Disease activated drugs: A new concept for the treatment of asthma

Article abstract of DOI:10.1016/S0968-0896(01)00077-3

Disease activated drugs (DAD) and pro-drugs of one active principle or combinations of two drugs, which have a proven efficacy for the treatment of the target disease. In opposition to pro-drugs, DAD are activated in inflamed but not normal tissues. Due to the disease specific activation, the amount of locally released drug(s) should be related directly to the severity of the inflammation. To test this concept in asthma a PDE4 inhibitor, an isoquinoline derivative, was chemically derivatized into pro-drugs or combined with corticosteroids. These new compounds were more readily cleaved into active PDE4 inhibitor, in bronchoalveolar lavage fluid (BALF) from Brown-Norway rats with lung inflammation than in BALF from rats without airway inflammation. The DAD concept (local selective release and improved therapeutic window) was validated in vivo using the inhibition of methachline induced bronchoconstriction in guinea pigs with or without ozone induced lung inflammation. An example of DAD hydrolysis (isoquinoline-dexamethasone) was also examined in BALF from asthmatics and healthy volunteers Copyright

Full text of DOI:10.1016/S0968-0896(01)00077-3

Bioorganic & Medicinal Chemistry 9 (2001) 1793–1805
Disease Activated Drugs: A New Concept for the
Treatment of Asthma
Brigitte Charpiot,a,* Francis Bitsch,a Karl-Heinz Buchheit,a Pascal Channez,b
a
Lazzaro Mazzoni,a Thomas Mueller,a Isabelle Vachierb andReto Naef
aResearch, Novartis Pharma AG CH-4002 Basle, Switzerland
bCHU Montpellier, Hoˆpital Arnaud de Villeneuve, 34295 Montpellier, France
Received12 September 2000; accepted14 March 2001
Abstract—Disease activated drugs (DAD) are pro-drugs of one active principle or combinations of two drugs, which have a proven
efficacy for the treatment of the target disease. In opposition to pro-drugs, DAD are activated in inflamed but not normal tissues.
Due to the disease specific activation, the amount of locally released drug(s) should be related directly to the severity of the
inflammation. To test this concept in asthma a PDE4 inhibitor, an isoquinoline derivative, was chemically derivatized into pro-
drugs or combined with corticosteroids. These new compounds were more readily cleaved into active PDE4 inhibitor, in bronch-
oalveolar lavage fluid(BALF) from Brown-Norway rats with lung inflammation than in BALF from rats without airway inflam-
mation. The DAD concept (local selective release and improved therapeutic window) was validated in vivo using the inhibition of
methacholine induced bronchoconstriction in guinea pigs with or without ozone induced lung inflammation. An example of DAD
hydrolysis (isoquinoline-dexamethasone) was also examined in BALF from asthmatics and healthy volunteers. # 2001 Elsevier
Science Ltd. All rights reserved.
Introduction
inflammatory conditions (e.g., asthma) would be the
ideal target to test the feasibility of such a new concept.
The use of phosphodiesterase type 4 (PDE4) inhibitors
andtheir combination with corticosteroisd appears
therefore as an attractive example for the evaluation of
the DAD concept.
The concept of disease activated drugs (DAD) is new,
andtheoretically, couldbe appliedto many chronic
inflammatory diseases. DAD are derivatized drugs or
combinations of two covalently linkedrdugs. These
derivatized drugs or drug combinations, per se inactive,
are to be activatedlocally by a disease specific mechan-
ism (e.g., proteolytic enzymes, free radical mediators,
alteration in local pH or redox potential) specifically in
the inflamedtissues (Fig. 1). This approach could
potentially present several advantages: reduced side
effects (since active drugs are released only in the
inflamedtissues) andthus improve therapeutic window,
possibility of modulating the pharmacokinetics of the
single drugs, potential additive or even synergistic
effects by combination of compounds with different
mechanism of action.
Cyclic 30,50-AMP (c-AMP) exerts its action in the cell by
binding to and activating protein kinase A. This family
of enzymes elicits the phosphorylation of specific target
proteins, leading to functional changes within the cell.
Changes in the c-AMP signalling system have been
notedin a number of idseases, such as asthma,
DAD might be appliedin various chronic idseases
with episodic progression but it seems that chronic
*Corresponding author. Tel.: +41-61-324-6208; fax: +41-61-324-
7875; e-mail: brigitte.charpiot@pharma.novatis.com
Figure 1. DAD concept.
0968-0896/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(01)00077-3

1794
B. Charpiot et al. / Bioorg. Med. Chem. 9 (2001) 1793–1805
inflammation andatopy. The only known way cells
inactivate c-AMP is by hydrolysing it to 50-AMP. This
is achievedthrough the action of a family of enzymes
calledcyclic AMP phosphoidesterases (EC 3.1.4.17).
PDE4 enzymes, a subfamily of the phosphodiesterases,
are able to hydrolyse specifically c-AMP and have
receivedconsiderable attention as molecular targets for
novel anti-asthmatic agents.1aꢀc PDE4 inhibitors have
been shown to impair the influx andactivation of neu-
trophils, eosinophils, lymphocytes andother cells of the
immune system.2 In the treatment of asthma, the appli-
cation by inhalation of a DAD containing a PDE4
inhibitor could further reduce the side-effects due to the
ubiquitousness of PDE activity, a common problem of
this inhibitor class.3 In addition the therapeutic asso-
ciation of a PDE4 inhibitor anda corticosteroid(dex-
amethasone, budesonide) in a DAD could possibly
potentiate the effect of the single components. For proof
of concept, compound 1, ethanol, 2-[1-(3,5-bis(1-methyl-
ethoxy)-phenyl)-3-ethyl-7-methoxy-6-isoquinolinyl] oxy],
a selective, potent inhibitor of PDE4 enzyme with high
anti-inflammatory activity,4 was selectedas parent
component. Derivatives of 1 andcombinations with
corticosteroids (dexamethasone and budesonide) were
synthesised as model molecules for the validation of the
DAD concept. Dexamethasone and budesonide were
chosen as corticosteroids since they were tested as equi-
potent with compound 1 in our in vivo model of airway
inflammation.
Chemistry
The general route for synthesising the acyl derivatives of
the PDE4 inhibitor, 1, is illustratedin Scheme 1. Cou-
pling via the free hydroxyl function of 1 with the
appropriate acidchloried occurredin presence of a
base, usually dried triethylamine and gave compounds
2–5 in moderate to good yields.
The carbonate derivative 6 was obtainedby reacting
ethyl chloroformate and 1. Formation of the mono
oxalic salts facilitatedthe purification of some
compounds or improved their physicochemical properties.
The use of 1,10-carbonyldiimidazole (respectively 1,10-
thiocarbonyldiimidazole for the synthesis of 9 and 11)
turnedout to be necessary for the synthesis of the
homodimer of 1, 8, andthe combination products of
1
with dexamethasone, 10, and 1 with budesonide, 12,
(Scheme 2). Other coupling reagents, for example phos-
gene, ethyl chloroformate, were unsuccessful. The
intermediate 7 is a stable compoundandcan be isolated.
The acylation of dexamethasone was only possible in
position 21. Although several dexamethasone-17 alkyl
carbonate derivatives are described in the literature,5
none of these methods (enzymatic alkoxycarbonylation
from vinyl carbonate or methyl carbonate catalyzedby
Candida antartica lipase,6aꢀc synthesis of dexame-
thasone-17,21-dialkylorthocarbonate and successive
The design of the linker was subjected to the following
considerations: the linker should not generate a toxic
entity during the hydrolysis process, it should not
increase significantly the synthesis cost of the final
molecule andfinally shouldallow an almost simulta-
neous release of both active compounds. Thus, we
focusedon a carbonate linker (for PDE4 inhibitor–
corticosteroidhybridmolecules) as carbon idoxied
release is of no consequence for the lungs.
In the first step, these molecules were usedto test the
feasibility of the concept in vivo andin vitro. A next
step, as part of a mechanistic approach, wouldaim at
the determination of the enzyme(s) involved in the clea-
vage process.
Scheme 2. Reagents and conditions used: (a) 1 equiv 1,10-carbodiimi-
dazole (1,10-thiocarbonyldiimidazole, respectively), 4-dimethylamino
pyridine (4 equiv); (b) 1 equiv dexamethasone (budesonide or 1).
Scheme 1. Reagents and conditions used: (a) 1.2 equiv RCOCl, 1.2
equiv NEt3;(b) oxalic acidin ether (except for 5).

