereoselectivity is
slightly lower (eudismic ratioϭ10). We currently have no ex-
planation for the differing stereoselectivities between (R)-
and (S)-rolipram on the one hand, and (R)- and (S)-11 on the
other. The recent X-ray crystal structure of PDE4D could be
used to gain a better understanding of this phenomenon,
however with uncertainty over how 11 is oriented in the ac-
tive site, any explanantion about the source of this selectivity
is speculative at this moment.

August 2001
1015
cases, binding of the second antibody was analyzed by stepwise incubation
with streptavidin–alkaline phosphatase conjugate (Mabtech, Stockholm,
Sweden) and 4-nitrophenylphosphate disodium salts (Sigma Chemical Co.).
Optical density was measured at 405 nm and cytokine concentrations were
calculated based on the results from serial dilutions of standard recombinant
Conclusion
In the current study we have reported the structure–activ-
ity relationship of a new series of rolipram based PDE4 in-
hibitors. The compounds were evaluated for inhibitory activ-
ity on all four PDE4 isozymes and for selectivity against mouse IL-4 and IFN-g, respectively. The sensitivity of the cytokine ELISAs
was around 10 pg/ml.
PDE3. At the same time the compounds were tested for cel-
Induction and Measurement of TNFa Production Mononuclear cells
lular activity in a human eosinophil oxidative burst assay and
were isolated as above and resuspended in RPMI1640 supplemented with
their potential to bind to the [3H]-rolipram binding site in rat
10% FCS. Cell density was adjusted to 1ϫ106 cell/ml. Cells were stimulated
brain was assessed. This investigation lead to the identifica-
tion of a highly potent PDE4 inhibitor (22), which showed
greater than 1000 fold selectivity versus PDE3 and was an
excellent inhibitor of eosinophil oxidative burst as well as
TNF-a, IL-4 and IFN-g production from human peripheral
blood mononuclear cells. Correlations of PDE4D inhibition
potency with rolipram binding data indicate that 22 has a
similar side effect profile as the currently leading PDE4 in-
hibitor SB-207499, while being considerable more potent as
a PDE4 inhibitor in vitro as well as in all the cellular assays
examined. These new potent rolipram derivatives exhibit
stereospecific inhibition of the PDE4 isozymes, however in
contrast to rolipram the (S)-enantiomer is the more potent in-
hibitor. This effect can also be observed in primary human
cells where the (S)-enantiomer is about 10 fold more potent
than the corresponding (R)-enantiomer.
with LPS (1 mg/ml) and IFN-g (10 ng/ml) and supernatants were harvested
after 24 h of incubation at 37 °C in a humidified incubator with 5% CO2.
Concentration of TNF-a in these supernatants was measured by sandwich
ELISA using two monoclonal antibodies recognizing different epitopes of
the specific cytokine (mAb357/101-4 and biotinylated 2-179/E11). Optical
density was measured at 405 nm and cytokine concentration was calculated
based on the results from serial dilutions of standard recombinant human
TNF-a.
Rolipram Receptor-Binding Assay Method Rat brains were sus-
pended in 10 ml/g of ice-cold buffer (tris 20 mM, MgCl2 1 mM, dithiothreitol
0.1 mM, pH 7.5), homogenized using 2 bursts (30 s each) of a Polytron ho-
mogenizer (Kinematica, Switzerland), centrifuged (600ϫg, 10 min, 4 °C).
The supernatants were pooled, and centrifuged (24000ϫg, 10 min, 4 °C).
The pellets were resuspended in 5 ml/g of buffer. Protein content was deter-
mined using the method of Bradford,38) according to the instructions of the
manufacturer (BioRad). The assay was carried out as described by Schneider
et al.39)
Oxidative Burst from Human Eosinophils Blood was obtained from
normal individuals. Granulocytes were separated from mononuclear cells by
Ficoll hypaque gradient centrifugation. Erythrocytes were lysed by two cy-
cles of hypotonic lysis and the remaining granulocytes were incubated with
Experimental
General Melting points are corrected. 1H-NMR (360 MHz) spectra anti-CD16 coated immunomagnetic particles (Miltenyi Biotec, Bergisch
were recorded on a Bruker AM-360 instrument, using residual solvent pro- Gladbach, Germany). Magnetically labeled neutrophils were then depleted
tons as references. Infrared spectra were recorded on a Bruker IFS 66 appa- by passing the granluocytes through a MACS (magnetic cell separation) col-
ratus and mass spectra on a VG 70-SE instrument. Chromatography was car- umn which resulted in a more than 98% pure eosinophil preparation. These
ried out using silica gel (Merck, Kieselgel 60 F254, 230—400 mesh) and ana- purified human eosinophils were diluted in HBSS and pipetted into 96 well
lytical TLC was performed using precoated Kieselgel 60 F254 glass plates. microtitreplates (MTP) at 104 cells/well. Each well contained a 200 ml sam-
AR grade solvents (Fluka) and commercial reagents (Fluka, Aldrich) were ple comprising: 100 ml eosinophil suspension, 50 ml HBSS, 10 ml lucigenin,
used without further purification in all cases. Brine refers to a saturated 20 ml activation stimulus, 20 ml compound of interest. The samples were in-
aqueous NaCl solution and concentration implies the use of a rotary evapo- cubated with compound or vehicle for between 10 and 30 min prior to addi-
rator at water aspirator pressure.
