A novel series of arylsulfonylthiophene-2-carboxamidine inhibitors of the complement component C1s

Article abstract of DOI:10.1016/j.bmcl.2006.01.036

Inhibiting the classical pathway of complement activation by attenuating the proteolytic activity of the serine protease C1s is a potential strategy for the therapeutic intervention in disease states such as hereditary angioedema, ischemia-reperfusion inj

Full text of DOI:10.1016/j.bmcl.2006.01.036

Bioorganic & Medicinal Chemistry Letters 16 (2006) 2200–2204
A novel series of arylsulfonylthiophene-2-carboxamidine
inhibitors of the complement component C1s
Nalin L. Subasinghe,* Jeremy M. Travins, Farah Ali, Hui Huang, Shelley K. Ballentine,
´
´
Juan Jose Marugan, Ehab Khalil, Heather R. Hufnagel, Roger F. Bone,
Renee L. DesJarlais, Carl S. Crysler, Nisha Ninan, Maxwell D. Cummings,
Christopher J. Molloy and Bruce E. Tomczuk
Johnson & Johnson Pharmaceutical Research and Development, L.L.C., 665 Stockton Drive, Exton, PA 19341, USA
Received 18 November 2005; revised 10 January 2006; accepted 11 January 2006
Available online 3 February 2006
Abstract—Inhibiting the classical pathway of complement activation by attenuating the proteolytic activity of the serine protease
C1s is a potential strategy for the therapeutic intervention in disease states such as hereditary angioedema, ischemia–reperfusion
injury, and acute transplant rejection. A series of arylsulfonylthiophene-2-carboxamidine inhibitors of C1s were synthesized and
evaluated for C1s inhibitory activity. The most potent compound had a Ki of 10 nM and >1000-fold selectivity over uPA, tPA,
FXa, thrombin, and plasmin.
Ó 2006 Elsevier Ltd. All rights reserved.
The complement cascade is a major component of the
innate immune system in mammals and other vertebrate
species.1 It plays a major role in the destruction of
invading microorganisms and the clearance of immune
complexes. Unregulated complement activation leading
to acute inflammation and tissue damage has been
implicated in the pathology of many disease states.2
Activation of the classical pathway has been implicated
in humorally mediated graft rejection,3 ischemia–reper-
fusion injury (IRI),4 hereditary angioedema (HAE),5
Vascular Leak Syndrome,6 and acute respiratory dis-
tress syndrome (ARDS).7 C1s is a trypsin-like serine
protease that is present as a proenzyme within the first
component of complement in the classical pathway.8
Under normal physiological conditions activated C1s
is inhibited by its endogenous inhibitor, C1 esterase
inhibitor (C1-INH). Pathological conditions result in
excessive activation of complement that is not sufficient-
ly inhibited by C1-INH. Furthermore, C1-INH can be
degraded by other proteases that are released during
complement-mediated inflammation. A small molecule
inhibitor of C1s would be a useful therapeutic agent in
the treatment of complement-mediated disease.9
We have previously reported the discovery of a novel
series of thiopheneamidine C1s inhibitors (1).10 These
compounds, while having very good selectivity over
thrombin, plasmin, and FXa, had poor selectivity over
the serine protease urokinase plasminogen activator
(uPA). Because the thiopheneamidine moiety appeared
to be a good S1 binding fragment for C1s, several com-
pound libraries containing this moiety were screened for
C1s inhibitory activity. Compound 2 was identified as
an 11 lM inhibitor of C1s, with 2-fold selectivity over
uPA. Substituting the isopropyl residue with a phenyl
residue provided compound 3 with Ki = 0.75 lM for
C1s and 13-fold selectivity over uPA.
MeS
S
S
N
N
N
N
HN
NH2
1 Ki (uPA) = 0.135 M; K
= 0.065 M
μ
μ
i (C1s)
Keywords: Classical pathway; Complement; C1s; Complement inhib-
itor; C1s inhibitor; Classical pathway inhibitor.
The observed selectivity profile for 1 and 3 may be relat-
ed, at least in part, to a difference at Lys614 (Gln192;
C1s numbering, followed by uPA/chymotrypsinogen
*
Corresponding author. Tel.: + 1 610 458 6066; fax: +610 458
8249; e-mail: nsubasin@prdus.jnj.com
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.01.036

