Discovery of thiadiazoles as a novel structural class of potent and selective PDE7 inhibitors. Part 2: Metabolism-directed optimization studies towards orally bioavailable derivatives

Article abstract of DOI:10.1016/j.bmcl.2004.07.009

The synthesis and optimization of pharmacokinetic parameters of structurally novel small PDE7 inhibitors is discussed.

Full text of DOI:10.1016/j.bmcl.2004.07.009

Bioorganic & Medicinal Chemistry Letters 14 (2004) 4615–4621
Discovery of thiadiazoles as a novel structural class of potent
and selective PDE7 inhibitors. Part 2: Metabolism-directed
optimization studies towards orally bioavailable derivatives
a
a
a
Fabrice Vergne,a, Patrick Bernardelli, Edwige Lorthiois, Nga Pham,
*
Emmanuelle Proust,a Chrystelle Oliveira,a Abdel-Kader Mafroud,a Pierre Ducrot,a
Roger Wrigglesworth,a Franc¸oise Berlioz-Seux,a Francis Coleon,a Eric Chevalier,a
Franc¸ois Moreau,a Moulay Idrissi,a Anita Tertre,a Arnaud Descours,a
Patrick Bernaa and Mei Lib
aPfizer Global Research and Development, 3-9 Rue de la loge 94265 Fresnes, France
bPfizer Global Research and Development, Eastern Point Road, Groton CT 063409, USA
Received 23 January 2004; revised 18 June 2004; accepted 2 July 2004
Available online 7 August 2004
Abstract—The synthesis and optimization of pharmacokinetic parameters of structurally novel small PDE7 inhibitors is discussed.
ꢀ 2004 Elsevier Ltd. All rights reserved.
1. Introduction
Phosphodiesterases (PDEs) are part of a family of en-
zymes, which play a role in signal transduction by regu-
lating cellular concentrations of cAMP. Aberrant
regulation of cAMP levels is implicated in many diseases
like asthma, inflammation, immuno-related disorders,
allergy and cancer.1 And, it has been shown that inhibi-
Figure 1. Schematic representation of 1.
tion of cAMP-specific PDE activity (e.g. with PDE4
inhibitors) causing an increase in intracellular concen-
trations of cAMP could be beneficial for the treatment
of several diseases.2 PDE7 is a recently described high
affinity cAMP-specific PDE (Km =200nM) whose func-
tional role in T-cells3 has been demonstrated. Recent
findings on tissue distribution support the hypothesis
that PDE7 could be a good target for the treatment of
airway diseases,4 T-cell related diseases,3 CNS disor-
ders,5 leukemia,6 cardiovascular diseases7 and fertility
disorders.8 Therefore, the identification of selective
inhibitors targeted against PDE7 enzyme has become
an attractive area of research.9 In this context, we have
decided to design bioavailable PDE7 inhibitors to pur-
sue our studies in animal models.
In our previous publication,10 we described the discovery
of highly potent and selective PDE7 inhibitors. Among
them, 1 (Fig. 1) is the representative of one chemical
sub-series. Herein, we would like to report our efforts di-
rected towards the discovery of orally bioavailable PDE7
inhibitors starting from this compound (Fig. 1).
2. Chemistry
The thiadiazole derivatives exemplified in this report
were prepared according to the routes described11 in
Scheme 1 (some of these compounds can also be synthe-
sized by alternative pathways).10 The common reaction
conditions leading to the key intermediate 5 were first
Keywords: PDE7; Phosphodiesterase; ADME; cAMP.
*
Corresponding author at present address: Aventis Pharma, 13 Quai
Jules Guesde, 94400 Vitry-sur-Seine, France. Tel.: +33-158932962;
fax: +33-158938014; e-mail: fabrice.vergne@aventis.com
0960-894X/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.07.009

4616
F. Vergne et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4615–4621
Scheme 1. Reagents and conditions: (a) CS2, MeI, KOH, EtOH, 0–15ꢁC; (b) R2PhCOCl, toluene, reflux; (c) Ac2O, HClO4, Et2O, À5 to 5ꢁC or
TMSOTf, CH2Cl2, rt (compounds 17 and 20); (d) R1NH2, alcohol (EtOH), Et3N, 40–80ꢁC; (e) H2O2 (30% in water), Na2CO3 (3N), EtOH, rt; (f)
KOH (6N), isopropanol, reflux; (g) SOCl2, toluene, reflux; (h) ammonia; (i) HNR3R4, BOP, i-Pr2NEt, HOAt, DMF.
