Novel, selective indole-based ECE inhibitors: Lead optimization via solid-phase and classical synthesis

Article abstract of DOI:10.1016/j.bmcl.2005.06.085

A novel class of indole-based endothelin-converting enzyme (ECE) inhibitors was identified by high throughput screening. We report systematic optimization of this compound class by means of classical and solid-phase chemistry. Optimized compounds with a bisarylamide side chain at the 2-position of the indole skeleton exhibit low-nanomolar activity on ECE.

Full text of DOI:10.1016/j.bmcl.2005.06.085

Bioorganic & Medicinal Chemistry Letters 15 (2005) 4201–4205
Novel, selective indole-based ECE inhibitors: Lead optimization
via solid-phase and classical synthesis
Michael Brands,* Jens-Kerim Erguden, Kentaro Hashimoto, Dirk Heimbach,
¨
Christian Schro¨der, Stephan Siegel, Johannes-Peter Stasch and Stefan Weigand
BAYER HealthCare AG, Business Group Pharma, Research & Development, D-42096 Wuppertal, Germany
Received 25 April 2005; revised 24 June 2005; accepted 27 June 2005
Abstract—A novel class of indole-based endothelin-converting enzyme (ECE) inhibitors was identified by high throughput screen-
ing. We report systematic optimization of this compound class by means of classical and solid-phase chemistry. Optimized com-
pounds with a bisarylamide side chain at the 2-position of the indole skeleton exhibit low-nanomolar activity on ECE.
Ó 2005 Elsevier Ltd. All rights reserved.
1. Introduction
Through screening of the Bayer compound collection,
indole derivative 1 has been identified to be a potent
ECE inhibitor.5 In a standard enzyme-linked immuno-
sorbent assay (ELISA) format,6 1 had an IC50 value of
1.8 lM. The interesting in vitro activity, the unprece-
dented structure as well as high selectivity observed for
this compound versus the related enzymes NEP and
ACE (IC50 values >10lM) rendered indole 1 into an
attractive starting point for a medicinal chemistry opti-
mization program.
Since its discovery in 1988, endothelin (ET-1), its metab-
olism, and potential implications in several diseases have
attracted considerable attention.1 Blockade of the action
of ET-1 by ET-receptor antagonists has been widely
studied in animal models as well as in clinical trials
aimed at evaluating their use in the treatment of various
diseases such as hypertension, congestive heart failure,
and cancer.2 Notably, bosentan, the first marketed ET-
receptor antagonist, was approved for the treatment of
pulmonary arterial hypertension in 2001.3 Since the
21-amino acid peptide ET-1 is formed by cleavage of
its precursor big-ET via the action of endothelin-con-
verting enzyme (ECE), inhibition of this metalloprotease
has become an attractive target to modulate ET-1 levels.
Several inhibitors of ECE have been described in recent
years.4 The majority of those investigational compounds
typically contains structural motifs such as thiols or
phosponates addressing the Zn-bearing catalytic center
of the enzyme. Such pharmacophoric groups, however,
can lead to detrimental pharmacokinetic properties. In
addition, selectivity versus related metalloproteases such
as neutral endopeptidase 24.11 (NEP) or angiotensin-
converting enzyme (ACE) has been reported to be diffi-
cult to achieve.4 Thus, there is still need to discover nov-
el, selective lead structures for the inhibition of ECE.
In this article, we describe a systematic approach to-
wards the optimization of ECE inhibitor 1 and disclose
the details of the resulting structure–activity relationship
(SAR).
2. Chemistry
As no obvious Zn-chelating group is present in ECE
inhibitor 1, we decided to systematically vary the struc-
tural features of the lead compound. For this, we fo-
cused on the three substituents at positions 1, 2, and 5
of the indole core (see Fig. 1) which can be subjected
Keywords: ECE inhibitor; Indole; Solid-phase chemistry.
*
Corresponding author. Fax: +49 202 364061; e-mail: michael.brands
@bayerhealthcare.com
Figure 1. Lead structure 1 from HTS.