tion of fMLP (0.01—10 mM). MTPs were agitated (Titertek MTP mixer) to
Inhibition of Human PDE3 PDE3 was prepared from human platelets facilitate mixing of the cells and medium, and the MTP placed into a Hama-
by ultrasonic homogenization. Platelets were washed once with PBS, sus- matsu luminometer. Total chemiluminescence and the temporal profile of
pended in 10 ml of buffer H (sucrose 0.25 M, EDTA 1 mM, tris–HCl 10 mM,
dithiothreitol 1 mM, pH 7.4) containing the following protease inhibitor solu-
each well was measured simultaneously over 20 min and the results ex-
pressed as a percentage of fMLP-induced chemoluminescence in the ab-
tions: 5 ml/ml of phenylmethyl-sulphonylfluoride (7 mg/ml in 2-propanol), sence of compound. Results were fitted to the Hill equation and IC50 values
1 mg/ml leupeptin and pepstatin A (1 mg/ml each in ethanol). After sonica- calculated. All compounds were dissolved in dimethyl sulphoxide and there-
tion (15 s at 4 °C, Branson probe sonicator), homogenates were centrifuged after diluted in buffer.
at 2200ϫg. The pellet was resuspended in the same volume of buffer H and
the sonication repeated. Pooled supernates were stored at Ϫ20 °C. Activity with Aryl Halides Method A: A mixture of rolipram (1), the appropriate
was assayed as described.34)
aryl halide (2—6 eq), Cu powder (1.85 eq), K2CO3 (1.5 eq) and KI (1.5 eq)
General Procedure for the Copper Catalyzed Coupling of Rolipram
Inhibition of Human cAMP-Specific Phosphodiesterase (PDE4) was heated without solvent under an Argon atmosphere at the time and tem-
Isoenzymes PDE4 activity was assessed as previously described.34) With perature indicated in Table 1.
the exception of PDE4B35) (rat; expressed in S. cerevisae), all isoenzyme
Method B: A mixture of 1, the appropriate aryl halide (1.5 eq), Cu powder
preparations were from human sources: PDE4A,36) PDE4C,34) PDE4D37) ex- (0.16 eq) and K2CO3 (1.4 eq) in dry DMF was heated under an Argon atmos-
pressed in S. cerevisae.
phere at the time and temperature indicated in Table 1.
Anti-CD3 Induced Proliferation and Cytokine Production Mononu-
Method C: A mixture of 1, the appropriate aryl halide (1.0 eq), Cu powder
clear cells (MNC) were isolated from blood of normal individuals by Ficoll- (22 eq) and K2CO3 (1.5 eq) in dry DMF was heated under an Argon atmos-
hypaque gradient centrifugation (20 min at 800ϫg). The interphase was col- phere at the time and temperature indicated in Table 1.
lected, washed twice in PBS and resuspended in RPMI1640 supplemented
Work up: The reaction mixture was filtered through a bed of hyflo and
with 10% FCS. Cell density was adjusted to 1ϫ106 cell/ml. One hundred concentrated in vacuo (Methods B and C) or treated with EtOAc whilst still
microliters of the MNC suspension was placed in 96 well culture plates and warm, worked-through with a spatula or glass rod and then filtered through
50 ml of either medium or compounds in the indicated concentrations were hyflo (Method A). The filtrate was then diluted with EtOAc, washed with
added. After a 10 min preincubation, cells were stimulated with 50 ml of water and brine, dried (Na2SO4), filtered and concentrated. The crude prod-
anti-CD3 monoclonal antibodies (OKT-3, 100 ng/ml), incubated for either ucts were purified by flash chromatography and/or recrystallization from an
24 h (cytokine production) or 48 h (proliferation assay) at 37 °C in a humidi- appropriate solvent.
fied incubator with 5% CO2. To determine the proliferative response of anti-
4-(3-Cyclopentyloxy-4-methoxyphenyl)-1-phenyl-pyrrolidin-2-one (6)
CD3 stimulated MNCs, [3H]-thymidine was added to the culture plates for mp 124—126 °C; H-NMR (CDCl3) d: 7.68 (2H, d, Jϭ7 Hz), 7.37 (2H, dd,
the last 6 h of the incubation period and incorporated radioactivity was mea- Jϭ7, 7 Hz), 7.12 (1H, dd, Jϭ7, 7 Hz), 6.99 (1H, d, Jϭ2 Hz), 6.89 (1H, d,
sured in an automated liquid scintillation counter. Cytokines were measured Jϭ8 Hz), 6.87 (1H, dd, Jϭ8, 2 Hz), 4.80 (1H, m), 4.15 (1H, dd, Jϭ9, 6 Hz),
in supernatants harvested after 24 h of incubation. IL-4 and IFN-g were 3.81 (1H, dd, Jϭ9, 9 Hz), 3.71 (3H, s), 3.62 (1H, dddd, Jϭ9, 9, 9, 6 Hz),
measured by sandwich ELISA using two monoclonal antibodies recognizing 2.83 (1H, dd, Jϭ16, 9 Hz), 2.70 (1H, dd, Jϭ16, 9 Hz), 1.95—1.47 (8H, m).
