Synthesis of 4-(8-benzo[1,2,5]oxadiazol-5-yl-[1,7]naphthyridine-6-yl)-benzoic acid: A potent and selective phosphodiesterase type 4D inhibitor

Article abstract of DOI:10.1016/S0960-894X(01)00720-X

The synthesis of a 6,8-disubstituted 1,7-naphthyridine 1 and its characterization as a potent and selective phosphodiesterase type 4D inhibitor (IC₅₀=1.5nM) are described. The compound inhibited TNFα-release from human peripheral blood mononuclear cells and was orally active in a model of adjuvant-induced arthritis in rats.

Full text of DOI:10.1016/S0960-894X(01)00720-X

Bioorganic & Medicinal Chemistry Letters 12 (2002) 233–235
Synthesis of 4-(8-benzo[1,2,5]oxadiazol-5-yl-[1,7]naphthyridine-
6-yl)-benzoic Acid: a Potent and Selective Phosphodiesterase
Type 4D Inhibitor
Rene Hersperger,a,* Janet Dawsona and Thomas Muellerb
aNovartis Pharma AG, Arthritis & Bone Metabolism, CH-4002 Basel, Switzerland
bNovartis Pharma AG, Research Information Management, CH-4002 Basel, Switzerland
Received 2 August 2001;revised 26 October 2001;accepted 29 October 2001
Abstract—The synthesis of a 6,8-disubstituted 1,7-naphthyridine 1 and its characterization as a potent and selective phosphodiesterase
type 4D inhibitor (IC50=1.5nM) are described. The compound inhibited TNFa-release from human peripheral blood mononuclear
cells and was orally active in a model of adjuvant-induced arthritis in rats. # 2002 Elsevier Science Ltd. All rights reserved.
Introduction
nausea and emesis. Although the mechanism which
leads to the induction of this untoward effect is not fully
understood, one hypothesis is that binding of inhibitors
to the so-called rolipram high-affinity binding site might
play an important role.6 ArifloTM (SB 207499, cilomi-
last), which is the most advanced PDE4 inhibitor and is
currently undergoing clinical trials for asthma and
COPD, was apparently selected based on an optimized
ratio of PDE4 catalytic activity relative to affinity to the
rolipram high-affinity binding site.7,8
Phosphodiesterase type 4 (PDE4) belongs to a family of
at least 10 structurally, biochemically and pharmaco-
logically distinct isoenzymes that catalyze the hydrolysis
of the second messengers adenosine 30, 50-cyclic mono-
phosphate (cAMP) or guanosine 30, 50-cyclic monophos-
phate (cGMP) to the corresponding inactive 50-
nucleotide products.1,2 PDE4 is mainly expressed in
lymphocytes, monocytes, neutrophils, eosinophils and
mast cells and, therefore, is the most important member
of the PDE family in regulating immune and inflamma-
tory responses.3 The anti-inflammatory effect of PDE4
inhibitors has been demonstrated in various animal
models and, consequently, inhibition of PDE4 has been
proposed as a new therapeutic approach for the treat-
ment of a variety of chronic inflammatory diseases such
as asthma, chronic obstructive pulmonary diseases
(COPD) and rheumatoid arthritis.3,4 Four different iso-
genes and splice variants of human PDE4 were descri-
bed (PDE4A-D). The enzymes are characterized by
their selective, high affinity hydrolysis of cAMP and
sensitivity to inhibition by rolipram, an archetype PDE4
inhibitor. With the exception of PDE4C mRNAs of all
isoforms are expressed in the majority of cells of the
immune system.5 Nevertheless, the precise function and
the relevance of the PDE4 subtypes remain still to be
established. A common side effect of PDE4 inhibitors is
The drug seems to have an acceptable therapeutic win-
dow.
In contrast to many other companies, we focused our
efforts on finding PDE4D subtype selective compounds.
We recently reported the discovery of 6,8-disubstituted
1,7-naphthyridines as a novel class of potent and selec-
tive PDE4D inhibitors with strong anti-inflammatory
activity in an in vivo model of allergic asthma.9 The
most interesting compound of this series was the ben-
zoic acid substituted 1,7-naphthyridine 2.