B. Charpiot et al. / Bioorg. Med. Chem. 9 (2001) 1793–1805
1795
rearrangement,5 silylation of the 21-hydroxyl and direct
acylation of the 17-position or reaction of the 17-
hydroxyl with alkyl halide and silver carbonate7) ledto
the 17-acylated desired product. The budesonide analo-
gue, 12, was synthesisedfollowing the preceding method.
andremain uncleavedin normal animals. The in vitro
assay which seems to be adequate for this purpose is the
hydrolysis study of the different compounds in bronch-
oalveolar lavage fluids (BALF) of BN rats sensitised
andchallengedwith ovalbumin (BALF+) comparedto
the BALF of sensitised, but non-challenged BN rats
(BALFꢀ). Compound 13,11 (Fig. 2), which has a che-
mical structure similar to 1, was usedas an internal
reference for the relative quantification of compounds in
the BALF samples.
Results and Discussion
DAD, as previously mentioned, should be per se inac-
tive molecules. In our context, they shouldnot inhibit
PDE4 enzyme andalso, to reduce the potential for car-
dio-vascular side-effects, should be inactive as PDE3
inhibitors.
Firstly, we investigatedwhether DAD 2–5 (ester deri-
vatives of 1) could be hydrolysed, that is activated, in
vitro in the two different BALF.
Data obtainedwith the esters was promising (Fig. 3).
After 1 h, hydrolysis of 2 yielded 38% of 1 in BALF+
and6% in BALF ꢀ. The same pattern of hydrolysis was
also observedwith other esters. A fine tuning of the
hydrolysis rate in BALF, while the selectivity is still
conserved, was possible by the introduction of sterically
hindered residues on the ester side (compound 2
comparedto 3–5).
Phosphodiesterase enzymatic assays
The compounds synthesised are far less potent inhibi-
tors of PDE4 than 1 andare mostly inactive against
PDE3 (Table 1). The compounds were tested against
both PDE4 enzymes, isolatedfrom human neutrophil
homogenates, andclonedrat PDE4B enzyme. Since
human PDE4B, a subtype of the PDE4 family, seems to
be the dominant PDE4 isogene in neutrophils8 andis
92% homologous to rat PDE4B,9 results from both
preparations should be comparable. Indeed compounds
1, 6 and 9, which were testedfor PDE4 andPDE4B
inhibition, showedsimilar activities in both assays.
The carbonate, 6, exhibiteda similar pattern of hydro-
lysis as compound 2 (Fig. 3). Carbamate derivatives of 1
(tert-butyl, benzyl) were stable during the 24 h incubation
in BALF+ andBALF ꢀ (data not shown). We then
lookedat DAD 8–12, which are inactive hybridmole-
cules of two per se active compounds. In general, a
dimer of 1, 8, and the hybrids dexamethasone-1, 10, and
budesonide-1, 12, were cleavedmore slowly than the
esters 2–4, nevertheless 1 was still formedmore rapidly
in BALF+ than in BALFꢀ (Fig. 4). Remarkably, com-
pound 8, which allows free access to the hydrolizable
carbonyl function was cleavedless efficiently than com-
pound 10, which is more sterically hindered. This find-
ing indicates that the hydrolysis rate is not entirely
driven by steric hindrance but probably also by enzyme
substrate recognition. Surprisingly, the thio-analogues
of 8, 9 (data not shown), and of 10, 11, remained
uncleavedafter 24 h incubation in both BALF. This
result reinforces the hypothesis that an enzymatic reac-
tion is involvedin the hydrolysis.
Hydrolysis studies in rat bronchoalveolar lavage fluids
(BALF)
Efflux of eosinophils into the airway lumen, following
inhalation of the antigen, is a distinctive feature of
asthma. Analogous to the effects in asthmatics, an
accumulation of inflammatory cells (neutrophils, eosi-
nophils andmononuclear cells) anda release of media-
tors in the airways occur when actively sensitisedBrown
Norway (BN) rats inhale modest amounts of allergen.
Such animal challenge models can be used to suggest
potential clinical efficacy of anti-asthma drugs.10 Based
on this, we aimedat developing an in vitro assay to test
our working hypothesis that DAD, administered by
inhalation, are cleavedin the lung of allergic animals
Table 1. Inhibition of phosphodiesterase isoenzymesa
Compounds
PDE3
PDE4
(human neutrophils)
PDE4B
(rat)
1
2
3
4
5
6
0.48ꢁ0.15 (3) 0.0036ꢁ0.0005 (10) 0.0042ꢁ0.0001 (130)
b
>100
3.71
2.3
—
—
—
—
—
—
—
60.2
>100
1.7
>100
2.04
8
9
11
12
—
0.076 (2)
>10
0.14 (2)
0.17 (2)
—
—
0.29 (2)
0.26 (2)
0.7 (2)
—
—
aData represent mean IC50 values mꢁSEM (mM) with the number of
observations shown in brackets.
bMeans not determined.
Figure 2. Structure of compound 13.

1796
B. Charpiot et al. / Bioorg. Med. Chem. 9 (2001) 1793–1805
Figure 3. Hydrolysis of DAD esters in BALF from sensitized-challenged Brown Norway rats (BALF+) and BALF from sensitized-non-challenged
Brown Norway rats (BALFꢀ). Shown are the means of two aliquots.

B. Charpiot et al. / Bioorg. Med. Chem. 9 (2001) 1793–1805
1797
Figure 4. Hydrolysis of DAD carbonates in BALF from sensitized-challenged Brown Norway rats (BALF+) and BALF from sensitized-non-
challengedBrown Norway rats (BALF ꢀ). Shown are the means of two aliquots.

1798
B. Charpiot et al. / Bioorg. Med. Chem. 9 (2001) 1793–1805
Bronchoconstriction in ozone- and non-ozone treated
guinea pigs
This anti-bronchoconstrictor effect can fully be ascribed
to the release of 1, if the DAD molecule does not inhibit
bronchoconstriction in non-ozone-treatedguinea pigs,
where no inflammation process is present. However, this
fully depends on the assumption that for 1, the released
active principle, an anti-bronchoconstrictor effect can
be observed in the chosen animal model. Our data
demonstrate a dose-dependent anti-bronchoconstrictor
effect of 1, which can be measuredboth in ozone- and
non-ozone-treatedguinea pigs. In ozone-pretreatedani-
mals, 1 inhibitedmethacholine-inudcedbronchocon-
striction by 73% (maximal effect, obtained30 min after
administration) at a dose of 30 mg/kg given intra-
tracheally. In guinea pigs, which were not treatedwith
ozone, the maximal effect of 1 was identical (78% inhi-
bition, 30 min after administration) (Fig. 5).
In a wide variety of species, including man,12 rats 13 and
guinea pigs,14 inhalation of ozone induces an inflam-
matory process in the lung, accompaniedby increased
contractile reaction of the airways to exogenous and
endogenous spasmogens such as methacholine. As a
result, the bronchoconstrictor effect to a given dose of,
for example methacholine, is markedly enhanced in
ozone- as comparedto non-ozone treatedguinea pigs.
Since PDE4 inhibitors, like 1, have the potential to
inhibit bronchoconstriction,15aꢀd the assessment of the
anti-bronchoconstrictor effect of DAD in ozone- and
non-ozone-treatedguinea pigs offers the possibility of
(1) determining the degree of hydrolysis of these com-
pounds in vivo and (2) proving the concept of DAD in
animals. If a DAD is cleavedto the active principle 1,
by enzymes releasedin the ozone-inudcedairway
inflammatory process, an inhibition of the bronchocon-
striction provokedby methacholine is to be expected.
In contrast to 1, compounds 2 and 6, which were readily
hydrolyzed in BN rat BALF+, also inhibited metha-
choline-induced bronchoconstriction in ozone-treated
guinea pigs but not in guinea pigs which were not trea-
tedwith ozone (Fig. 5). The doses that were necessary
Figure 5. Time course of the effect of compounds 1, 2, 6, 8 andtheir vehicle ( *) on methacholine induced bronchoconstriction in ozone-pretreated
(&) andnon-ozone-pretreated( &) guinea pigs. Shown is the inhibition of the bronchoconstrictor response in % of the response to methacholine in
the absence of test compounds as meanꢁSEM in three to five animals. The compounds were given it at timepoint 0. Significant differences between
effects of the vehicle and the test compounds are indicated: *p<0.05; **p< 0.01; ***p< 0.001 (Student’s t-test).