1
different epitopes of the specific cytokine. Antibodies used for measuring
IL-4 and IFN-g were purchased from Mabtech, Stockholm, Sweden. In all 75.19; H, 7.17; N, 3.99. Found: C, 74.90; H, 7.00; N, 4.00.
IR (KBr) cmϪ1: 1685. MS m/z: 351 (Mϩ). Anal. Calcd for C22H25NO3: C,

1016
4-(3-Cyclopentyloxy-4-methoxyphenyl)-1-(3,5-dimethoxyphenyl)-
Vol. 49, No. 8
HCl (15 ml) and extracted with EtOAc (3ϫ). The organic phases were
washed with water, sat. NaHCO3 and brine. Drying (Na2SO4), filtration and
concentration provided a yellow solid which was recrystallized from diiso-
propyl ether to yield a white crystalline solid (148 mg, 72%): mp 144—
146 °C: 1H-NMR (CDCl3) d: 8.20 (2H, m), 7.63 (2H, m), 7.50 (3H, m), 7.43
(1H, m), 7.03 (1H, m), 6.83—6.77 (3H, m), 4.76 (1H, m), 4.19 (1H, dd,
Jϭ9, 7 Hz), 3.85 (1H, dd, Jϭ9, 7 Hz), 3.82 (3H, s), 3.61 (1H, m), 3.00 (1H,
dd, Jϭ16, 9 Hz), 2.78 (1H, dd, Jϭ16, 9 Hz), 1.95—1.51 (8H, m). IR (KBr)
cmϪ1: 1726, 1699. MS m/z: 471 (Mϩ). Anal. Calcd for C29H29NO5: C, 73.87;
H, 6.20; N, 2.97. Found: C, 73.60; H, 6.30; N, 3.00.
pyrrolidin-2-one (8) mp 52—54 °C. 1H-NMR (CDCl3) d: 6.99 (1H, d,
Jϭ1 Hz), 6.95—6.82 (4H, m), 6.32 (1H, br s), 4.80 (1H, m), 4.14 (1H, br dd,
Jϭ9, 9 Hz), 3.75 (1H, m), 3.74 (6H, s), 3.72 (3H, s), 3.60 (1H, m), 2.83 (1H,
dd, Jϭ16, 9 Hz), 2.70 (1H, dd, Jϭ16, 9 Hz), 1.95—1.47 (8H, m). IR (KBr)
cmϪ1: 1692. MS m/z: 411 (Mϩ). Anal. Calcd for C24H29NO5: C, 70.05; H,
7.10; N, 3.40. Found: C, 69.70; H, 7.40; N, 3.60.
4-(3-Cyclopentyloxy-4-methoxyphenyl)-1-(3-nitrophenyl)-pyrrolidin-
2-one (9) mp 99—101 °C. 1H-NMR (CDCl3) d: 8.36 (1H, dd, Jϭ1, 1 Hz),
8.21 (1H, ddd, Jϭ7, 1, 1 Hz), 8.00 (1H, ddd, Jϭ7, 1, 1 Hz), 7.55 (1H, dd,
Jϭ7, 7 Hz), 6.88—6.79 (3H, m), 4.78 (1H, m), 4.23 (1H, dd, Jϭ9, 7 Hz),
3.90 (1H, dd, Jϭ9, 7 Hz), 3.85 (3H, s), 3.69 (1H, dddd, Jϭ9, 9, 9, 9 Hz),
3.05 (1H, dd, Jϭ16, 9 Hz), 2.83 (1H, dd, Jϭ16, 9 Hz), 1.98—1.50 (8H, m).
IR (KBr) cmϪ1: 1698. MS m/z: 396 (Mϩ).