Although this compound displayed a very promising
biological profile it contained a sterically non-hindered
aromatic nitro functionality. Aromatic amines and nitro
groups represent a clear structural alert and such com-
pounds are known to be potentially mutagenic and car-
cinogenic.10,11 Consequently, we were looking for an
appropriate replacement of the nitro group. Herein, we
describe the synthesis of the 6-benzo[1,2,5]oxadiazole
substituted 1,7-naphthyridine 1 and demonstrate that the
*Corresponding author. Fax:+41-61-324-3565;e-mail: rene.
hersperger@pharma.novartis.com
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00720-X

234
R. Hersperger et al. / Bioorg. Med. Chem. Lett. 12 (2002) 233–235
compound potently and selectively inhibited PDE4D and
thus, had a similar profile as its nitro analogue 2 (Chart 1).
standard reaction conditions for Suzuki12 or Stille-type
couplings13 were applied. Intermediate 10 was trans-
formed to the corresponding triflate 11 which was then
coupled to 4-carboxyphenylboronic acid by a second
Suzuki-type reaction.
Chemistry
The title compound 1 was synthesized by two altern-
ative routes using palladium-catalyzed cross-coupling
reactions as key steps. The necessary tin or boronic acid
benzofurazan reagents 7 and 8 were synthesized in good
overall yields from the commercially available amine 3
(Scheme 1). Bromination of 3 followed by cyclization
afforded the N-oxide 5 which was easily reduced to the
benzofurazan 6. Palladium-catalyzed stannylation led to
the tin compound 7. Formation of the boronic acid 8
was more tricky and using organolithium- or Grignard
reagents followed by quenching with triethyl borate gave
complex mixtures when using standard reaction condi-
tions. Very low-temperature, addition of N,N,N0,N0-tetra-
methylethylenediamine (TMEDA) and, particularly,
in-situ trapping of the organolithium intermediate by
triethyl borate were mandatory to prepare compound 8
in acceptable yield.
Alternatively, the benzoic acid moiety was attached to
the 1,7-naphthyridine ring first as depicted in Scheme 3.
In this case regioselective cyclization of the dinitrile 12
gave the amine 13 which after transformation into the
corresponding triflate, was subjected to the first Suzuki
coupling. The ether 14 was readily converted into the bro-
mide 15 which was then coupled to the boronic acid 8
under standard Suzuki conditions. In this way, multi-gram
quantities of 1 were synthesized. Interestingly, the free acid
1 could be transformed into the water-soluble N-methyl-
d-glucamine salt 1b by crystallization from methanol.
Results and Discussion
Compound 1 was measured against a panel of human
phosphodiesterases (Table 1). As predicted, 1 selectively
inhibited PDE4D (IC50=1.5 nM) and was 23-, 400-,
and 820-fold less potent on PDE4B, PDE4A, and
A first synthesis of compound 1 is depicted in Scheme 2,
using bromide 99 and either the easily available but
potentially toxic tin compound 7 or, alternatively, the
more difficult to prepare boronic acid 8. In both cases,
Scheme 2. Synthesis of 4-(8-benzo[1,2,5]oxadiazol-5-yl-[1,7]naphthyri-
dine-6-yl)-benzoic acid 1: (a) 7, Pd(dba)2, P(Ph)3, DMF, 120 ꢀC, 3 h;
(b) 8, Pd(dba)2, P(o-tol)3, toluene/DMF, 2 N Na2CO3, 110 ꢀC, 3 h;(c)
NaNO2, CF3SO3H/DMF, rt 3 h;(d) 4-carboxyphenylboronic acid,
Pd(dba)2, P(Ph)3, 2 N Na2CO3, DMF, 80 ꢀC, 2 h.
Chart 1.