B. Charpiot et al. / Bioorg. Med. Chem. 9 (2001) 1793–1805
1799
for a comparable degree of inhibition were marginally
Therapeutic window
higher than those usedin the case of 1. The lack of
inhibition obtainedin healthy guinea pigs with these
two compounds clearly indicates their inability to relax
airway smooth muscle when they are not cleavedinto 1.
This observation is consistent with their low inhibitory
activity against PDE3 (IC50>100 mM) or PDE4 (Table 1)
as bronchodilatation is often associated with inhibition
of PDE3 or/and PDE4 activity. In addition it shows
that the molecules were not hydrolysed in non-ozone-
treatedguinea pigs. As a consequence, it can be con-
cluded that the inhibitory effect obtained in ozone-trea-
tedanimals is due to the release of 1 in the lung. Even
though higher doses are required as for 1 itself, the dif-
ference might be explainedby the kinetics of the enzy-
matic hydrolysis and the pharmacokinetics of 1 and
DAD respectively in guinea pigs. For example, if the
rate of hydrolysis of compounds 2 and 6 to 1 is slow
and—at the same time—their elimination is fast, higher
doses of compounds 2 and 6 are necessary to achieve the
same degree of inhibition as observed with 1. In con-
trast to compounds 2 and 6, compound 8, which is
slowly cleavedin BN rat BALF+(less than 10% after
24 h), is almost inactive in ozone-treatedguinea pigs as
an anti-bronchoconstrictor agent (e.g., 30% inhibition
at a dose of 150 mg/kg it).
In order to determine the therapeutic window, the effect
of compound 1 andthe DAD on bloodpressure was
comparedto the anti-bronchoconstrictor effect in gui-
nea pigs in vivo (Table 2). At 30 mg/kg it, compound 1
inhibitedbronchoconstriction in ozone- andnon-ozone-
treatedguinea pigs andincreasedbloodpressure by
about 20%. In contrast, compounds 2 and 6, both at
100 mg/kg it causeda similar degree of inhibition of
bronchoconstriction in ozone-treatedguinea pigs but
were devoid of effects on blood pressure, both in ozone-
andnon-ozone-treatedanimals (Table 2). This indicates
an improvement of the therapeutic window for both
DAD as comparedto the locally administeredparent
compound 1.
Hydrolysis studies in human BALF supernatants
The results obtainedwith some DAD in vitro andin
vivo encouragedus to focus on compound 10 andcarry
on our studies using human material. Consequently we
usedhuman BALF supernatants from asthmatic
patients andfrom healthy volunteers.
The hydrolysis rate and quantity of 1 formedafter
incubation of 10 in BALF supernatants of seven asth-
matic patients was comparedwith those obtainedin
BALF supernatants of 8 healthy volunteers. Age, sex,
clinical severity of asthma andpulmonary function
assessedby forcedexpiratory volume in one second,
FEV1, characterizedthe patients. Moreover, BALF cell
content was determined on cytospin slides (Table 3).
The number of cells was similar in the asthmatic andin
the control groups, 126ꢁ20 and133 ꢁ16, respectively
(meanꢁSEMꢂ103/mL). Calibration curves of 1 were
performedfor each inidviudal experiment andthe
curves were usedto determine the quantity of 1 formed
in the corresponding BALF supernatant. The hydrolysis
of 10 in BALF supernatants from asthmatic patients
yielded a small but significant amount of 1 after 24 h of
incubation (14 ng, 3%) (Fig. 6). On the contrary, the
hydrolysis yield in BALF supernatants from healthy
volunteers after 24 h was less than 0.5%, similar to the
hydrolysis observed in saline (Fig. 6). As already observed
in BALF from BN rats, compound 10 was cleavedmore
rapidly in BALF supernatants from asthmatic patients
than in BALF supernatants from non-asthmatics over a
These results indicate that there is a correlation between
the in vitro hydrolysis rate of compounds in BALF
from challengedrats andtheir ability to inhibit metha-
choline-induced bronchoconstriction in ozone-treated
guinea pigs in vivo, that is compounds, which show
rapidhydrolysis in vitro, are active as anti-broncho-
constrictor agents in ozone-treatedanimals, whereas
compounds, which are not or only slowly hydrolysed in
vitro, are inactive in the guinea pig ozone model. Thus,
it seems that both the in vitro andthe in vivo model can
be useful to screen DAD.
Furthermore, the data that DAD are not cleaved in
lungs from non-ozone-treatedanimals but are hydro-
lysedin guinea pigs where an inflammatory process had
been evokedby pretreatment with ozone, provide evi-
dence that the concept of disease activated drugs may
work in vivo. A limitation of this in vivo model is that
only compounds with rapid cleavage can be tested,
since—for technical reasons—the model only allows
observation of animals for 1 h.
a
Table 2. Therapeutic window of compounds 1, 2, 6 and 8 in ozone- andnon-ozone-treatedguinea pigs following it administration
Compounds
Effect on blood pressure
Inhibition of bronchoconstriction
Non-ozone-treatedOzone-treatedNon-ozone-treatedOzone-treated
Vehicle
ꢀ1ꢁ5 (4)
+20ꢁ7 (3)
+6ꢁ2 (3)
+6ꢁ4 (3)
n.d.b
+4ꢁ3 (5)
+21ꢁ5 (3)
+3ꢁ3 (3)
+4ꢁ5 (3)
+4ꢁ3 (3)
19ꢁ15 (4)
8ꢁ10 (5)
68ꢁ11 (3)
75ꢁ8 (3)
76ꢁ7 (3)
18ꢁ15 (3)
1 (30 mg/kg)
2 (100 mg/kg)
6 (100 mg/kg)
8 (150 mg/kg)
45ꢁ22 (3)
0.3ꢁ7 (3)
21ꢁ15 (3)
n.d.
aShown are effects on mean arterial bloodpressure (as % change from baseline values) andinhibition of methacholine-inducedbronchoconstriction
(as% inhibition relative to the response before drug administration) for a selected dose of the compounds, 15 min after administration, as meansꢁSEM
(number of experiments in brackets). For effects on bloodpressure, ‘ ꢀ’denotes reduction and ‘+’ increase of blood pressure. The increase in blood
pressure induced by compound 1 is statistically significant (p<0.01; Student’s t-test).
bn.d., no data.

1800
B. Charpiot et al. / Bioorg. Med. Chem. 9 (2001) 1793–1805
periodof 24 h. The quantitative difference observed
between the BN rat BALF hydrolysis studies and the
human BALF supernatant studies is probably due to
the presence of a much larger number of inflammatory
cells in the BN rat BALF as comparedto the human
BALF supernatants.
necessarily argue against the concept of reduced side
effects with DAD. Although the compounds are cleaved
in the plasma andpossibly in the blood, systemic expo-
sure to active principles after inhalation of DAD should
be much lower compared with the drug levels produced
in the lung (as indicated by the improved therapeutic
window for DAD in the in vivo model).
In connection with the preceding studies, compound 10
was also incubatedin human plasma obtainedfrom two
healthy donors. Increased plasma protein extravasation
is a characteristic feature of asthma.16 The fact that 10 is
rapidly hydrolysed in human plasma (donor 1: 80%
after 1 h; donor 2: 40% after 1 h) indicates that, in the
diseased lung, plasma proteins (possibly among other
proteins) are directly involved in the hydrolysis of 10.
The rapid hydrolysis of the DAD in plasma does not
In conclusion, these data demonstrate that such DAD
couldoffer interesting novel therapies for chronic
inflammatory diseases. Reduced systemic side effects of
DAD couldresult from the local release of active drug
within the inflammatory site as well as from the potential
synergy associatedwith the simultaneous release of two
anti-inflammatory principles (e.g., PDE4 inhibitors and
steroids), resulting in reduced clinical effective doses.
Table 3. Characteristics of asthmatic andcontrol subjects andBALF cytology
Donors
Macr.a (%)
Eos.a (%)
PMNa (%)
Lymp.a (%)
Epith.a (%)
Squam.a (%)
Sex
Aas scores
FEV1 (%)b
Age
1 (asthma)
2 (asthma)
3 (asthma)
4 (asthma)
5 (asthma)
6 (asthma)
7 (asthma)
10 (control)
20 (control)
30 (control)
40 (control)
50 (control)
60 (control)
70 (control)
80 (control)
74
16
80
48
96
38
50
88
93
67
88
82
51
90
83
—
26
2
—
—
11
4
—
—
—
—
—
—
—
—
12
50
13
28
—
1
3
6
1
—
7
9
6
—
—
3
49
6
—
5
33
1
15
45
10
16
5
2
5
21
—
1
17
6
1
—
4
1
3
—
—
—
3
M
F
F
F
F
1
3
3
2
1
1
2
100
40
65
70
109
97
78
94
90
96
100
117
98
34
56
41
54
42
61
37
31
25
24
21
25
30
32
21
1
—
—
—
—
—
—
—
—
—
—
F
M
M
M
M
M
M
M
M
F
2
1
—
1
—
—
110
102
aMacr., Macrophages; Eos., eosinophils; PMN, peripheral mononuclear cells; Lymp., lymphocytes; Epith., epithelial cells; Squam., squamous cells
(other cells).
bForcedexpiratory volume in 1 s (% of predictedvalues).
Figure 6. Comparison of the hydrolysis of compound 10 (1.25 mg/250 mL) to compound 1, in BALF supernatants from asthmatics (n=7) (~),
healthy volunteers (n=8) (*) andin saline 0.9% ( n=2) (^). Shown are meansꢁSEM. (ꢃꢃꢃ denotes significant differences, p< 0.001, between
asthmatics andhealthy volunteers).