4-(3-Cyclopentyloxy-4-methoxyphenyl)-1-[3-(p-methoxybenzoy-
loxyphenyl)]-pyrrolidin-2-one (15) To
a mixture of NaH (18 mg,
0.45 mmol, 60% dispersion in oil) in dry DMF (2 ml) under argon was added
the phenol (7) (150 mg, 0.41 mmol) dissolved in dry DMF (2 ml). The re-
sulting gray-brown suspension was stirred for 1 h at room temperature, dur-
ing which time the solid material slowly went into solution, and then p-
methoxybenzoyl chloride (91 mg, 0.53 mmol, 1.3 eq) was added. After stir-
ring at room temperature for 1 h, more NaH (5 mg, 0.12 mmol) and p-
methoxybenzoyl chloride (20 mg, 0.12 mmol, 0.3 eq) were added and stirring
continued for 30 min. The solvent was removed in vacuo and the oily residue
dissolved in EtOAc, washed with water, 2 N aq. NaOH and brine. Drying
(Na2SO4), filtration and concentration provided an opaque yellow oil (240
mg) which was chromatographed on silica (eluent: hexane/EtOAc, 2/1) to
give a colorless oil that crystallized slowly. Recrystallization from 2-
propanol afforded white needles in two crops (150 mg, 73 %): mp 90—
92 °C: 1H-NMR (CDCl3) d: 8.15 (2H, d, Jϭ7 Hz), 7.60 (1H, dd, Jϭ1, 1 Hz),
7.52 (1H, ddd, Jϭ7, 1, 1 Hz), 7.41 (1H, dd, Jϭ7, 7 Hz), 7.01 (1H, ddd, Jϭ7,
1, 1 Hz), 6.99 (2H, d, Jϭ7 Hz), 6.87—6.78 (3H, m), 4.77 (1H, m), 4.19 (1H,
dd, Jϭ9, 7 Hz), 3.90 (3H, s), 3.86 (1H, dd, Jϭ9, 7 Hz), 3.83 (3H, s), 3.62
(1H, dddd, Jϭ9, 9, 9, 9 Hz), 3.00 (1H, dd, Jϭ16, 9 Hz), 2.78 (1H, dd, Jϭ16,
9 Hz), 1.98—1.51 (8H, m). IR (KBr) cmϪ1: 1731, 1701. MS m/z: 501 (Mϩ).
Anal. Calcd for C30H31NO6: C, 71.84; H, 6.23; N, 2.79. Found: C, 71.60; H,
6.04; N, 2.76.
4-(3-Cyclopentyloxy-4-methoxyphenyl)-1-[3-(p-methoxybenzami-
dophenyl)]-pyrrolidin-2-one (16) A solution of the nitro compound (9)
(2 g, 5.05 mmol) in MeOH (100 ml) containing three drops of NEt3 was hy-
drogenated for 2 h under 1 atm at room temperature in the presence of 10%
Pd–C (200 mg). Filtration through hyflo and concentration afforded a light-
gray foam (1.85 g, 100%). To a portion of this crude aniline (200 mg, 0.546
mmol) in dry pyridine (2 ml) was added, all at once, p-methoxybenzoyl chlo-
ride (121 mg, 0.71 mmol, 1.3 eq). After stirring at room temperature for 2 h,
the reaction mixture was poured into 3 N HCl (15 ml) and extracted with
EtOAc (3ϫ). The organic phases were washed with water, sat. NaHCO3 and
brine. Drying (Na2SO4), filtration and concentration provided an opaque,
light-gray semi-solid residue (320 mg) which was chromatographed twice on
silica (eluent: hexane/EtOAc, 1/1) to give a colorless foam (210 mg, 99%):
1H-NMR (CDCl3) d: 8.04 (1H, br s), 7.87 (1H, br s), 7.85 (2H, d, Jϭ9 Hz),
7.55 (1H, m), 7.37 (2H, m), 6.98 (2H, d, Jϭ9 Hz), 6.88—6.79 (3H, m), 4.78
(1H, m), 4.21 (1H, dd, Jϭ9, 7 Hz), 3.88 (1H, dd, Jϭ9, 7 Hz), 3.88 (3H, s),
3.84 (3H, s), 3.62 (1H, dddd, Jϭ9, 9, 9, 9 Hz), 3.00 (1H, dd, Jϭ16, 9 Hz),
2.78 (1H, dd, Jϭ16, 9 Hz), 1.98—1.51 (8H, m). IR (KBr) cmϪ1: 3300, 1700,
1674. MS m/z: 501 (MHϩ). Anal. Calcd for C30H32N2O5: C, 71.98; H, 6.44;
N, 5.60. Found: C, 71.74; H, 6.32; N, 5.77.
4-(3-Cyclopentyloxy-4-methoxyphenyl)-1-[3-(p-methoxybenzyloxy)-
phenyl]-pyrrolidin-2-one (17) To a 60% dispersion of NaH (18 mg, 0.45
mmol, 1.1 eq) in dry DMF (2 ml) under argon was added during 5 min a so-
lution of the phenol (7) (150 mg, 0.41 mmol) in dry DMF (2 ml). The result-
ing gray-brown suspension was stirred for 1 h at room temperature during
which time the solid material slowly went into solution. This was now
treated all at once with p-methoxybenzyl chloride (91 mg, 0.53 mmol,
1.3 eq) and stirred for 1 h. More NaH (5 mg, 0.12 mmol, 0.3 eq) and p-
methoxybenzyl chloride (20 mg, 0.12 mmol, 0.3 eq) were added sequentially
before stirring for another 30 min. The reaction mixture was then concen-
trated in vacuo, taken up in EtOAc and washed with water, 2 N NaOH (2ϫ)