Scheme 1. Synthesis of 5-trimethylstannanyl-benzo[1,2,5]oxadiazole 7
and 5-benzo[1,2,5]oxadiazole boronic acid 8. Reaction conditions: (a)
N-bromosuccinimide, AcOH, 60 ꢀC, 2 h;(b) KOH, EtOH, 60 ꢀC, 1 h;
(c) aq NaOCl, 0 ꢀC, 1 h;(d) P(Ph) 3, xylene, 140 ꢀC, 3 h, (e) hexa-
methyl-distannane, Pd(dba)2, P(Ph)3, toluene, 110 ꢀC, 3 h;(f) n-BuLi,
N,N,N’,N0-tetramethylethylenediamine, B(OEt)3, THF/n-pentane,
ꢁ100 ꢀC, 5 min.
Scheme 3. Alternative synthesis of 4-(8-benzo[1,2,5]oxadiazol-5-yl-
[1,7]naphthyridine-6-yl)-benzoic acid N-methyl-d-glucamine salt 1b:
(a) Na, MeOH, rt, 16 h;(b) aq NaNO , CF3SO3H/H2O, 0 ꢀC, 3 h;(c)
2
4-carboxyphenylboronic acid, Pd(dba)2, P(Ph)3, 2 N Na2CO3, DMF,
80 ꢀC, 2 h;(d) PBr 3, DMF, 100 ꢀC, 30 min;(e) 8, Pd(dba)2, P(o-tol)3,
2 N Na2CO3, DMF, 110 ꢀC, 2 h;(f) N-methyl-d-glucamine, MeOH.

R. Hersperger et al. / Bioorg. Med. Chem. Lett. 12 (2002) 233–235
235
Table 1. Inhibition of human phosphodiesterase activity, [3H]-roli-
pram binding and inhibition of TNFa release from human peripheral
blood mononuclear cells by compounds 1, 2 and SB207499
rheumatoid arthritis14 we tested our PDE4D selective
compound 1b in a rat model of adjuvant-induced
arthritis (Fig. 1). When dosed for 17 days after adjuvant
injection (5 mg/kg bid, po), compound 1b inhibited paw
swelling by 50%. The control compound, a COX-2
inhibitor, was very potent in this model which is known
to be very sensitive to this class of compounds.
1 (nM)a
2 (nM)a
SB207499 (nM)a
PDE4Ab
PDE4B
PDE4C
PDE4D
PDE1
PDE2
PDE3
PDE5
[3H]-rolipramc
TNFa-released
602 (ꢂ25)
34 (ꢂ0.5)
1230 (ꢂ39)
1.5 (ꢂ0.1)
>10,000
88 (ꢂ14)
49 (ꢂ7)
68 (ꢂ9)
1.0 (ꢂ0.2)
—
—
>10,000
—
398 (ꢂ7)
288 (ꢂ7)
813 (ꢂ13)
63 (ꢂ2)
>10,000
>10,000
>10,000
>10,000
40 (ꢂ13)
3467 (64)
>10,000
>10,000
Conclusion
>10,000
1.0 (ꢂ0.3)
We have demonstrated that the 6-benzo[1,2,5]oxadiazole
substituted 1,7-naphthyridine 1 could be readily prepared
by two alternative synthetic routes using palladium-cata-
lyzed cross-coupling reactions as key steps. In vitro, com-
pound 1 potently and selectively inhibited PDE4D and
thus had a very similar profile to its nitro analogue 2.
Compound 1 was negative in in vitro assays of mutageni-
city and genotoxicity. Oral anti-inflammatory activity of
1 was demonstrated in a rat model of adjuvant-induced
arthritis where it dose-dependently inhibited paw-swel-
ling. The full therapeutic potential of 1, particularly its
effect in models of other chronic inflammatory diseases
such as asthma and COPD, will be reported elsewhere.
0.6 (ꢂ0.2)
68 (ꢂ1)
115
aData indicated as mean IC50 of at least three experiments;S.E. mean
is given in parentheses.
bInhibition of PDE activity, see ref 9.
cInhibition of [3H]-rolipram binding.
dInhibition of LPS/interferon g stimulated TNFa release from human
peripheral blood mononuclear cells.
Acknowledgements
The authors thank E. Bacher, B. Glasser, and A.
Ursprung for their excellent technical assistance, and
wish to acknowledge the contributions of Novartis
Central Technologies for spectral data.