B. Charpiot et al. / Bioorg. Med. Chem. 9 (2001) 1793–1805
1801
Experimental
perature, the solvent was removedandthe residue pur-
ifiedby flash chromatography to give a white solid(125
mg, 41%). H NMR (CDCl3) d 1.35 (d, 12H), 1.38 (t,
Chemistry
1
1H NMR spectra were recorded with a Bruker AM-360
or a Varian Gemini-200; chemical shifts are given in
ppm (d) relative to Me4Si as internal standard. J values
are given in Hertz (Hz). Analyses were performedby the
analytical department of Novartis, Basle and are within
the ꢁ0.4% range of the theoretical value. Analytical
thin layer chromatography (TLC) was performedon
silica gel 60 F254 glass plates (HPTLC, E. Merck). Pre-
parative column chromatography was performedon
silica gel (230–400 mesh ASTM) under pressure. Solvents
were AR grade and used without further purification.
N,N-dimethylformamide, triethyl amine, dichloro-
methane were dried over molecular sieves before use.
Melting points were determined with a Buchi 512 appa-
ratus and not corrected. Mass spectra (MS) were recorded
in either the EI mode (on a VG TS 250) or the FAB mode
in a matrix of thioglycerol or nitrobenzyl alcohol using a
VG 70-SE mass spectrometer. All chemical yields repor-
ted are unoptimized. Extracts were dried over sodium
sulphate andsolvents removedunder reducedpressure.
3H), 1.6 (s, 6H), 2.98 (q, 2H), 3.8 (s, 3H), 4.3 (m, 2H),
4.45–4.65 (m, 4H), 6.52 (t, 1H), 6.75 (d, 2H), 7 (s, 1H),
7.1.4 (m, 7H). The oxalic salt, 3, was preparedas
described for 2 (70%); mp 124–126 ꢄC. 1H NMR
(CDCl3) d 1.35 (d, 12H), 1.45 (t, 3H), 1.6 (s, 6H), 3.2 (q,
2H), 3.85 (s, 3H), 4.38 (m, 2H), 4.5.6 (m, 4H), 6.65 (t,
1H), 6.74 (d, 2H), 7.1 (s, 1H), 7.12.4 (m, 6H), 7.55 (s,
1H). MS m/z 586 (MH+), 544, 396. Anal.
.
(C38H45NO10 0.5 H2O) C: calcd66.65; found66.89; H:
calcd6.77 found6.71; N: calcd2.05; found1.985.
1-phenyl-cyclopentanecarboxylic acid 2-[1-(3,5- bis(1 -
methylethoxy)-phenyl)-3-ethyl-7-methoxy-isoquinolin-6-
yloxy]-ethyl ester (4). This compoundwas synthesised
following the procedure described for 3 using the same
1
molar ratio: 47% of a white solid. H NMR (CDCl3) d
1.3.45 (d+t, 15H), 1.6.8 (m, 4H), 1.7–2 (m, 2H), 2.6–2.8
(m, 2H), 2.95 (q, 2H), 3.8 (s, 3H), 4.3 (m, 2H), 4.5 (m,
2H), 4.6 (m, 2H), 6.55 (t, 1H), 6.75 (d, 2H), 7 (s, 1H),
7.1–7.4 (m, 7H). The oxalic salt 4 was preparedas
already described for 2 (43%); mp 101–103 ꢄC. 1H
NMR (CDCl3) d 1.35 (d, 12H), 1.45 (t, 3H), 1.65–1.8
(m, 4H), 1.85–2 (m, 2H), 2.6–2.7 (m, 2H), 3.2 (q, 2H),
3.85 (s, 3H), 4.3–4.4 (m, 2H), 4.45–4.6 (m, 4H), 6.65 (t,
1H), 6.75 (d, 2H), 7.1 (s, 1H), 7.12–7.4 (m, 6H), 7.55 (s,
1H). MS m/z 612 (MH+), 440, 396, 217. Anal.
2,2-dimethyl-propionic
acid
2-[1-(3,5-bis(1-methyl-
ethoxy)-phenyl)-3-ethyl-7-methoxy-isoquinolin - 6 - yloxy]-
ethyl ester (2). A mixture of 1 (500 mg, 1.14 mmol) and
triethyl amine (320 mL, 2.3 mmol) in dichloromethane
(15 mL) was stirredunder argon at room temperature.
Pivaloyl chloride (215 mL, 1.75 mmol) was added to the
solution andthe reaction was stirredovernight at room
temperature. After disappearance of 1 (TLC (ethyl ace-
tate/hexanes 1:4) the solvent was removedandthe resi-
due purified by flash chromatography to give a white
solid(460 mg, 77%). 1H NMR (CDCl3) d 1.2 (s, 9H),
1.35 (d, 12H), 1.4 (t, 3H), 2.98 (q, 2H), 3.85 (s, 3H), 4.4
(m, 2H), 4.55 (m, 2H), 4.58 (m, 2H), 6.55 (t, 1H), 6.76
(d, 2H), 7.1 (s, 1H), 7.3 (s, 1H), 7.35 (s, 1H). MS m/z
524 (MH+), 396.
.
(C40H47NO10 0.5 H2O) C: calcd67.59; found67.33; H:
calcd6.8 found6.72; N: calcd1.97; found1.89.
2,2-diphenyl-propionic acid 2-[1-(3,5-bis(1-methyleth-
oxy)-phenyl)-3-ethyl-7-methoxy-isoquinolin-6-yloxy]-ethyl
ester (5). Compound 5 was synthesisedfollowing the
procedure reported for 3, using the same molar ratio:
1
87% of a colorless oil, H NMR (CDCl3) d 1.38 (d,
12H), 1.4 (t, 3H), 1.95 (s, 3H), 2.97 (q, 2H), 3.8 (s, 3H),
4.3 (m, 2H), 4.5–4.65 (m, 4H), 6.55 (t, 1H), 6.75 (d, 2H),
7 (s, 1H), 7.15–7.25 (m, 11H), 7.35 (s, 1H). MS m/z 648
(MH+), 440, 396, 181. Anal. (C41H45NO6.0.5 H2O) C:
calcd74.97; found74.6; H: calcd7.06 found7.1; N:
calcd2.13; found2.
Addition of a solution of oxalic acid (80 mg, 0.88 mmol)
in ether (3 mL) to a stirredsolution of the preceding
compound(460 mg, 0.88 mmol) in ether (2 mL) gave a
white precipitate, 2, which was recoveredby filtration.
(500 mg, 84%); mp 130 ꢄC (decomp.), 1H NMR
(DMSO) d 1.12 (s, 9H), 1.25.4 (d+t, 12H+3H), 2.87 (q,
2H), 3.75 (s, 3H), 4.37 (m, 2H), 4.45 (m, 2H), 4.65 (m,
2H), 6.57 (t, 1H), 6.75 (d, 2H), 7.3 (s, 1H), 7.45 (s, 1H),
7.55 (s, 1H). MS m/z 523(M+), 508, 481, 129. Anal.
Carbonic acid 2-[1-(3,5-bis(1-methylethoxy)-phenyl)-3-
ethyl-7-methoxy-isoquinolin-6-yloxy]-ethyl ester ethyl-
ester (6). Compound 6 was synthesisedaccording to
the method already described for 2 using the same
1
molar ratio: 30% of a white solid, H NMR (CDCl3) d
.
(C33H43NO10 0.5 H2O) C: calcd63.65; found63.76; H:
calcd7.12; found6.87.
1.2–1.45 (d+2t, 12H+6H), 2.95 (q, 2H), 3.8 (s, 3H),
4.25 (q, 2H), 4.4 (m, 2H), 4.5–4.8 (m, 4H), 6.52 (t, 1H),
6.75 (d, 2H), 7.32 (s, 1H), 7.35 (s, 1H), 7.5 (s, 1H). Anal.
(C29H37NO7) C: calcd68.08; found67.94; H: calcd7.29
found7.29; N: calcd2.74; found2.66. The oxalic salt
was prepared as already described for 2 (78%); mp 93–
2-Methyl-2-phenyl-propionic acid 2-[1-(3,5- bis(1-methyl-
ethoxy)-phenyl)-3-ethyl-7-methoxy-isoquinolin-6-yloxy]-
ethyl ester (3). A mixture of 2-methyl-2-phenyl pro-
pionic acid(100 mg, 0.6 mmol), oxalyl chloride (60 mL,
0.7 mmol) and few drops of DMF in dichloromethane
(2 mL) was stirredat room temperature for 4 h. After
removal of the solvent, a solution of the dried residue in
dichloromethane (4 mL) was slowly added to a solution
of 1 (220 mg, 0.5 mmol), triethyl amine (86 mL, 0.6
mmol) in dichloromethane (4 mL). (TLC (ethyl acetate/
hexanes 1:4). After 24 h under stirring at room tem-
98 ꢄC. H NMR (CDCl3) d 1.3–1.4 (d+t, 12H+3H),
1
1.45 (t, 3H), 3.2 (q, 2H), 3.85 (s, 3H), 4.25 (q, 2H), 4.47
(m, 2H), 4.57 (m, 2H), 4.65 (m, 2H), 6.65 (t, 1H), 6.75
(d, 2H), 7.2 (s, 1H), 7.35 (s, 1H), 7.55 (s, 1H). MS m/z
512 (MH+), 440, 396.
Carbonic acid bis-{2-[1-(3,5-bis(1-methylethoxy)-phenyl)-
3-ethyl-7-methoxy-isoquinolin-6-yloxy]-ethyl} ester (8).