and brine. Drying (Na2SO4), filtration and concentration provided an
opaque, yellow oil (240 mg). Chromatography on silica (eluent: hexane/
EtOAc, 2/1) gave a colorless oil which was crystallized from i-PrOH in two
crops affording the ether 17 (150 mg, 78%) as white needles: mp 130—
133 °C; 1H-NMR (CDCl3) d: 7.52 (2H, d, Jϭ7 Hz), 7.41 (1H, dd, Jϭ7,
7 Hz), 7.01 (1H, ddd, Jϭ7, 1, 1 Hz), 6.99 (2H, d, Jϭ7 Hz), 6.88—6.78 (5H,
m), 5.04 (2H, s), 4.77 (1H, m), 4.19 (1H, dd, Jϭ9, 7 Hz), 3.90 (3H, s), 3.86
(1H, dd, Jϭ9, 7 Hz), 3.83 (3H, s), 3.62 (1H, dddd, Jϭ9, 9, 9, 9 Hz), 3.00
(1H, dd, Jϭ16, 9 Hz), 2.78 (1H, dd, Jϭ16, 9 Hz), 1.98—1.51 (8H, m). IR
(KBr) cmϪ1: 1685. MS m/z: 487 (Mϩ). Anal. Calcd for C30H33NO5: C,
1-(3-Benzyloxyphenyl)-4-(3-cyclopentyloxy-4-methoxyphenyl)-pyrro-
1
lidin-2-one (10) mp 103—105 °C. H-NMR (CDCl3) d: 7.48—7.23 (7H,
m), 7.16 (1H, dd, Jϭ7, 2 Hz), 6.88—6.75 (4H, m), 5.09 (2H, s), 4.77 (1H,
m), 4.15 (1H, dd, Jϭ9, 7 Hz), 3.85 (3H, s), 3.82 (1H, dd, Jϭ9, 7 Hz), 3.61
(1H, dddd, Jϭ9, 9, 9, 9 Hz), 2.99 (1H, dd, Jϭ16, 9 Hz), 2.76 (1H, dd, Jϭ16,
9 Hz), 1.98—1.51 (8H, m). IR (KBr) cmϪ1: 1686. MS m/z: 458 (MHϩ).
Anal. Calcd for C29H31NO4: C, 76.12; H, 6.83; N, 3.06. Found: C, 76.10; H,
6.90; N, 3.30.
4-(3-Cyclopentyloxy-4-methoxyphenyl)-1-[3-(3-methoxybenzyloxy)-
1
phenyl]-pyrrolidin-2-one (11) mp 73—74 °C. H-NMR (CDCl3) d: 7.43
(1H, dd, Jϭ1, 1 Hz), 7.32—7.25 (2H, m), 7.16 (1H, m), 7.01 (2H, m),
6.89—6.75 (5H, m), 5.05 (2H, s), 4.77 (1H, m), 4.15 (1H, dd, Jϭ9, 7 Hz),
3.85 (3H, s), 3.83 (1H, dd, Jϭ9, 7 Hz), 3.83 (3H, s), 3.61 (1H, dddd, Jϭ9, 9,
9, 9 Hz), 3.00 (1H, dd, Jϭ16, 9 Hz), 2.77 (1H, dd, Jϭ16, 9 Hz), 1.95—1.50
(8H, m). IR (KBr) cmϪ1: 1687. MS m/z: 487 (Mϩ). Anal. Calcd for
C30H33NO5: C, 73.89; H, 6.77; N, 2.87. Found: C, 73.64; H, 6.74; N, 3.04.
Alternative Preparation of (11) A mixture of the phenol (7, vide infra)
(200 mg, 0.545 mmol), m-methoxybenzylchloride (116 mg, 0.746 mmol,
1.3 eq) and Cs2CO3 (266 mg, 0.817 mmol, 1.5 eq) in isobutylmethyl ketone
(12 ml) was stirred at room temperature for 2 h, at 60 °C for 1.5 h and at
80 °C for 15 h. The resulting light-gray suspension was filtered through hyflo
and the filtrate concentrated in vacuo. The residue was taken up in EtOAc
and washed with 2 N NaOH and brine. Drying (Na2SO4), filtration and con-
centration provided a yellow oil (302 mg) which was chromatographed on
silica (eluent: hexane/EtOAc, 3/1) to give a clear, colorless oil, crystalliza-
tion of which from i-PrOH furnished the ether (11) as white blocks (150 mg,
57%).
R-(؉)-4-(3-Cyclopentyloxy-4-methoxyphenyl)-1-[3-(3-methoxybenzyl-
oxy)-phenyl]-pyrrolidin-2-one ((R)-(؉)-11) For spectral data see above
for compound 11: mp 61—62 °C (i-Pr2O/hexane/pentane). [a]D20 ϩ7.1°
(cϭ0.31, MeOH). CD Spectrum.30)
S-(؊)-4-(3-Cyclopentyloxy-4-methoxyphenyl)-1-[3-(3-methoxybenzyl-
oxy)-phenyl]-pyrrolidin-2-one ((S)-(؊)-11) For spectral data see above
for compound 11: mp 57—59 °C (i-Pr2O/hexane/pentane). [a]D20 Ϫ6.2°
(cϭ0.50, MeOH). CD Spectrum.30)
Methyl 3-Benzyloxy-5-[4-(3-cyclopentyloxy-4-methoxyphenyl)-2-oxo-
pyrrolidin-1-yl]-benzoate (12) Amorphous solid. 1H-NMR (CDCl3) d:
7.87 (1H, m), 7.65 (1H, m), 7.50—7.30 (6H, m), 6.90—6.70 (3H, m), 5.13
(2H, s), 4.77 (1H, m), 4.20 (1H, dd, Jϭ10, 8 Hz), 3.90 (3H, s), 3.88 (1H, dd,
Jϭ10, 8 Hz), 3.82 (3H, s), 3.64 (1H, m), 2.99 (1H, dd, Jϭ16, 9 Hz), 2.79
(1H, dd, Jϭ16, 9 Hz), 2.00—1.50 (8H, m). IR (KBr) cmϪ1: 1704. MS m/z:
516 (MHϩ). Anal. Calcd for C31H33NO6: C, 72.21; H, 6.45; N, 2.72. Found:
C, 72.33; H, 6.61; N, 2.75.