reported by Molina et al.12 Methyl hydrazine (2) was
reacted with carbon disulfide and with methyl iodide
in presence of potassium hydroxide to form the hydr-
azine carbodithioate (3). Acylation of 3 followed by
intramolecular cyclization in presence of acetic anhy-
dride and perchloric acid in diethylether (or using
TMSOTf in CH2Cl2) as reagents gives the correspond-
ing intermediate 1,3,4-thiadiazolium salts 5 (ClOÀ4 or
OTfÀ, respectively). Compound 5 is then reacted with
suitable amines to yield the targeted thiadiazoles (6,
10–12, 17, 20–21 and 21a). Reaction of the appropriate
benzonitrile 6 with KOH (6N) in isopropanol at reflux
gave the carboxylic acid derivative 7 and 8, respectively
(7, when R1 =cyclohexyl and 8, when R1 =trans-3-hy-
droxycyclohexyl). Compounds 9, 13–16 and 16a–b were
prepared from 6 by transformation of the nitrile to the
amide function with H2O2 (30% in water) as reagent in
presence of a solution of sodium carbonate. The cyclo-
hexyl derivative 7 is easily converted to compound 1
by transformation to the corresponding benzoyl chlo-
ride 22 with thionyl chloride, followed by a condensa-
tion with ammonia. 18 and 19 were, respectively,
prepared by reaction of 7 and 8 with the corresponding
primary amines in presence of the BOP coupling agent
in dimethylformamide.
3. Biological results13 and discussion
We have recently reported that the thiadiazole amide
sub-series, represented by the general structure 23
(Fig. 2), show nanomolar inhibitory activities.10 How-
ever, in rats, low bioavailabilities (<20%) and high
clearances (>100mL/min/kg) were observed for some
of these derivatives. Additional data (not shown) sug-
gested that the marginal bioavailabilities were not re-
lated to permeability, physicochemical properties (such
as solubility or dissolution) or to transporters issues
therefore suggesting a predominant role for a clearance
route. Most of the in vivo clearances were much higher
than the clearances predicted from in vitro rat micro-
somal stability studies (<46.3mL/min/kg). Therefore,
the high in vivo total clearance was suspected to be a
combination of an extra-hepatic and hepatic metabo-
lism. Consequently, we decided to identify the major
in vivo metabolites in order to remove or alter the func-
tionalities associated with the rapid rate of metabolism.
The identification of the major metabolites after po and
iv rat administration of 1 revealed two principal sites
of metabolism: oxidation of the cyclohexyl ring to
form 9 and hydrolysis of the amide function to give 7
(Fig. 2).

F. Vergne et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4615–4621
4617
Figure 2. Schematic representation of the principle sites of metabolism.
Hydroxylation occurs mainly at the 3-position of the
cyclohexyl ring to form the cis-stereoisomer with a
marginal formation of 4-hydroxylated derivatives
(stereochemistry not confirmed for the latter). This
result was anticipated because this region of the mole-
cule was highly hydrophobic and could be a good sub-
strate for cytochrome P-450Õs, which have lipophilic
binding sites. The extra-hepatic metabolic pathway
was the hydrolysis of the amide function in rat whole
blood, which gave the carboxylic acid derivative 7,
known to be a PDE7 inhibitor (IC50 =550nM). It is
worth noting that reduction of the global lipophilicity
(decrease of logD by amino side chains as R groups
(23), Fig. 2) resulted in an excellent in vitro metabolic
stability (data not shown) but was without any effect
on the in vivo total clearance.
potential metabolites of 1) to assess their properties in
terms of inhibitory activities and pharmacokinetics; (c)
find stable amide bioisosteres without loss of inhibitory
activity. These metabolism-directed optimization studies
should lead to appropriate modifications of the meta-
bolically labile structural parts to generate compounds
with improved pharmacokinetics.
We first assessed the pharmacokinetic profile of some
benzoic acid derivatives illustrated by representative
examples 10–12 of this chemical sub-series (Table 1).
Upon oral dosing of these compounds to rats, com-
pound 10 exhibited the highest bioavailability
(F=49%) with a 2h plasma half-life. As illustrated by
11 and 12, the nature and the relative position of R2
did not appear to affect the enzymatic activity, suggest-
ing no or marginal favourable interactions with residues
of the protein. The major contribution in terms of inhib-
itory activity probably resulted from interactions be-
tween the benzoic acid structural part and the enzyme.