1802
B. Charpiot et al. / Bioorg. Med. Chem. 9 (2001) 1793–1805
A mixture of 1 (1.3 g, 3 mmol), 1,10-carbonyldiimidazole
1.55 (s, 3H), 1.6–1.85 (m, 2H), 2.15 (m, 1H), 2.2–2.5 (m,
4H), 2.6 (ddd, J=6, 15, 15, 1H), 2.95 (q, 2H), 3.1 (m,
1H), 3.85 (s, 3H), 4.23 (m, 1H), 4.45–4.7 (m +q, 4H),
4.9 (m, 1H), 5 (m, 1H), 5.1 and5.5 (AX, J=18, 2H), 6.1
(s, 1H), 6.34 (dd, J=2, 7, 1H), 6.55 (t, 1H), 6.78 (d, 2H),
7.1 (s, 1H), 7.15 (d, J=9, 1H), 7.35 (s, 1H), 7.38 (s, 1H).
MS m/z 874 (MH+), 500, 440, 396. Anal.
(251 mg, 1.5 mmol) andNa CO3 (1.24 g, 9 mmol) in 2-
2
butanone (15 mL) was heatedat 65 ꢄC overnight. The
suspension was filtratedthrough Celite andthe Celite
washedseveral times with acetone. After solvent eva-
poration, the residue was dissolved in ethyl acetate,
washed with water and dried. Chromatography (eluent:
ethyl acetate/hexanes 1:1) yielded 8 as a white solid
.
(C49H60FNO10S 0.5 H2O) C: calcd66.65; found66.61;
H: calcd6.96; found6.98.
(1.06 g, 78%); mp 79–82 ꢄC; H NMR (CDCl3) d 1.35
1
(d, 12H), 1.4 (t, 3H), 2.95 (q, 2H), 3.84 (s, 3H), 4.4 (m,
2H), 4.55 (m, 2H), 4.65 (m, 2H), 6.5 (t, 1H), 6.75 (d,
2H), 7.05 (s, 1H), 7.3 (s, 1H), 7.35 (s, 1H). MS m/z 905
(MH+), 891, 863, 440. Anal. (C53H64N2O11 0.5 H2O)
C: calcd69.56; found69.5; H: calcd7.27; found7.26 N:
calcd3.06; found2.91.
Thiocarbonic acid bis-{2-[1-(3,5-bis(1-methylethoxy)-
phenyl)-3-ethyl-7-methoxy-isoquinolin-6-yloxy]-ethyl} ester
(9). The thiocarbonyl homodimer of 1, less polar than
11, was isolated as side product (28%) and char-
.
acterised: mp 78–82 ꢄC, H NMR (CDCl3) d 1.35 (d,
1
12H), 1.4 (t, 3H), 2.95 (q, 2H), 3.8 (s, 3H), 4.45 (m, 2H),
4.55 (m, 2H), 4.95 (m, 2H), 6.5 (t, 1H), 6.75 (d, 2H),
7.05 (s, 1H), 7.3 (s, 1H), 7.35 (s, 1H). MS m/z 921
(MH+), 817, 440, 396. Anal. (C53H64N2O10S 1 H2O) C:
calcd67.84; found68.06; H: calcd7.089; found7.26 N:
calcd2.99; found2.73.
Carbonic acid 2-[1-(3,5-bis(1-methylethoxy)-phenyl)-3-
ethyl-7-methoxy-isoquinolin-6-yloxy]-ethyl ester 2-(9-fluoro-
ꢀ,17-dihydroxy-10,13,16-trimethyl-3-oxo-6,7,8,9,10,11,12,
13,14,15,16,17-dodecahydro-3H-cyclopenta[a]phenantren-
17-yl)-2-oxo ethyl ester (10). To a solution of 1 (336
mg, 0.765 mmol), 4-dimethylaminopyridine (385 mg, 3.1
mmol) in methylene chloride (6 mL) was added 1,10-
carbonyldiimidazole (128 mg, 0.765 mmol). The reac-
tion was stirredat room temperature for 5 h. The dis-
.
Carbonic acid 2-[1-(3,5-bis(1-methylethoxy)-phenyl)-3-
ethyl-7-methoxy-isoquinolin-6-yloxy]-ethyl ester 2-(5-
hydroxy-4a,6a-dimethyl-2-oxo-8-propyl-2,4a,4b,5,6,6a,9a,
10,10a,10b,11,12-dodecahydro-7,9-dioxa-pentaleno[2,1-a]-
phenanthren-6b-yl)-2-oxo-ethyl ester (12). Compound 12
was synthesisedfollowing the procedure reportedfor
10, using budesonide in a third of the dexamethasone
molar ratio: 72% of a white solid(150 mg); mp 120 ꢄC
(decomp); 1H NMR (DMSO) d 0.85 (s+t, 3+3H), 0.8–
2.4 (m, 13H), 1.3 (d+t, 12+3H), 1.4 (s, 3H), 2.6 (m,
1H), 2.85 (q, 2H), 3.8 (s, 3H), 4.3 (m, 1H), 4.4 (m, 2H),
4.55 (m, 2H), 4.6–4.73 (m, 4H), 4.83 (d, 1H exchange-
able), 4.71.87 (AX, 2 d, 1H) and 5.2 (AX, 2 d+2 t, 2H),
5.9 (s, 1H), 6.15 (d, J=9, 1H), 6.55 (t, 1H), 6.72 (d, 2H),
7.31(d, J=9, 1H), 7.33 (s, 1H), 7.4 (s, 1H), 7.47 (s, 1H).
MS m/z 896 (MH+), 580, 484, 440. Anal.
appearance of
intermediate 7 was monitoredby TLC (ethyl acetate/
1
andformation of the stable
1
hexanes 2:1). [7, H NMR (CDCl3) d 1.35–1.4 (t+d,
3H+12H), 2.85 (q, 2H), 3.7 (s, 3H), 4.5–4.7 (m, 4H), 4.8
(m, 2H), 6.55 (t, 1H), 6.75 (d, 2H), 7 (d, 1H), 7.33 (s,
1H), 7.43 (s, 1H), 7.46 (s, 1H), 7.5 (d, 1H), 8.1 (d, 1H).
MS m/z 533 (M+), 518, 491]. Dexamethasone (300 mg,
0.765 mmol) and 6 mL of dichloromethane were added
to the solution andthe suspension stirredat 40 ꢄC for
48 h. TLC showedthe formation of two less polar
compounds. After evaporation of the solvent, the resi-
due was dissolved in ethyl acetate, neutralised to pH 7,
washedwith water andbrine, rdiedandchromato-
.
graphied(solvent: ethyl acetate/hexanes 1:1 to 2:1): 55%
of 10 as a white solid; mp 135 ꢄC (decomp); H NMR
(C52H65NO12 0.25 H2O) C: calcd69.35; found69.24; H:
calcd7.33; found7.35.
1
(CDCl3) d 0.9 (d, J=7, 3H), 1.05 (s, 3H), 1.1–1.9 (m,
5H), 1.38 (d, J=7, 12H), 1.4 (t, 3H), 1.55 (s, 3H), 2.2
(m, 2H), 2.21 (s, 1H exchangeable), 2.25.5 (m, 3H), 2.6
(ddd, J=6, 10.8, 11, 1H), 2.97 (q, 2H), 3.1 (m, 1H), 3.85
(s, 3H), 4.23 (m, 1H), 4.45 (m, 2H), 4.5.7 (m, 4H), 4.72
and4.97 (AX, J=17, 2H), 6.1 (s, 1H), 6.34 (dd, J=2, 9,
1H), 6.55 (t, 1H), 6.8 (d, 2H), 7.1 (s, 1H), 7.13 (d, J=9,
1H), 7.33 (s, 1H), 7.38 (s, 1H). MS m/z 858 (MH+),
Biological methods
Phosphodiesterase enzymatic assays. PDE3 enzyme pre-
paration (from human platelets). Platelet concentrate was
obtainedfrom the local bloodtransfusion center. Plate-
lets were washedonce with phosphate bufferedsaline
(PBS; NaCl 0.14 M, KCl 2.7 mM, KH2PO4 1.5 mM,
Na2HPO4 8.1 mM, adjusted to pH 7.4 at room tem-
perature) suspended in 10 mL of buffer H [sucrose
0.25 M, ethylenediamine-tetraacetic acid (EDTA) 1
mM, tris 40 mM adjusted to pH 7.4 with HCl], dithio-
threitol 1 mM andthe following protease inhibitor
solutions: 5 mL/mL of phenylmethyl-sulphonylfluoride
(7 mg/mL in 2-propanol), 1 mL/mL leupeptin andpep-
statin A (1 mg/mL each, in ethanol). After sonication
(15 s at 4 ꢄC, Branson probe sonicator), homogenates
were centrifugedat 2200 g. The pellet was resuspended
in the same volume of buffer H andthe sonication
repeated. Pooled supernates were stored at ꢀ20 ꢄC until
use as PDE enzyme source
.
484, 440, 396. Anal. (C49H60FNO11 1 H2O) C: calcd
67.18; found67.08; H: calcd7.13; found7.30. The car-
bonyl homodimer of 1, 8 (20%), less polar than 10, was
also isolated as side product.
Thiocarbonic acid 2-[1-(3,5-bis(1-methylethoxy)-phenyl)-
3-ethyl-7-methoxy-isoquinolin-6-yloxy]-ethyl ester 2-(9-
fluoro-11ꢀ,17-dihydroxy-10,13,16-trimethyl-3-oxo-6,7,8,
9,10,11,12,13,14,15,16,17-dodecahydro-3H-cyclopenta[a]-
phenantren-17-yl)-2-oxo ethyl ester (11). Compound 11
was synthesisedfollowing the procedure reportedfor 10
using 1,10-thiocarbonyldiimidazole and the same molar
ratio: 32% of a white solid(210 mg); mp 150 ꢄC
1
(decomp); H NMR (CDCl3) d 0.93 (d, J=7, 3H), 1.05
(s, 3H), 1.2–1.8 (m, 4H), 1.38 (d, J=7, 12H), 1.4 (t, 3H),