4-(3-Cyclopentyloxy-4-methoxyphenyl)-1-(3-hydroxyphenyl)-pyrro-
lidin-2-one (7) A solution of the benzyl ether (10) (900 mg, 1.97 mmol) in
MeOH (60 ml) was hydrogenolysed for 18 h under 1 atm at room tempera-
ture in the presence of 10% Pd–C (90 mg). Filtration through hyflo and con-
centration afforded a light-gray solid (730 mg, 100%). A portion of this
product was recrystallized from i-PrOH/i-Pr2O to give the pure phenol (7)
as white needles: mp 158—160 °C. 1H-NMR (CDCl3) d: 7.96 (1H, dd, Jϭ1,
1 Hz), 7.78 (1H, s), 7.22 (1H, dd, Jϭ7, 7 Hz), 6.86—6.66 (5H, m), 4.76 (1H,
m), 4.20 (1H, dd, Jϭ9, 7 Hz), 3.87 (1H, dd, Jϭ9, 7 Hz), 3.84 (3H, s), 3.63
(1H, dddd, Jϭ9, 9, 9, 9 Hz), 3.05 (1H, dd, Jϭ16, 9 Hz), 2.83 (1H, dd, Jϭ16,
9 Hz), 1.98—1.51 (8H, m). IR (KBr) cmϪ1: 3650, 1669. MS m/z: 367 (Mϩ).
Anal. Calcd for C22H25NO4: C, 71.91; H, 6.86; N, 3.81. Found: C, 71.82; H,
6.79; N, 4.10.
4-(3-Cyclopentyloxy-4-methoxyphenyl)-1-(3-benzoyloxyphenyl)-
pyrrolidin-2-one (14) To the phenol (7) (160 mg, 0.44 mmol) in dry pyri-
dine (2 ml) was added benzoyl chloride (74 mg, 0.52 mmol, 1.2 eq). After
stirring at room temperature for 2 h, the reaction mixture was poured into 3 N

August 2001
1017
73.89; H, 6.77; N, 2.87. Found: C, 73.60; H, 6.61; N, 2.83.
Higgs G., Br. J. Pharmacol., 118, 1183—1191 (1996).
8) Barnette M. S., Christensen S. B., Underwood D. C., Torphy T. J.,
Pharmacol. Rev. Comm., 8, 65—73 (1996).
3-Benzyloxy-5-[4-(3-cyclopentyloxy-4-methoxyphenyl)-2-oxo-pyrro-
lidin-1-yl]-benzoic acid (18) The methyl ester (12) (110 mg, 0.2 mmol)
was dissolved in methanol (5 ml) and excess 2 N NaOH (0.5 ml) was added.
The resulting solution was stirred for 48 h at room temperature and the reac-
tion mixture was then evaporated to dryness. The residue was dissolved in
CH2Cl2 and washed with 2 N HCl and brine. Drying (Na2SO4), filtration and
9) Kidney J., Dominguez M., Taylor P. M., Rose M., Chung K. F., Barnes
P. J., Am. J. Resp. Crit. Care Med., 151, 1907—1914 (1995).
10) Raeburn D., Underwood S. L., Lewis S. A., Woodman W. R., Battram
C. H., Tomkinson A., Sharma S., Jordan R., Souness J. E., Webber S.
E., Karlsson J., Br. J. Pharmacol., 113, 1423—1431 (1994).
11) Stafford J. A., Valvano N. L., Feldman P. L., Sloan-Brawley E., Cowan
D. J., Domanico P. L., Leesnitzer M. A., Rose D. A., Simpson S. A.,
Strickland A. B., Unwalla R. J., Verghese M. W., Bioorg. Med. Chem.
Lett., 5, 1977—1982 (1995).
12) Masamune H., Cheng J. B., Cooper K., Eggler J. F., Marfat A., Mar-
shall S. C., Shirley J. T., Tickner J. E., Umland J. P., Vazquez E.,
Bioorg. Med. Chem. Lett., 5, 1965—1968 (1995).
13) Kleinman E. F., Campbell E., Giordano L. A., Cohan V. L., Jenkinson
T. H., Cheng J. B., Shirley J. T., Pettipher E. R., Salter E. D., Hibbs T.
A., DiCapua F. M., Bordner J., J. Med. Chem., 41, 266—270 (1998).
14) Wei H., Huang F., Hanney B., Souness J., Miller B., Liang G., Mason
J., Djuric S., J. Med. Chem., 41, 4216—4223 (1998).
1
concentration gave a white solid (75 mg, 75%), mp 200—202 °C: H-NMR
(CDCl3) d: 7.92 (1H, s), 7.60—7.30 (7H, m), 6.98 (1H, s), 6.87 (2H, m),
5.18 (2H, s), 4.80 (1H, m), 4.10 (1H, m), 3.83 (1H, m), 3.70 (3H, s), 3.64
(1H, m), 2.86 (1H, dd, Jϭ16, 9 Hz), 2.72 (1H, dd, Jϭ16, 9 Hz), 1.90—1.50
(8H, m). IR (KBr) cmϪ1: 1735. MS m/z: 502 (MHϩ).