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.06.085

4202
M. Brands et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4201–4205
to variations independently from each other. In addi-
tion, changes in the indole scaffold itself and position
shifts of the side chains were envisaged.
using a split-and-pool IRORI KanÒ method.7 Using this
synthetic route, 1332 new compounds were prepared,
which gave satisfactory spectral and analytical data,
and were therefore tested for their ECE inhibitory
activity.
Synthesis of indole derivatives was achieved by starting
from commercially available indole 2 through a stan-
dard sequence (Scheme 1). Alkylation of the indole
nitrogen in 2 was accomplished by treatment with 2-flu-
orobenzylbromide under basic conditions. After reduc-
tion, the resulting aminoindole 3 was treated with
pivaloyl chloride to give the corresponding amide 4.
Saponification of the ester moiety and coupling with
aniline yielded the final indole-2-carboxamide 1.
For the synthesis of the inverse amide at the 5-position
(12, Scheme 3), commercially available 4-hydrazinoben-
zoic acid (9) was refluxed with ethylpyruvate in i-PrOH/
AcOH, followed by coupling with neopentyl amine. The
resulting phenylhydrazone was treated under Fischer-in-
dole type reaction conditions using polyphosphoric
acid8 to yield the desired indole-5-carboxamide 10.
Alkylation of indole nitrogen and subsequent installa-
tion of a C-2 amide side chain was carried out as de-
scribed in Scheme 1.
Synthetic throughput was dramatically increased by
employing solid-phase organic chemistry as outlined in
Scheme 2. Thus, library synthesis began with indole 5
that was obtained from reductive amination of 4-(4-for-
myl-3-methoxyphenoxy)butyrylamide resin with meth-
Synthesis of benzimidazole derivative 17 (Scheme 4)
was achieved starting from commercially available 4-
yl-5-amino-1H-indole-2-carboxylate. Acylation of
5
nitrobenzene-1,2-diamine (13) through a standard
with diverse carboxylic acid chlorides generated 6. N-al-
kylation with benzyl and alkyl halides using LiHMDS
followed by ester hydrolysis afforded the indole carbox-
ylic acid 7. Amide formation of 7 with anilines was most
efficient using HATU (method A), whereas coupling of
amino phenols worked best with PPh3 and NBS (meth-
od B). Finally, 8 was released from the resin with TFA
in CH2Cl2. The combinatorial library was synthesized
synthesis sequence.9 After treatment with a trichloro-
acetimidate, the resulting intermediate 14 was convert-
ed into ester 15 under standard conditions.
N-benzylation of 15 employing basic conditions result-
ed in a 1:1 mixture of regioisomers from which the de-
sired 5-nitrobenzimidazole isomer was separated by
flash chromatography (cyclohexane/ethyl acetate) and
Scheme 3. Synthesis of indole-5-carboxamide 12. Reagents and
conditions: (a) ethyl pyruvate, AcOH, i-PrOH, reflux, 76%; (b) TBTU,
DIPEA, C(CH3)3CH2NH2, CH2Cl2, 0 °C to rt, 59%; (c) PPA, 120 °C,
5 h, 34%; (d) KOt-Bu, 18-crown-6, 0 °C, 2-fluorobenzylbromide, THF,
rt, 55%; (e) LiOH, MeOH/H2O 3:1, rt, 98%; and (f) HATU, pyr/DMF
2:1, aniline, rt, 47%.
Scheme 1. Synthesis of indole-2-carboxamide 1. Reagents and condi-
tions: (a) KOt-Bu, 18-crown-6, 0 °C, 2-fluorobenzylbromide, THF, rt,
86%; (b) Pd/C, ammonium formate, EtOH/ethyl acetate, 95%; (c)
C(CH3)3CH2COCl, NEt3, CH2Cl2, 0 °C to rt, 98%; (d) LiOH, MeOH/
H2O 3:1, rt, 89%; and (e) HATU, DIPEA, DMF, aniline, rt, 84%.