B. Charpiot et al. / Bioorg. Med. Chem. 9 (2001) 1793–1805
1803
PDE4 enzyme preparation (from human neutrophils).
Neutrophils were isolatedfrom human bloodusing
spontaneous sedimentation in 1% dextran and iso-
pycinic centrifugation over Histopaque 1119 and1077.
After washing, a suspension of 3ꢂ107 cells/mL PBS was
sonicated4 times 15 s at 4 ꢄC (Branson probe soni-
cator). Immediately after sonication, protease inhibitor
solutions (5 mL/mL of phenylmethyl-sulphonylfluoride
(7 mg/mL in 2-propanol), 1 mL/mL leupeptin andpep-
statin A (1 mg/mL each, in ethanol), dithiothreitol (0.15
[HBSSꢂ10 (Hank’s balancedsalt solution 10 ꢂ con-
centrate), 100 mL; EDTA 100 mM, 100 mL; HEPES (4-
(2-hydroxyethyl)piperazine-1-ethylsulfonic acid) 1 M, 10
mL; H2O, 1 l]. The lavages from sensitizedchallenged
animals (BALF+) were pooledandusedimmediately
or storedat ꢀ20 ꢄC for at least 2 months without loss of
activity. BALF from sensitizednon challengedanimals
(BALFꢀ) was obtainedin the same way from sensitized
animals, which were treatedon ady 28 with saline
insteadof OA.
mg/mL) andNa EDTA (0.37 mg/mL) were added. The
2
centrifugation supernatant (1500g, 10 min, 4 ꢄC) was
made 25% (v/v) in ethylene glycol and 1 mg/mL
CHAPS (3-[(3-cholamidopropyl)dimethylammonio]-1-
propanesulfonate). This preparation was storedat
ꢀ20 ꢄC until use as PDE enzyme source. Since human
PDE4B is 92% homologous to rat PDE4B,9 the rat
enzyme was substitutedfor the human PDE4B in order
to improve the availability of the enzyme.
Eosinophils in BALF (meanꢁSE). The number of
eosinophils was 25.18ꢁ4.97ꢂ106 in BALF+ and
0.17ꢁ0.054ꢂ106 cells /mL in BALFꢀ.
DAD solutions. Each compound(approximately 10
mg, exactly weighed) and a reference compound 13
(approximately 20 mg, exactly weighed) were dissolved
in 10 mL DMSO for the hydrolysis studies.
PDE4B cloning and expression. PDE4 cDNA coding
for the isoenzyme rat PDE4B17 was integrated(PDE4B;
single copy) at the pep4 locus of a Saccharomyces cere-
visiae strain lacking both of the wild-type yeast PDE
genes.18 The yeast strain was grown in 1 L cultures at
30 ꢄC, pelletedandfrozen until homogenization. Pel-
letedyeast (5 mL) was suspendedin 50 mL of buffer (10
mM Tris–hydroxymethyl-aminomethane, 1 mM EDTA,
1 mg/mL each of leupeptin andpepstatin A, 175 mg/mL
phenylmethyl-sulphonyl fluoride, 1 mM dithiothreitol,
pH 7.4 with HCl). After centrifugation, 15 g of glass
beads (425–600 mm, acid washed, Sigma Chemical Co.)
washedwith buffer were adedto the pellet. To this
slurry, 1 mL of buffer and60 mg of cholamidopropane
sulphonic acid were added and the slurry was vigorously
agitatedduring 4 h at 4 ꢄC. Disintegration of the yeast
cells was observedmicroscopically (phase-contrast
optics) as dark cells and was >30% (usually 50%). The
slurry was transferredto a coarse glass funnel andthe
homogenate collectedby suction andwashing of the
glass beads with a total of 15 mL buffer. Cell fragments
were separatedfrom cytosol by centrifugation (2000 g,
10 min, 4 ꢄC). The pellet was resuspended in 15 mL of
buffer andassayedfor PDE activity together with the
cytosol.
Hydrolysis studies. Five microlitres of a solution of
DAD+13 in DMSO was added to 500 mL of BALF or
Hank’s solution, as control, andthe samples were incu-
ꢄ
batedat 37 C for 1, 2, 4 and24 h. Thereafter, 250 mL of
a mixture of isopropanol and tert-butyl methyl ether (1/
4) was added to the sample. The suspension was mixed
for 2 min using a vortex-genie (Bender & Holbein AG)
speedcontrol 7. The sample was then centrifugedat
4000 rpm for 2 min andthe organic phase was removed.
The procedure was repeated three times. The dried
combinedorganic layers were dissolvedin 500
mL ace-
tonitrile. Each set of experiments was run twice.
Extraction of 250 mL of BALFꢁ without compound
was also usedas control.
HPLC analysis. Twenty microlitres of the 500 mL
acetonitrile sample were injectedonto the HPLC system
[LiChrosorb-250-RP-select-B (5 mm, 250ꢂ4 mm, E.
Merck) fittedwith a C-8 precolumn]. Isocratic elution
was achievedwith 30% water containing 10-cam-
phorsulfonic acid(0.5 g/L) and70% acetonitrile. The
flow was typically 1.5 mL/min, the temperature of the
column 40 ꢄC, andthe UV detection at 254 nm. Under
these conditions, the internal standard 13, 1, andthe
study compounds had retention times of 4, 5.3 and <20
min, respectively, except for 8 and 9 (the two homo-
dimers of 1) which had23 and48 min, respectively, as
retention time. Peaks corresponding to each compound
were integratedandthe ratio between area of 1 andarea
of 1+ area of DAD was usedfor quantitation. The
extraction yieldof the compounds was controlledby
comparison of the compound 13 peak area in the sam-
ples to the peak area produced by 5 mL of the mother
solution in 500 mL acetonitrile. The formation of dex-
amethasone was investigatedby retention time compar-
ison with a commercial sample, addition of a small
amount of commercial dexamethasone to the aliquot
andconfirmation that, under less polar conditions of
elution, no additional peak was formed.
Assay of PDE activity. Activity was assayedeither
19
by the methodof Thompson
Colicelli.17
or as publishedby
DAD hydrolysis in BN rat BALF
BALF. Male Brown-Norway (BN) rats (approxi-
mately 200 g) were usedfor the study of cell accumula-
tion in the lungs. Ovalbumin (OA: 10 mg/0.5 mL) was
mixedwith aluminium oxide (10 mg/0.5 mL) andinjec-
ted(sc) simultaneously with Bortadella pertussis vaccine
(0.25 mL/animal ip). Injection of OA, together with
adjuvant, was repeated 15 and 21 days later. On day 28,
the sensitizedanimals were restrainedin plastic tubes
andexposed(60 min) to an aerosol of OA (3.2 mg/mL)
using a nose only exposure. Animals were killed48 h
later with pentobarbital (250 mg/kg ip). The lungs were
lavagedusing three aliquots (4 mL) of Hank’s solution
In vivo model
Measurement of lung function and mean arterial blood
pressure in guinea pigs. Lung function andbloodpressure