Methyl 5-[4-(3-Cyclopentyloxy-4-methoxyphenyl)-2-oxo-pyrrolidin-1-
yl]-3-(3-methoxybenzyloxy)-benzoate (20) This compound was prepared
according to the abovementioned alternative procedure described for 11
(from 10), starting in this case from methyl 3-benzyloxy-5-[4-(3-cyclopenty-
loxy-4-methoxyphenyl)-2-oxo-pyrrolidin-1-yl]-benzoate (12), via hy-
drogenolysis to 19 (cf. 10→7) followed by 3-methoxybenzylation to 20 (cf.
7→11). Amorphous solid: 1H-NMR (DMSO-d6) d: 7.98 (1H, m), 7.56 (1H,
m), 7.50—7.20 (2H, m), 7.10—6.80 (6H, m), 5.15 (2H, s), 4.71 (1H, m),
4.18 (1H, t, Jϭ8 Hz), 3.83 (3H, s), 3.83 (1H, t, Jϭ8 Hz), 3.75 (3H, s), 3.71
(3H, s), 3.64 (1H, m), 2.85 (1H, dd, Jϭ16, 8 Hz), 2.74 (1H, dd, Jϭ16, 8 Hz),
1.90—1.50 (8H, m). IR (KBr) cmϪ1: 1705. MS m/z 546 (MHϩ). Anal. Calcd
for C32H35NO7: C, 70.44; H, 6.47; N, 2.57. Found: C, 70.22; H, 6.45; N,
2.40.
5-[4-(3-Cyclopentyloxy-4-methoxyphenyl)-2-oxo-pyrrolidin-1-yl]-3-(3-
methoxy-benzyloxy)benzoic Acid (21) Compound (20) (390 mg, 0.71
mmol) was dissolved in ethanol (15 ml) and treated with 2 N NaOH (1.25 ml,
2.5 mmol). After 5 h of stirring at room temperature the volume of the reac-
tion mixture was reduced to ca. 1 ml and 2 N HCl was added (50 ml). The
mixture was extracted with CH2Cl2 (2ϫ) and the combined organic extracts
dried (Na2SO4), filtered and concentrated to afford a yellow solid which was
recrystallized from 2-propanol/diisopropylether (260 mg, 68%): 1H-NMR
(DMSO-d6) d: 13.0 (1H, s), 7.93 (1H, s), 7.56 (1H, s), 7.30 (2H, m), 7.1—
6.8 (6H, m), 5.12 (2H, s), 4.80 (1H, m), 4.18 (1H, m), 3.83 (1H, m), 3.74
(3H, s), 3.71 (3H, s), 3.63 (1H, m), 2.85 (1H, dd, Jϭ16, 9 Hz), 2.71 (1H, dd,
Jϭ16, 9 Hz), 1.90—1.50 (8H, m). IR (KBr) cmϪ1: 3300—2700, 1712, 1660.
MS m/z: 532 (MHϩ). Anal. Calcd for C31H33NO7: C, 70.04; H, 6.26; N, 2.63.
Found: C, 69.76; H, 6.33; N, 2.52.
15) Silvestre J., Graul A., Castanier J., Drugs Future, 23, 607—615
(1998).
16) Christensen S. B., Guider A., Forster C. J., Gleason J. G., Bender P. E.,
Karpinski J. M., DeWolf Jr., W. E., Barnette M. S., Underwood D. C.,
Griswold D. E., Cieslinski L. B., Burman M., Bochnowicz S., Osborne
R. R., Manning C. D., Grous M., Hillegas L. M., Bartus-O’Leary J.,
Ryan M. D., Eggleston D. S., Haltiwanger R. C., Torphy T. J., J. Med.
Chem., 41, 821—835 (1998).
17) Yamamoto T., Kurata Y., Can. J. Chem., 61, 86—91 (1983).
18) Aebischer E., Bacher E., Demnitz F. W. J., Keller Th. H., Kurzmeier
M., Ortiz M. L., Pombo-Villar E., Weber H.-P., Heterocycles, 48,
2225—2229 (1998).
19) Counsell R. E., Smith T. D., Ranade V. V., Noronha O. P. D., Desai P.,
J. Med. Chem., 16, 684—687 (1973).
20) Ashton M. J., Cook D. C., Fenton G., Karlsson J. A., Palfreyman M.
N., Raeburn D., Ratcliff A. J., Souness J. E., Thurairaitnam S., Vicker
N., J. Med. Chem., 37, 1696—1703 (1994).
21) Wilhelm R. S., Chin R. L., Devens P. H., Alvarez R., PCT Int. Appl.,
WO 9319068.
22) a) Baures P. W., Eggleston D. S., Erhard K. F., Cieslinski L.B., Torphy
T. J., Christensen S. B., J. Med. Chem., 36, 3274—3277 (1993); b)
Marivet M. C., Bourguignon J. J., Lugnier C., Mann A., Stoclet J. C.,
Wermuth C. G., ibid., 32, 1450—1457 (1989).
23) Müller T, Engels P., Fozard J. R., Trends Pharmacol. Sci., 17, 295—
298 (1996).
24) Bacher E., Boer Ch., Bray-French K., Demnitz F. W. J., Keller Th. H.,
Mazzoni L., Mueller Th., Walker C., Bioorg. Med. Chem. Lett., 8,
3229—3234 (1998).
25) Clog P: 10ϭ5.3; 11ϭ5.1; 22ϭ4.2. Clog D (pH 7.4): 21ϭ2.4. Solubil-
ity of compounds is given in ref. 24).