Scheme 4. Synthesis of benzimidazole 17. Reagents and conditions: (a)
methyl 2,2,2-trichloroacetimidate, AcOH, rt, 3 h, 92%; (b) AgNO3,
EtOH, reflux, 15 h, 99%; (c) KOt-Bu, 18-crown-6, 0 °C, o-fluor-
obenzylbromide, THF, rt, 29% desired regioisomer; (d) aniline, NaH,
THF, reflux, 65%; (e) SnCl2, EtOH, 32%; and (f) C(CH3)3CH2COCl,
NEt3, CH2Cl2, 0 °C to rt, 20%.
Scheme 2. Solid-phase synthesis of 8. Reagents and conditions: (a)
R1C(O)Cl, acetone, Et3N, rt; (b) LiHMDS, R2CH2Br, DMF, rt; (c)
KOH, MeOH/dioxane (1:2), rt; (d) method A (anilines): HATU,
DMF/Pyr (1:2), rt, or method B (amino phenols): PPh3, NBS, CH2Cl2,
Pyr, THF, À25 °C to rt; and (e) TFA, CH2Cl2, rt.

M. Brands et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4201–4205
4203
Table 1. Modification of the N-1 and C-5 side chains of the indole 1
(R1 at N-1/R2 at C-5 position of 1, IC50 values determined via
fluorescence read-out10.)
readily converted into amide 16. Reduction of the nitro
group and amide formation yielded benzimidazole
derivative 17.
Compound R1
R2
Method
IC50
(lM)
3. Pharmacology
1
Scheme 1
0.22
0.86
Novel compounds have been tested for their ability to
inhibit ECE-1 by using a high throughput assay employ-
ing an artificial, bradykinin-derived substrate allowing
for the rapid readout of a fluorescence signal.10 Selected
compounds have been assessed in a lower throughput
ELISA assay using big-ET-1 as the endogenous substrate
of ECE-1.6 Since the assay employing the fluorogenic sub-
strate turned out to be more sensitive (e.g., IC50 for lead
compound 1 is 0.22 lM in this assay vs 1.8 lM in the
ELISA), it was well suited for the ranking of compounds,
and therefore it was used for the optimization program.
18
19
20
21
22
23
24
25
26
27
28
29
30
12
31
Scheme 2
Scheme 111 5.9
Scheme 2
Scheme 1
Scheme 2
Scheme 2
Scheme 2
Scheme 1a
Scheme 2
2.2
5.1
First, we focused on changing the nature of the benzylic
group at the N-1 position of 1 (Table 1). Whereas replace-
ment of the phenyl group by cyclohexyl leads to an only
moderate drop in activity (18), direct attachment of a
phenyl group (19) as well as extension of the linker length
by one additional methylene unit (21) result in less active
compounds. A small substituent like a methyl group in m-
position of the benzyl group is not tolerated (23), indicat-
ing a tight binding pocket at that site, which is further
underlined by the low activity of a corresponding thiazole
bioisoster replacement (20). Thus, the only variations tol-
erated were the attachment of an additional fluorine atom
in the second ortho position (22) and its replacement by a
trifluoromethyl group (24).
0.65
3.0
0.17
0.74
0.30
We next explored the neopentyl group in the amide sub-
stituent at the 5-position. Chain elongation by one
methylene unit as in 25 proved to be detrimental for
activity, whereas the bulky bornane residue turned out
to be the only group tolerated in this position not bear-
ing a quaternary carbon atom (26). The necessity for the
quaternary center is clearly highlighted through compar-
ison of derivatives 27 and 28. Further alterations in the
C-5 side chain revealed that variations in the amide link-
er group are not tolerated: thioamide 29 shows some
residual activity, whereas replacement of the amide by
a sulfonamide (30) or its inverted counterpart (12) leads
to a dramatic loss in activity. Finally, the shift of the side
chain to the C-4 position of the indole core (31) deleted
inhibitory activity.