Then, in an attempt to increase the half-life of 10 the
R2 groups were modified (with R5 =H) to reduce the
plasma protein binding (PPB). Indeed, it is known that
binding of acidic compounds to albumin is dependent
on lipophilicity, hence lipophilic acids generally will
tend to bind extensively and consequently will exhibit
These results reflected that the extra-hepatic route was
the primary contributor to the higher rat in vivo
clearance.
To improve these pharmacokinetic parameters, the com-
bined strategies adopted were to: (a) study and optimize
the pharmacokinetics of the previously identified meta-
benzoic acid sub-series (preceding publication)10 which
do not have these structural metabolism sites; (b) pre-
pare hydroxylated cyclohexyl derivatives (which are

4618
F. Vergne et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4615–4621
Table 1. Rat pharmacokinetic parameters for compounds 1 and 10–12
Compd
PDE7A1
IC50, lMa
PDE4D3
IC50, lMa
Cl
(mL/min/kg)b
Vd
(L/kg)b
t1/2
(h)c
F
(%)c
Solubility
pH7.4
(lg/mL)d
1
0.061
15.30
179.5
7.65
BLQe
BLQ
4
R5
R2
10
11
12
4-F
H
4-Cl
4-SO2Me
3-CN
0.16
0.12
0.19
39.66
65.5
35.67
3.9
15.8
2.65
0.187
7.58
0.994
1.96
2.7
49.3
29.3
37.0
145
370
295
H
4.75
a Measured against the human full length enzyme produced in baculovirus infected sf9 cells. Values are means of three experiments.
b Dose 0.5mg/kg iv.
c Dose 2.5mg/kg po.
d Determined by stirring in pH7.4 Na2HPO4/NaH2PO4 buffer for 24h.
e BLQ=Below limit of quantification.
volumes of distribution close to plasma or blood volume
(0.1–0.2L/kg).14 The highly hydrophobic compound 10
is a perfect representative example of this effect. As,
the half-life is a function of the clearance and volume
of distribution (according to the following equation:
half-life=0.693·Vd/Cl), half-life of these related com-
pounds can be improved by increasing the volume of
distribution or reducing the plasma clearance.15 Since
10 displayed low clearance, we decided to increase the
Vd by reducing the plasma protein binding of related
analogues. The approach was to decrease the global
lipophilicity by introduction of polar groups (R2) on
the phenyl part of the molecule. These modifications
illustrated by compounds 11 and 12 resulted indeed in
a decrease of plasma protein binding (respectively 7%
and 2% less than the PPB for 10). Consequently 11
and 12 exhibited a 40-fold and 5.3-fold increase respec-
tively in the volume of distribution (Vd) compared to 10
and accordingly improved the corresponding half-lives.
Only a slight decrease of bioavailability was observed
compared to 10. In conclusion from these first studies,
the design of new benzoic acid derivatives allowed a sig-
nificant improvement of the pharmacokinetic profiles
(t1/2 > 2h, F > 29%) with good solubilities (>145lg/mL
at pH7.4) compared to the reference compound 1.
the racemic forms of the hydroxylated derivatives re-
sulted in a only moderate decrease in inhibitory activity
(3-fold) as exemplified by 15 and 16 compared to 1.
While a relatively more significant loss in activity was
observed with compounds 13, 14 and 9 versus 1. It is
worth noting that the position of the OH groups on
the cyclohexyl ring and the stereochemistry modulate
the level of inhibitory activity as exemplified by compar-
ison of 13–14 (2-position) with 15 (4-position) or 16 (3-
position) and by comparison of 9 (cis-isomer) with 16
(trans-isomer). The trans-isomer 16 displayed 190nM
IC50 while the cis analog 9 (major in vivo metabolite)
is around 9-fold less potent.
All the derivatives exhibited a high selectivity against the
PDE4 enzyme. In conclusion from these preliminary
studies, two new potent PDE7 inhibitors 15 and 16
(IC50 =190nM) have been identified.
The in vivo rat pharmacokinetic profiles were deter-
mined for the selected racemic compounds 9 and 14–
16 (Table 2). All the hydroxylated derivatives tested
led to a significant decrease in the total clearances (range
of reduction between 62–81% compared to 1).