N. L. Subasinghe et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2200–2204
2201
Table 1. SAR of sulfone substitution
Lys614
O
S
Ser617
MeS
S
O
R
HN
NH2
Arg563
Compound
R
C1s inhibition
Ki (lM)
Cys-Cys
Asp548
Arg557
2
3
4
5
6
7
8
CH2(CH3)2
Phenyl
11.0
0.75
0.41
2.29
1.88
1.79
2.24
2-Naphthyl
2-Pyridyl
1-Methylimidazol-2-yl
2-Methylfuran-3-yl
2-Thiopheneyl
Figure 1. Model of sulfone 3 bound to C1s. This figure shows selected
residues of C1s (PDB ID 1ELV; residue numbering is according to
1ELV; Ser617 is the catalytic serine, Asp548 is the canonical S1 acid
residue of trypsin-like serine proteases), with superposition of the
thiophene amidine as described previously.10 Cys-Cys indicates the
cystine disulfide formed by Cys613 and Cys644, and the dashed lines
indicate key intermolecular electrostatic contacts of the binding model
(see text). In this model, the side-chain conformation of Lys614 has
been adjusted from that described in the published crystal structure,11
to favor interaction with the sulfone moiety of 3. The distance between
Lys614:NZ and the two inhibitor sulfone oxygen atoms is 3.0 and
H
N
9
0.40
2.0
N
Br
10
Benzyl
phene-2-carboxylate 3613 with an appropriate thiol at
ꢀ78 °C resulted in substitution at the 4-position to give
the corresponding thioether, which was oxidized to the
sulfone 37 by treating with MCPBA in refluxing
DCM. Treating the sulfone 37 with sodium thiomethox-
ide at ꢀ78 °C affords primarily the nitro-displacement
product, which was converted to the amidine 38 by
treating with trimethylaluminum and ammonium chlo-
ride in refluxing toluene.14 When the thiomethoxide
addition was performed at room temperature, mixtures
of nitro and sulfone displacement products were
observed.
˚
3.5 A, respectively.
numbering in parentheses). In the original description of
the C1s structure, Gaboriaud et al.11 noted the potential
of a role for Lys614 in binding specificity. Lys614 of C1s
is adjacent to S1 and is positioned so that it could play a
role in restricting access to S1 and/or the catalytic
region. Our binding models (Fig. 1),10,12 with inhibitors
modeled into S1 of the apo C1s structure, are consistent
with the previously hypothesized importance of Lys614
in active-site binding. We extend this hypothesis to
propose that Lys614 of C1s and the corresponding
uPA residue Gln192 are involved in defining C1s selec-
tivity for compound 3. Figure 1 details some of the
key features of our proposed C1s binding model for 3.
Arg557 and Arg563 are positioned to constrain the side
chain of Lys614 to point in the direction of S1, and
Lys614:NZ can form a hydrogen bond with one or pos-
sibly both of the sulfone oxygen atoms of 3 as it is posi-
tioned in our model. Such interactions cannot be readily
modeled for 1. Additionally, Gln192, the corresponding
residue in uPA, is too short to achieve a similar interac-
tion and also appears to be restricted from leaning into
S1 due to a hydrogen bonding interaction with Lys143
(not shown). Therefore, it seems reasonable to propose
that C1s shows an increased preference for 3 due to
hydrogen bonding interactions with Lys614:NZ.
Compounds 9, 11, 12, 34, and 35 were synthesized
according to Scheme 2. The diazonium salt of amine
3915 was treated with sulfur dioxide in the presence of
cupric chloride to give the sulfonyl chloride 40.16 Sulfo-
nyl chloride 40 was reduced with sodium sulfite to give
the sulfinate 41.17 Addition of 41 to 5,7-dihalo-benz-
imidazole (42) afforded the sulfone 43.18 Sulfone 43
was reduced with sodium dithionite to 44, which upon
heating with formic acid gave the benzimidazole 45.
Benzimidazole 45 was directly converted to the amidine
to give compound 9. Alkylation of 45 with the appropri-
ate alkylhalide to give the regioisomers 46 and 47, fol-
lowed by amidination, gave compounds 11, 12, 34,
and 35. To overcome the poor yields associated with this
method of benzimidazole synthesis, an alternative route
(Scheme 3) was also pursued.
Several other aryl- and alkyl-sulfonyl derivatives
(Table 1) were synthesized in order to identify a core
fragment with significantly potent inhibitory activity to-
ward C1s. The benzimidazole 9 was chosen as a starting
point for further lead optimization. In this paper, we de-
scribe structure–activity relationships and in vitro bio-
logical results for this series of C1s inhibitors.
The sulfinate 48 (R2 = Br) was obtained by converting
the corresponding sulfonamide19 to the sulfonyl chloride
by treating with chlorosulfonic acid,19 followed by
reduction with sodium sulfite. Sulfinate 48 (R2 = H)
was obtained by reducing the commercially available
4-acetylamino-3-nitrobenzenesulfonyl chloride, fol-
lowed by hydrolysis of the acetamide. Addition of 48
to the thiophene 36 gave a mixture of 49 and 50. Thio-
Compounds 3–8 and 10 (Table 1) were synthesized
according to Scheme 1. Treating 4-Br-5-NO2-thio-