Scheme 112 0.12
Scheme 2
Scheme 1b
Scheme 1c
Scheme 3
6.4
0.66
6.7
7.0
Having already observed two tightly binding groups in
lead structure 1, we hoped that the C-2 side chain would
tolerate more variations allowing for increase in in vitro
potency and for tuning of physicochemical properties
(Table 2). Simple replacement of the phenyl group of
the C-2 amide by heterocycles led to a drop in activity
(32/33), with the pyridyl group (34) being the best
replacement. More interestingly, the 4-position tolerated
a range of substituents (35–42). In addition, some deriv-
atives in which an additional group was attached to the
3-position of the phenyl ring (entries 43/44) yielded high-
ly active compounds. Several of those compounds with
diverse substituents in the 3- and 4-positions of the C-
Scheme 113 7.4
a 4,4-Dimethylpent-2-ynoic acid was used in the amide formation step
followed by catalytic hydrogenation.
b LawessonÕs reagent was used for thioamide formation.
c 2,2-Dimethyl-propane-1-sulfonyl chloride was prepared from 2,2-di-
methyl-propane-Grignard reagent and sulfuryl chloride in THF.

4204
M. Brands et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4201–4205
Table 2. Modification of the C-2 side chain of the indole 1 (IC50 values lM determined
Table 3. Second generation SAR (IC50 values determined in ELISA6.)
via fluorescence read-out10.)
Compound
R
Method
IC50
>10
32
Scheme 2
33
Scheme 2
Scheme 2
Scheme 2
Scheme 2
Scheme 1
Scheme 1
Scheme 1
Scheme 2
Scheme 2
Scheme 2
2.3
Compound
R
IC50 (lM)
34
0.85
0.22
0.21
0.39
1.2
48
49
50
51
SO2NH2
SO2NHCH3
SOCH3
0.039
0.150
0.066
0.120
35
SO2CH3
2 arylamide have been found through the solid-phase
combinatorial approach described in Scheme 2. Thus
for the first time, analogs with improved in vitro activity
compared to lead compound 1 could be identified as
shown by their IC50 values in the big-ET ELISA6 (1:
1.8 lM; 41: 0.22 lM; 43: 0.18 lM; and 44: 0.50 lM).
On the contrary, all efforts to change the nature of the
C-2 amide linker led to inactive compounds as exempli-
fied by amine 45 and amide bioisoter 46. As already ob-
served for the C-5 side chain, a shift in the attachment
point of the arylamide side chain from C-2 to C-3 of
the indole skeleton deletes all inhibitory activity (47).
36
37a
38a
39a
40
0.17
0.20
0.44
0.13
Finally, a variation of the scaffold itself was explored.
However, benzimidazole analog 17 (Scheme 4) turned
out to be less active (IC50 = 1.3 lM fluorescence read-
out) compared to lead compound 1. The same trend
was observed for other benzimidazole derivatives that
were prepared (data not shown).
41
42
Thus in summary, the systematic optimization effort re-
vealed that major structural variations of lead structure
1 were only tolerated in the aryl part of the C-2 amide
side chain. This prompted us to perform a second-gener-
ation lead structure optimization program. With exten-
sive use of solid-phase synthesis, a comprehensive
program of derivatization concentrating on this site of
the molecule was carried out. Through this, several com-
pounds with a second arylamide group attached to the
3-position of the phenyl group were found to exhibit
high activity in the big-ET ELISA assay. Of particular
interest were those derivatives bearing an additional sul-
fonamide, sulfoxide or sulfone group in para position as
exemplified in Table 3, compounds 48–51. The highest
activity was observed for sulfonamide 48, for which in vi-
tro inhibitory activity on ECE was optimized by a factor
of nearly 50 when compared to screening lead 1 (IC50 in
ELISA is 1.8 lM). As observed for lead structure 1,
none of the described compounds including 48 show
any appreciable activity toward inhibition of ACE or
NEP at a concentration of 10 lM.
43a
Scheme 1
Scheme 1
(74)e
44b
(90)e
45c
Scheme 1
Scheme 1
>10
>10
46d
4714
Scheme 1
>10
a Amino or carboxy groups introduced as part of the anilines were appropriately
protected (N-Boc or N-Cbz; methyl ester) and deprotected under standard
conditions.
b Phenyl analog of 43 was coupled with methylpiperazine using HATU in DMF.
c 2-Hydroxymethyl indole derivative obtained from LiAlH4 reduction of 4 was con-
4. Conclusion
verted to 45 using PPh3, NBS, and aniline in THF.
d Oxadiazole formation was achieved by reaction of the carboxylic acid derived from 4
and phenyl amidoxime using PyBOP, NEt3, DME, and reflux.
e % inhibition at 5 lM.