We next examined another approach (c) directed to-
wards the assessment of hydroxylated cyclohexyl deriv-
atives in terms of inhibitory activities and
pharmacokinetics. The objective was to identify within
these new derivatives, potent PDE7 inhibitors with a re-
duced rate of metabolism at the cyclohexyl site. Table 2
summarizes a broad survey around the effect of cyclo-
hexyl hydroxylation on PDE7 inhibitory activity,
selectivity versus PDE4 and pharmacokinetics. Interest-
ingly, some of the preliminary modifications based on
The optimal compounds combining good half-lives and
bioavailabilities were the 3-hydroxy derivatives 9 and 16
with F=61.2% and 38.5%, respectively (t1/2 =1.42 and
1.80h, respectively). These studies clearly demonstrated
that the rat in vivo clearance could be considerably re-
duced by introduction of an OH group at the 3-position.
In addition, compound 16 was identified as a tool for in
vivo experiments. Considering the real benefit of the 3-
hydroxylation (16) in terms of pharmacokinetics and
considering the importance of the stereochemistry (trans

F. Vergne et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4615–4621
4619
Table 2. In vitro data and in vivo rat pharmacokinetic parameters for compounds 1, 9 and 13–16(a,b)
Compd R6
Stereochemistry PDE7A1
PDE4D3
Cl (mL/min/kg)b Reduction in
clearance versus 1 (%)
Vd (L/kg)b t1/2 (h)c F (%)c
IC50, lMa IC50, lMa
1
H
––
2-OH cis
0.061
0.61
0.55
0.19
1.65
0.19
0.088
1.74
15.30
28.50
>101
46.33
>101
71.64
61.62
>101
179.5
ND
38.2
66.9
34.4
59.9
ND
ND
0
7.65
ND
1.47
0.98
1.39
1.78
ND
ND
BLQ
ND
0.31
0.27
1.42
1.8
BLQ
13
14
15
9
ND
79
ND
24.7
7
2-OH trans
4-OH trans
3-OH cis
62
81
61.2
38.5
ND
ND
16
16a
16b
3-OH trans
3-OH (R,R)
3-OH (S,S)
67
ND
ND
ND
ND
a Measured against the human full length enzyme produced in baculovirus infected sf9 cells. Values are means of three experiments.
b Dose 0.5mg/kg iv.
c Dose 2.5mg/kg po.
vs cis) in terms of potency, we decided to assess the
inhibitory activity of both enantiomers of 16. These
compounds 16a and 16b were evaluated and the results
are indicated in Table 2. This study allowed to identify
the potent enantiomer 16a (eutomer) with an IC50 of
88nM (2-fold increase compared to 16).
tory activities and the pharmacokinetics of some
selected analogs from this study.
Simple modifications of the amide function without
changes at the cyclohexyl site, exemplified by 17 (reverse
amide) and 18 (increased steric hindrance) showed better
metabolic stability compared to 1 (clearance reduced by
58 % and 75 %, respectively). Interestingly, 17 displayed
the same level of inhibitory activity compared to 1.
These new results led us to direct our efforts towards the
discovery of hybrid structures combining stable amide
bioisosteric groups with the novel (R)-hydroxy cyclo-
hexyl part, to reduce more significantly the clearances
and accordingly to increase again the half-lives and the
bioavailabilities. Table 3 lists the structures, the inhibi-
As illustrated by 18, the high oral bioavailability
(F=85%) is likely to be the result of a lower clearance
and a better solubility compared to compounds 1, 17
Table 3. In vitro data and in vivo rat pharmacokinetic parameters for compounds 1 and 16, 17–21(a)
Compd R6
Stereo-
chemistry
R2
PDE7A1
PDE4D3
Cl
Reduction
Vd
t1/2
F
Solubility
IC50, lMa IC50, lMa (mL/min/kg)b in clearance (L/kg)b (h)c
versus 1 (%)
(%)c pH 7.4
(lg/mL)d
1
H
H
––
––
4-CONH2
4-NHCOMe
4-CONH2
0.061
0.068
0.19
15.30
34.25
71.64
>101
179.5
74.7
59.9
44.8
0
58
67
75
7.65
3.3
BLQ BLQ
4
8
17
16
18
0.74
1.8
19.6
38.5
85
3-OH trans
––
1.78
2.59
40
350
H
4-CO(4-methyl- 0.42
piperazin-1-yl)
1.37
19
3-OH trans
4-CONH(2-
hydroxy-1,1-
dimethylethyl)
4-NHCOMe
4-SO2Me
0.38
>101
24.9
86
0.43
0.46
63
58
20
21
3-OH trans
3-OH trans
3-OH (R,R)
0.085
0.31
0.052
40.80
79.17
20.0
13.3
12.2
15.1
92
93
91
2.02
0.92
0.62
2.3
2
76
350
135
160
100
100
21a
4-SO2Me
2.5
a Measured against the human full length enzyme produced in baculovirus infected sf9 cells. Values are means of three experiments.
b Dose 0.5mg/kg iv.
c Dose 2.5mg/kg po.
d Determined by stirring in pH7.4 Na2HPO4/NaH2PO4 buffer for 24h.