1804
B. Charpiot et al. / Bioorg. Med. Chem. 9 (2001) 1793–1805
20
were measuredas describedpreviously.
The animals
either a reversible airway obstruction characterisedby
an increase of 15% of FEV1 after inhalation of 200 mg
of salbutamol or a positive challenge with carbachol.23
None of the subjects hadsmokedwithin the previous 2
years. No subject hadany bronchial or respiratory tract
infection during the month preceding the test and had
never taken systemic or inhaledcorticosteroids of any
form. The BALF supernatants of eight healthy volun-
teers (aged21–32 years; mean ꢁSD, 26ꢁ4 years) were
usedas a control group. Their pulmonary function was
within the normal range. None of the subjects were
current or previous smokers. No subject hadany bron-
chial or respiratory tract infection during the month
preceding the test. The study was done after informed
consent andapproval from the ethic committee of the
University of Montpellier.
were anaesthetized, paralysed and ventilated via a tra-
cheal cannula. Air flow was measuredat the trachea by
a pneumotachograph connectedto a differential pres-
sure transducer. Pressure changes within the thorax
were monitoredvia an intrathoracic cannula, connected
to a differential pressure transducer to determine the
pressure difference between the trachea and thorax
(transpulmonary pressure). From measurements of air
flow andtranspulmonary pressure, airway resistance
[RL, cm H2O/(L/s)] anddynamic compliance (C dyn, mL/
cm H2O) were calculatedafter each respiratory cycle
with a digital electronic respiratory analyzer. Mean
arterial bloodpressure was continuously recordedfrom
the carotidartery using a pressure transducer.
Anti-bronchoconstrictor effects in ozone-treated guinea
pig. Conscious male Dunkin–Hartley guinea pigs were
treatedwith ozone (3 ppm, 30 min). After a recovery
phase of 60 min, the animals were preparedfor lung
function measurement as described above. Transient
bronchoconstriction was induced twice in 10 minute
intervals by iv injections of methacholine (3.2 mg/kg).
The increase in RL obtainedby the secondinjection was
taken as 100% response. The test compoundwas admi-
nisteredintratracheally (0.1 mL) following the second
injection of methacholine. 5, 15, 30 and60 min there-
after, bronchoconstriction was induced by single iv
injections of methacholine. Anti-bronchoconstrictor
activity was expressedas % inhibition of the broncho-
constriction to methacholine relative to the response
before administration of the test compound.
Fiberoptic bronchoscopy and bronchoalveolar lavage
(BAL). Fiberoptic bronchoscopy was performedas pre-
viously described.23 Briefly, after premedication with
atopine and diazepam, and local anaesthesia with lido-
caine, 2% appliedto the upper respiratory tract, a
BFTR fiberoptic bronchoscope (Olympus, Tokyo,
Japan) was insertedinto the trachea. The BAL was
performedidentically in all patients in one of the sub-
segmental bronchi of the middle lobe with the injection
of several aliquots of sterile saline (total volume of 150
mL) reaspiratedby gentle syringe suction. Each
bronchoscopy was done with oxygen, and adrenaline
was readily available. A nebulization with 1 mg of sal-
butamol was performed after the procedure. Immedi-
ately after lavage, cytocentrifuge slides were made with
an aliquot of the lavage, andthe remaining BALF was
centrifugedat +4 ꢄC for 400 g for 5 min. The super-
natant was storedat ꢀ20 ꢄC until further utilisation.
Anti-bronchoconstrictor effects in healthy guinea pigs.
The anti-bronchoconstrictor effect of the test com-
pounds was also assessed in normoreactive guinea pigs.
In this case, the animals were treatedwith air insteadof
ozone anda higher dose of methacholine (10 mg/kg iv)
was administered in order to obtain the same degree of
bronchoconstriction as in ozone-treatedanimals.
Otherwise, the same procedure, as described above, was
used.
Total cell number in BALF. The total cell number was
126ꢁ20ꢂ103 cells/mL in BALF from asthmatic patients
and133 ꢁ16ꢂ103 cells/mL in BALF from the control
group (meansꢁSEM).
Asthma scores. The severity of asthma was rated
according to the clinical score of Aas24 usedto grade
chronic asthma from very mildform (score of 1) to
incapacitating disease requiring severe medication
(score 5; Table 3).
Data evaluation. Data from individual experiments
were presentedas means ꢁSEM andunpairedStudent’s
t-test was usedfor comparison between treatments at
individual time points. Significance was assumed at the
5% probability level.
Pulmonary function. Pulmonary function was assessed
by measuring% FEV1 andwas expressedas % of
predicted values as described.25
Test compound and route of administration. For intra-
tracheal administration, the test compounds were sus-
pended in a medium of tragacanth gum 0.5% in distilled
water. This was sonicatedfor 30–60 s at 100 W before
administration in order to reduce the particle size. The
solutions andsuspensions were preparedfreshly on the
day of the experiment.
Hydrolysis studies of 10. A solution of 10 [12.5 mL
(1.25mg] in DMSO (100 mg/mL) was added to 250 mL of
human BALF supernatant andincubatedat 37 ꢄC for 0
min, 5 min, 4 h, 15 h, 24 h, 48 h. After addition of
250mL isopropanol and tert-butyl methyl ether (1:4), the
sample was agitated2 min in a vortex-genie (Bender &
Holbein AG) speedcontrol 7–8. The sample was then
centrifugedat 4000 rpm for 2 min andthe organic phase
was removed. The procedure was repeated three times.
The organic layers were combined and dried under
argon flow at 37 ꢄC. Each set of experiments was run
twice anda set of experiments in saline 0.9% was run as
Hydrolysis of 10 in human BALF supernatants
Subjects (Table 3). The hydrolysis studies were per-
formedin BALF supernatants of seven patients with
asthma (aged34–61 years; mean ꢁSD, 41ꢁ14 years).
Asthma was defined according to the criteria of the
American Thoracic Society 21,22 andall the patients had