Systematic variation of indole lead structure 1 led to the
discovery of highly potent and selective ECE inhibitors

M. Brands et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4201–4205
4205
5. Ergueden, J. K.; Krahn, T.; Schro¨der, C.; Stasch, J.-P.;
Weigand, S.; Wild, H.; Brands, M.; Siegel, S.; Heimbach,
D.; Keldenich, J. WO 03/28719, 2003; Chem. Abstr. 2003,
138, 304157.
6. Schro¨der, A.; Tajimi, M.; Matsumoto, H.; Schro¨der, C.;
Brands, M.; Andersson, K.-E. J. Urol. 2004, 172, 1171.
7. Nicolaou, K. C.; Xiao, X.-P.; Parandoosh, Z.; Senyei, A.;
Nova, M. P. Angew. Chem. 1995, 107, 2476; Angew.
Chem., Int. Ed. Engl. 1995, 34, 2289.
8. (a) Samsoniya, S. A.; Mumladze, E. A.; Chikvaidze, I. S.;
Suvorov, N. N. Khim. Geterosikl. Soedin. 1984, 20, 349–
351; [Chem. Heterocycl. Compd. (Engl. Transl.) 1984, 20,
280–282]; (b) Murakami, Y.; Tani, M.; Tanaka, K.;
Yokoyama, Y. Chem. Pharm. Bull. 1988, 36, 2023.
9. Louvet, P.; Lallement, G.; Pernot-Marino, I.; Luu-Dud,
C.; Blanchet, G. Eur. J. Med. Chem. 1993, 28, 71.
10. Johnson, G. D.; Ahn, K. Anal. Biochem. 2000, 286, 112.
11. Kato, Y.; Conn, M. M.; Rebek, J., Jr. J. Am. Chem. Soc.
1994, 116, 3219.
such as sulfonamide 48 representing a novel type of ECE
inhibitor that is structurally different from all the classes
described so far. The absence of any Zn-chelating ele-
ments in its structure is a distinct feature and gives rise
to the speculation that this compound class might exhib-
it an unexpected binding mode to the enzyme.15 There-
fore, compounds such as sulfonamide 48 and its close
analogs are expected to serve as valuable tools for struc-
tural biology as well as for pharmacological studies.
More details on the biological characterization and
pharmacology of these compounds will be published in
due course.
References and notes
1. Recent review article: Masaki, Y. Trends Pharmacol. Sci.
2004, 25, 219.
12. 1-Methylcyclopentyl acetic acid was prepared according
to: Bott, K. Chem. Ber. 1967, 100, 978.
13. Synthesis starts from 4-nitroindole-2-carboxylic acid ethy-
lester prepared according to method A: Parmerter, S. M.;
Cook, A. G.; Dixon, W. B. J. Am. Chem. Soc. 1958, 80,
4621.
2. Kirchengast, M.; Luz, M. J. Cardiovasc. Pharmacol. 2005,
45, 182; Krum, H.; Liew, D. Curr. Opin. Invest. Drugs
2003, 4, 298; Wu-Wong, J. R. Curr. Opin. Invest. Drugs
2002, 3, 1234; Ergul, A. Expert Opin. Ther. Patents 2003,
13, 33.
3. Cohen, H.; Chahine, C.; Hui, A.; Mukherji, R. Am. J.
Health-Syst. Pharm. 2004, 61, 1107; Chin, K.; Channick,
R. Exp. Rev. Cardiovasc. Ther. 2004, 2, 175.
4. Dogrell, S. A. Expert Opin. Ther. Patents 2004, 14, 655;
Jeng, A. Y. Curr. Opin. Invest. Drugs 2003, 4, 1076.
14. Synthesis of indole-3-carboxamides: Hassinger, H. L.; Soll,
R. M.; Gribble, G. W. Tetrahedron Lett. 1998, 39, 3095.
15. Computational chemistry studies toward an understand-
ing of the binding mode of compound 48 will be published
soon.