4620
F. Vergne et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4615–4621
and also 16. This result reflects the benefit of amide
modifications to optimize the pharmacokinetic profiles
of thiadiazole derivatives. However, as mentioned in
the preceding publication,10 the loss of HBD properties
(18) caused a decrease in inhibitory activity. Compound
19 was then designed to assess the importance of the
combination of a novel hindered amide function (with
HBD property) with the hydroxy cyclohexyl ring on
the rate of metabolism. As expected, this modification
resulted in a more pronounced reduction in clearance
(86%) compared to 16 (67%), 17 (58%) and 18 (75%).
This result confirmed the choice of a combined ap-
proach. However, the moderate volume of distribution
for 19 resulted in a short half-life (t1/2 =0.46h). Interest-
ingly, the equivalent inhibitory activity displayed by 19
compared to 18, suggested that steric bulk at the prox-
imity of the amide H atom, is also detrimental to the
HBD property and consequently to potency. It is worth
noting that compounds 20 and 21 (21a) offered the
appropriate structural combination resulting in a signif-
icant improvement of pharmacokinetic profiles com-
pared to the starting point 1. In addition, replacing the
racemic form 21 by the enantiomer 21a proved to be
favourable for inhibitory activity (6-fold increase to
reach 52nM).
and the Evotec OAI scientists for their assistance in this
project. We also thank Angus Nedderman and Michael
Ritzau (PGRD, Sandwich) for their assistance in this
project.
References and notes
1. (a) Francis, S. H.; Turko, I. V.; Corbin, J. D. Prog. Nucleic
Acid Res. Mol. Biol. 2001, 65, 1–52; (b) Conti, M. In
Principles of Molecular Regulation; Conn, P. M., Means,
A. R., Eds.; Humana, 2000; pp 261–275; (c) Soderling,
S. H., J. A.; Beavo, J. A. Current Opinion in Cell Biology
2000, 12, 174–179.
2. (a) Conti, M.; Jin, S. L. C. Prog. Nucleic Acid Res. Mol.
Biol. 1999, 63, 1–38; (b) Essayan, D. M. J. Allergy Clin.
Immunol. 2001, 108, 671–680.
3. (a) Li, L.; Yee, C.; Beavo, J. A. Science 1999, 283,
848–851; (b) Nakata, A.; Ogawa, T.; Sasaki, T.; Koyama,
N.; Wada, K.; Kotera, J.; Kikkawa, H.; Omori, K.;
Kaminuma, O. Clin. Exp. Immunol. 2002, 128, 460–466.
4. Smith, S. J.; Brookes-Fazakerley, S.; Donnelly, L. E.;
Barnes, P. J.; Barnette, M. S.; Giembycz, M. A. Am. J.
Physiol. Lung Cell. Mol. Physiol. 2003, 284, L279–L289.
5. (a) Perez-Torres, S.; Cortes, R.; Tolnay, M.; Probst, A.;
Palacios, J. M.; Mengod, G. Exp. Neurol. 2003, 182,
322–334; (b) Miro, X.; Perez-Torres, S.; Palacios, J. M.;
Puigdomenech, P.; Mengod, G. Synapse 2001, 40,
201–214; (c) Sasaki, T.; Kotera, J.; Omori, K. Biochem.
J. 2002, 361, 211–220.
Compound 18 exhibited the higher Vd in contrast to
some of the neutral compounds 16, 19–21 and (21a).
This is probably due to the basic property of the 4-meth-
ylpiperazine moiety, suggesting that the half-lives could
be again optimized by incorporation of basic functions
into the molecules. As anticipated, the Vd of neutral
hydroxy compounds such as 16, and 20 are lower com-
pared, respectively, to 1 and 17 most probably reflecting
a reduced lipophilicity (presence of polar functions)
compared to compounds 1 and 17. As expected, this
effect is even more pronounced with 19 or 21 (21a).
These results are in agreement with the overall trend
for neutral compounds to display increased Vd with
higher logD,14 with the risk also to increase the suscep-
tibility to hepatic clearance as observed with 1 and 17.