B. Charpiot et al. / Bioorg. Med. Chem. 9 (2001) 1793–1805
1805
control. Extraction of 250 mL of each BALF without
compoundwas also usedas control. It was checkedthat
no residual peak after extraction of the different super-
natant samples interferedin the UV detection of 1 and
that incubation of 10 in saline 0.9% yielded only a
negligible amount of 1.
3. Wright, L. C.; Seybold, J.; Robichaud, A.; Adcock, I. M.;
Barnes, P. J. Am. J. Physiol. 1998, 275, 694.
4. Naef, R. DE 4438737, 1995.
5. Stache, U.; Fritsch, W.; Alpermann, H. G. DE 27 35 110,
1979.
6. (a) Pozo, M.; Pulido, R.; Gotor, V. Tetrahedron 1992, 48,
6477. (b) Moris, F.; Gotor, V. Tetrahedron 1992, 48, 9869. (c)
Pozo, M.; Gotor, V. Tetrahedron 1993, 49, 10725.
7. Teranishi, K.; Kayakiri, T.; Mizutani, M.; Hisamatsu, M.
Biosci. Biotech. Biochem. 1994, 58, 1537.
Calibration curve of 1. Several concentrations of 1 (0,
10, 25 and50 ng) were spikedin 250
mL of BALF
supernatant of each individual, extracted and dried as
described above.
8. Engels, P.; Lubbert, H.; Fichtel, K. FEBS Lett. 1994, 350,
291.
9. Obernolte, R.; Bhakta, S.; Alvarez, R.; Bach, C.; Zuppan,
P.; Milkins, M.; Jarnagin, K.; Shelton, E. R. Gene 1993, 129,
239.
10. Tominaga, T.; Watanabe, A.; Tsuji, J.; Haruki, H.; Koda,
A. Gen. Pharmacol. 1995, 26, 353.
11. Ethanol, 2-[1-[3-(1-methylethoxy)-5-hydroxy-phenyl)-3-
isopropyl-7-methoxy-3,4-dihydro-6-isoquinolinyl] oxy]. 1H
NMR (DMSO) d 1 (d, 3H), 1.05 (d, 3H), 1.2–1.3 (2d, 6H), 1.9
(m, 1H), 2.4 (AX, J=18 and12 Hz, 1H), 2.64 (AX, J=18 and
6 Hz, 1H), 3 (m, 1H), 3.65 (s, 3H), 3.75 (m, 2H), 4.05 (m, 2H),
4.55 (m, 1H), 4.9 (t, 1H exchangeable), 6.4 (t, 1H), 6.5 (d, 1H),
6.58 (d, 1H), 6.8 (s, 1H), 7 (s, 1H). MS m/z 414 (MH+). Anal.
(C24H31NO5) C: calcd69.71 found69.68; H: calcd7.56; found
7.60; N: calcd3.39; found3.36.
HPLC-MS analysis. Samples were dissolved in 100
mL of acetonitrile/water (1:1) containing 200 ng/mL of
the internal standard terbinafine (MW 291). Twenty
microlitres were injectedonto the HPLC system. The
separation was achievedusing a Nucleosil C18-AB col-
umn (2.1ꢂ50 mm) packedwith C-18 reverse phase and
fittedwith a C-18 precolumn (Optimize Technologies).
Flow was typically 300 mL/min andtemperature of the
column was kept at 60 ꢄC. Eluent A was water contain-
ing 0.05% trifluoroacetic acid. Eluent B was acetonitrile
containing 0.05% trifluoroacetic acid. Percentage of B
was kept at 10% for 1 min then a gradient was run from
10 to 95% of B in 9 min. Under these conditions 1, ter-
binafine and 10 hadretention times of 6.7, 7.0 and8.3
min, respectively. Flow was split 1:4 prior to ionisation
allowing a flow of approximately 75 mL/min into the
electrospray (ESI) source. Molecules were ionisedby
ESI. Heatedcapillary temperature was kept at 220 ꢄC.
Under these conditions, 10, 1 andterbinafine mainly
yielded the MH+ ion (m/z 859, 440 and292, respec-
tively). SelectedIon Monitoring was usedas a scan/
detection method (span 0.3 amu, 1.5 s/scan). Multiplier
was set at 2000 volts. Peaks corresponding to each
12. Seltzer, J.; Bigby, B. G.; Stulbarg, M.; Holtzman, M. J.;
Nadel, J. A.; Ueki, I. F.; Leikauf, G. D.; Goetzl, E. J.;
Boushey, H. A. J. Appl. Phys. 1986, 60, 1321.
13. Evans, T. W.; Brokaw, J. J.; Chung, K. F.; Nadel, J. A.;
McDonald, D. M. Am. Rev. Respir. Dis. 1988, 138, 140.
14. Yeadon, M.; Wilkinson, D.; Darley Usmar, V.; O’Leary,
V. J.; Payne, A. N. Pulm. Pharmacol. 1992, 5, 39.
15. (a) Heaslip, R. J.; Lombardo, L. J.; Golankiewicz, J. M.;
Ilsemann, B. A.; Evans, D. Y.; Sickels, B. D.; Mudrick, J. K.;
Bagli, J.; Weichman, B. M. J. Pharmacol. Exp. Ther. 1994,
268, 888. (b) Hughes, B.; Howat, D.; Lisle, H.; Holbrook, M.;
James, T.; Gozzard, N.; Blease, K.; Hughes, P.; Kingaby, R.;
Warrellow, G. Br. J. Pharmacol. 1996, 118, 1183. (c) Raeburn,
D.; Underwood, S. L.; Lewis, S. A.; Woodman, V. R.; Bat-
tram, C. H.; Tomkinson, A.; Sharma, S.; Jordan, R.; Souness,
J. E.; Webber, S. E. Br. J. Pharmacol. 1994, 113, 1423. (d)
Turner, C. R.; Cohan, V. L.; Cheng, J. B.; Showell, H. J.;
Pazoles, C. J.; Watson, J. W. J. Pharmacol. Exp. Ther. 1996,
278, 1349.
compoundwere integratedandratio between
terbinafine was usedfor quantitation.
1 and
Statistical evaluation. The F-test was usedfor deter-
mination of the type of data distribution within the
various groups. Comparison between different groups
was performed using Student’s test (one-tailed); p values
<0.05 were taken as significant.
16. Glynn, A. A.; Michaels, L. Thorax 1960, 15, 142.
17. Colicelli, J.; Birchmeier, C.; Michael, T.; O’Neill, K.;
Riggs, M.; Wigler, M. Proc. Natl. Acad. Sci. U.S.A. 1989, 86,
3599.
18. Sullivan, M.; Egerton, M.; Shakur, Y.; Marquardsen, A.;
Houslay, M. D. Cell Signal 1994, 6, 793.
19. Thompson, W. J.; Terasaki, W. J.; Epstein, P. M.; Strada,
S. J. Adv. Sec. Mess. Phosphoprot. Res. 1979, 10, 69.
20. Buchheit, K. H.; Hofmann, A. Naunyn-Schmiedeberg’s
Arch. Pharmacol. 1996, 354, 355.
21. American Thoracic Society. Committee on Diagnostic
Standards for Non-tuberculous Respiratory Diseases. Am.
Rev. Respir. Dis. 1962, 85, 762.
Acknowledgements
We wish to express our gratitude to J. Brun, I. Donze,
M. Stefani, A. Dunbar for technical assistance andDr.
R. Hersperger for helpful andinsightful discussions.
References and Notes
22. Chai, H.; Farr, R. S.; Froelich, L. A.; Mathison, D. A.;
McLean, J. A.; Rosenthal, R. R.; Sheffer, A. L.; Spector, S. L.;
Townley, R. G. J. Allergy Clin. Immunol. 1975, 56, 323.
23. Godard, P.; Aubas, P.; Calvayrac, P.; Taib, J.; Michel,
F. B. Nouv. Presse Med. 1981, 10, 3141.
24. Aas, K. Heterogeneity of bronchial asthma: sub-popula-
tions—or different stages of the disease Allergy 1981, 36, 3.
25. Van Vyve, T.; Chanez, P.; Bousquet, J.; Lacoste, J. Y.;
Michel, F.-B.; Godard, P. Am. Rev. Respir. Dis. 1992, 146,
116.
1. (a) Torphy, T. J.; Undem, B. J. Thorax 1991, 46, 512. (b)
Raeburn, D.; Souness, J. E.; Tomkinson, A.; Karlsson, J.-A.
Prog. Drug Res. 1993, 40, 9. (c) Christensen, S. B.; Torphy, T.
Annu. Rep. Med. Chem. 1994, 29, 185.
2. Palfreyman, M. N.; Souness, J. E. Phosphodiesterase Type
IV Inhibitors. In Progress in Medicinal Chemistry; Ellis, G. P.,
Luscombe, D. K., Eds.; Elsevier Science: Amsterdam, 1996;
pp 1–52.