Figure 1. Activity of 1b in adjuvant-induced arthritis in rats. aWater.
bControl compound DuP 697;5-bromo-2-(4-fluoro-phenyl)-3-(4-
methanesulfonyl-phenyl)-thiophene. cThe compounds were given
orally for 17 days at the indicated dose and schedule, from the day of
adjuvant injection;** P<0.01 ANOVA followed by Dunnett’s test
post hoc. NS, not significant.
References and Notes
1. Soderling, S. H.;Beavo, J. A. Curr. Opin. Cell Biol. 2000,
12, 174.
2. Beavo, J. A. Physiol. Rev. 1995, 75, 725.
3. Teixeira, M. M.;Gristwood, R. W.;Cooper, N.;Hellewell,
P. G. Trends Pharmacol. Sci. 1997, 18, 164.
4. Doherty, A. M. Curr. Opin. Chem. Biol. 1999, 3, 466.
5. Muller, T.;Engels, P.;Fozard, J. R. Trends Pharmacol. Sci.
1996, 17, 294.
6. Barnette, M. S.;Christensen, S. B.;Underwood, D. C.;
Torphy, T. J. Pharmacol. Rev. Comm. 1996, 8, 65.
7. Torphy, T. J.;Barnette, M. S.;Underwood, D. C.;Gris-
wold, D. E.;Christensen, S. B.;Murdoch, R. D.;Nieman,
R. B.;Compton, C. H. Pulm. Pharmacol. Ther. 1999, 12, 131.
8. Christensen, S. B.;Guider, A;Forster, C. J.;Gleason, J. G.;
Bender, P. E.;Karpinski, J. M.;DeWolf, W. E., Jr.;Barnette,
M. S.;Underwood, D. C.;Griswold, D. E.;Cieslinski, L. B.;
Burman, M.;Bochnowicz, S.;Osborn, R. R.;Manning, C. D.;
Grous, M.;Hillegas, L. M.;Bartus, J. O.;Ryan, M. D.;
Eggleston, D. S.;Haltiwanger, R. C.;Torphy, T. J. J. Med.
Chem. 1998, 41, 821.
PDE4C, respectively. Furthermore, at a concentration
of 10 mM, 1 did not inhibit a variety of other phos-
phodiesterase family members (PDE1, 2, 3 and 5). Thus,
compound 1 had virtually the same activity profile
against all measured PDEs as the nitro analogue 2.
Compared to the reference compound SB 207499, 1 was
40-fold more potent in inhibiting PDE4D and showed a
significant better selectivity profile toward the other
PDE4 subtypes. Both compounds, 1 and 2 were not
selective in terms of inhibition of PDE4D catalytic
activity versus their ability to block the high-affinity
3[H]-rolipram binding site. In this respect, compound 1
had a very similar profile to SB 207499 which was also
not selective in our hands. On the cellular level, 1
potently inhibited LPS/interferon-gamma stimulated
TNFa-release from human peripheral blood mono-
nuclear cells (IC50=68 nM). Thus again, the compound
was roughly equipotent to the nitro-analogue 2. It is
noteworthy that the reference compound SB 207499 was
significantly less active in this assay.
9. Hersperger, R.;Bray-French, K.;Mazzoni, L.;Mueller, T.
J. Med. Chem. 2000, 43, 675.
10. Purohit, V.;Basu, A. K. Chem. Res. Toxicol. 2000, 13, 673.
11. Ashby, J.;Tennant, R. W. Mutat. Res. 1991, 257, 229.
12. Miyaura, N.;Suzuki, A. Chem. Rev. 1995, 95, 2457.
13. Stille, J. K. Angew. Chem., Int. Ed. Engl. 1986, 25, 508.
14. Francischi, J. N.;Yokoro, C. M.;Poole, S.;Tafuri, W. L.;
Cunha, F. Q.;Teixeira, M. M. Eur. J. Pharmacol. 2000, 399,
243.
Since the subtype non-selective, archetype PDE4 inhi-
bitor rolipram was reported to be active in a model of