The compounds 20–21 and 21a displayed very high bio-
availabilities (76–100%) with acceptable half-lives
(t1/2 > 2h) resulting not only from optimized low clear-
ances but also due to the enhanced solubilities
(>135lg/mL) and to high permeabilities confirmed by
measures across Caco-2 cells (Papp >30·10À6 cm/s).
6. Lee, R.; Wolda, S.; Moon, E.; Esselstyn, J.; Hertel, C.;
Lerner, A. Cell. Signalling 2002, 14, 277–284.
7. (a) Hetman, J. M.; Soderling, S. H.; Glavas, N. A.; Beavo,
J. A. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 472–476; (b)
Wang, P.; Wu, P.; Egan, R. W.; Billah, M. M. Biochem.
Biophys. Res. Commun. 2000, 276, 1271–1277; (c) Sasaki,
T.; Kotera, J.; Yuasa, K.; Omori, K. Biochem. Biophys.
Res. Commun. 2000, 271, 575–583.
8. (a) Kotera, J. PDE7-news and views, William Harvey
Research Conferences, Porto, Portugal, December 5–7,
2001; (b) Kluxen, F. W. PCT Int. Appl. WO 0183772 A1,
2001.
9. (a) Barnes, M. J.; Cooper, N.; Davenport, R. J.; Dyke, H.
J.; Galleway, F. P.; Galvin, F. C. A.; Gowers, L.;
Haughan, A. F.; Lowe, C.; Meissner, J. W. G.; Montana,
J. G.; Morgan, T.; Picken, L. C.; Watson, R. Bioorg. Med.
Chem. Lett. 2001, 11, 1081–1083; (b) Castro, A.; Abasolo,
M. I.; Gil, C.; Segarra, V.; Martinez, A. Eur. J. Med.
Chem. 2001, 36, 333–338; (c) Martinez, A.; Castro, A.; Gil,
C.; Miralpeix, M.; Segarra, V.; Domenech, T.; Beleta, J.;
Palacios, J. M.; Ryder, H.; Miro, X.; Bonet, C.; Casacu-
berta, J. M.; Azorin, F.; Pina, B.; Puigdomenec, P. J. Med.
Chem. 2000, 43, 683–689; (d) Pittz, W. J.; Vaccaro, W.;
Huynh, T.; Leftheris, K.; Roberge, J. Y.; Barbosa, J.;
Guo, J.; Brown, B.; Watson, A.; Donaldson, K.; Starling,
G. C.; Kiener, P. A.; Poss, M. A.; Dodd, J. H.; Barrish,
J. C. Bioorg. Med. Chem. Lett. 2004, 14, 2955–2958.
10. Vergne, F.; Bernardelli, P.; Lorthiois, E.; Pham, N.;
Proust, E.; Oliveira, C.; Mafroud, A.-K.; Royer F.;
Wrigglesworth, R.; Schellhass, J. K.; Barvian, M. R.;
Moreau, F.; Idrissi, M.; Tertre, A.; Bertin, B.; Coupe, M.;
Berna, P.; Soulard, P. Bioorg. Med. Chem. Lett. 2004, 14,
preceding communication. doi:10.1016/j.bmcl.2004.07.008.
11. Vergne, F.; Ducrot, P.; Andrianjara, C.; Bernardelli, P.;
Lorthiois, E. PCT Int. Appl. WO 0228847 A1, 2002.
12. (a) Molina, P.; Tarraga, A.; Espinosa, A. Synthesis 1988,
690–693; (b) Molina, P.; Tarraga, A.; Espinosa, A.
Heterocycles 1989, 29, 2301–2316.
4. Conclusion
In summary, we have rationally designed a structurally
novel series of potent PDE7 inhibitors displaying good
oral pharmacokinetics in rat. Compounds 20 and 21a
perfectly illustrated a successful metabolism-directed
optimization approach that allowed to design deriva-
tives with improved in vivo pharmacokinetics.
Acknowledgements
We thank the scientists from the Analytical Support
Group for their help in compound characterization

F. Vergne et al. / Bioorg. Med. Chem. Lett. 14 (2004) 4615–4621
4621
13. All the compounds illustrated in this report were not
selective for either enzyme subtype PDE7Aor PDE7B.
Consequently, only PDE7Ainhibitory activities are
presented.
14. Van de Waterbeemd, H.; Smith, D. A.; Beaumont, K.;
Walker, D. K. J. Med. Chem. 2001, 44, 1–21.
15. Van de Waterbeemd, H.; Gifford, E. Nat. Rev. Drug Disc.
2003, 2, 192–204.