26) Rocque W. J., Tian G., Wiseman J. S., Holmes W. D., Zajac-Thompson
I., Willard D. H., Patel I. R., Wisely G. B., Clay W. C., Kadwell S. H.,
Hoffman C. R., Luther M. A., Biochemistry, 36, 14250—14261
(1997).
27) Barnette M. S., Manning C. D., Cieslinski L. B., Christensen S. B.,
Torphy T. J., J. Pharmacol. Exp. Therap., 273, 674—679 (1995).
28) Romagnani S., Curr. Opin. Immunol., 6, 838—846 (1994).
29) Keatings V. M., O’Connor B. J., Wright L. G., Huston D. P., Corrigan
C. J., Barnes P. J., J. Allergy Clin. Immunol., 99, 693—698 (1997).
30) Demnitz J., LaVecchia L., Bacher E., Keller Th. H., Muller Th.,
Schürch F., Weber H.-P., Pombo-Villar E., Molecules, 3, 107—119
(1998).
31) The eudismic ratio is the ratio of activities (affinities, potencies, etc.)
of the two enantiomers in a specific biological action.32)
32) Lehman F., Quant. Struct.-Act. Relat., 6, 57—65 (1987).
33) Souness J. E., Scott L. C., Biochem. J., 291, 389—395 (1993).
34) Engels P., Sullivan M., Müller Th., Lübbert H., FEBS Lett., 358, 305—
310 (1995).
5-[4-(3-Cyclopentyloxy-4-methoxyphenyl)-2-oxo-pyrrolidin-1-yl]-3-(3-
methoxybenzyloxy)benzoic Acid N؅,N؅-Dimethylhydrazide (22) The
acid (21) (200 mg, 0.38 mmol) was dissolved in DMF (20 ml) and 1,1-di-
methylhydrazine (0.11 ml, 1.5 mmol), 4-hydroxybenzotriazol (62 mg, 0.46
mmol), N-methyl morpholine (0.05 ml, 0.46 mmol) and EDC (180 mg, 0.93
mmol) were added. The yellow solution was stirred overnight and the sol-
vent evaporated to dryness. The residue was dissolved in EtOAc, washed
with 2 N HCl and brine, dried (Na2SO4), filtered and concentrated to afford a
brown solid which was purified by column chromatography (silica gel;
1
EtOAc/MeOH 9 : 1) to give 22 as an amorphous solid (156 mg, 72%): H-
NMR (DMSO-d6) d: 9.39 (1H, s), 7.66 (1H, s), 7.55 (1H, s), 7.33—6.80
(8H, m), 5.13 (2H, s), 4.82 (1H, m), 4.16 (1H, m), 3.82 (1H, m), 3.73 (3H,
s), 3.71 (3H, s), 3.62 (1H, m), 2.85 (1H, dd, Jϭ16, 9 Hz), 2.73 (1H, dd,
Jϭ16, 9 Hz), 2.58 (6H, s), 1.90—1.50 (8H, m). IR (KBr) cmϪ1: 1701, 1646.
MS m/z: 596 ([MϩNa]ϩ), 574 (MHϩ). Anal. Calcd for C33H39N3O6: C,
69.09; H, 6.85; N, 7.32. Found: C, 69.41; H, 6.97; N, 7.54.
References and Notes
1) Present address: Universidade Federal de Pernambuco, Departamento
de Química Fundamental, Cidade Universitária, 50.670–901 Recife-
Pe, Brasil.
2) Holgate S. T., J. Allergy Clin. Immunol., 98, S7—S16 (1996).
3) Calpin C., Macarthur C., Stephens D., Feldman W., Parkin P. C., J. Al-
lergy Clin. Immunol., 100, 452—457 (1997).
4) The international study of asthma and allergies in childhood (ISAAC)
steering committee. World-wide variation in prevalence of symptoms
of asthma, allergic rhinoconjunctivitis, and atopic exzema, ISSAC,
Lancet, 351, 1225—1232 (1998).
5) Schudt C., Dent G., Rabe K. F. (eds.), “Phosphodiesterase Inhibitors,”
Academic Press, London, 1996.
35) Colicelli J., Birchmeier C., Michael T., O’Neill K., Riggs M., Wigler
M., Proc. Natl. Acad. Sci. U.S.A., 86, 3599—3603 (1989).
36) Sullivan M., Egerton M., Shakur Y., Marquardsen A., Houslay M. D.,
Cell Signal., 6, 793—812 (1994).
6) Torphy T. J., Cieslinski L. B., Mol. Pharmacol., 37, 206—214 (1990).
7) Hughes G., Howat D., Lisle H., Holbrook M., James T., Gozzard K.,
Blease K., Hughes P., Kingaby R., Warrellow G., Alexander R., Head
J., Boyd E., Eaton M., Perry M., Wales M., Smith B., Owens R., Cat-
terall C., Lumb S., Russell A., Allen R., Merriman M., Blowham D.,
37) Baecker P. A., Obernolte R., Bach C., Yee C., Shelton E. R., Gene,
138, 253—256 (1994).
38) Bradford M. M., Anal. Biochem., 72, 248—254 (1976).
39) Schneider H. H., Schmiechen R., Brezinski M., Seidler J., Eur. J.
Pharmacol., 127, 105—115 (1986).