2202
N. L. Subasinghe et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2200–2204
O
S
O
S
O2N
S
MeS
O
O2N
S
O
1) NaSMe
1) R1SH / THF/ -78o C
Br
THF/MeOH
-78 C, 30-50%
R1
R1
S
2) MCPBA / DCM Δ
MeO2C
MeO2C
2) AlMe3 / NH4Cl
Toluene, Δ
30−60%
HN
37
NH2
36
38
40-70%
2 steps
Scheme 1.
MeS
MeS
MeS
CuCl2 2H2O
SO2
HCl
SO2Na
SO2Cl
NH2
Na2SO3
S
S
S
NaHCO3
87%
t-Butyl nitrite
MeO2C
35%
MeO2C
MeO2C
40
41
39
R2
O
S
MeS
S
O
O
O
S
MeS
N
N
R2
R2
R2
S
Na2S2O4
HCO2H
heat
42
MeO2C
N
N
MeO2C
EtOH/H2O
NH2
H2N
EtOH/H2O, AcOH
R2 = Br, Cl 60%
44
43
O
S
O
S
MeS
MeS
S
O
O
O
S
MeS
O
R2
R2
R3
S
R2
1) K2CO3, R3-X
52-100%
S
MeO2C
MeO2C
R3
47
N
N
N
N
MeO2C
NH
N
46
45
57% 2 steps
Scheme 2.
Br
O2N
O
O
O
O
O
O
R3
O O
SMe
S
S
NO2
S
R3
S
R3
CO2Me
DMF, 0o C --> rt
S
R3
NaSMe
R3SO2Na
S
S
S
+
THF/MeOH
-78o
C
O2N
48
O
OMe
O
OMe
O
OMe
R3
51
31% 3 steps
50
H2N
=
R2 = Br, H
49
R2
R2
R2 = Br, H
O
O
O
O
S
SMe
R2
SMe
1) Me3Al, NH4Cl
PhMe, 90o
2) Boc2O
41%
S
1) Fe, NH4Cl
2) HCOOH,
C
S
S
N
HN
NH
N
O
OMe
NH
Boc
HN
52
R
2 = Br, H
53
Scheme 3.
methoxide addition occurs selectively at the 5-position
of both 49 and 50 to give compound 51.
Compounds 14, 21, 22, and 33 were regioselectively
accessed (Scheme 4) starting from intermediate 51.
Compound 51 was diazotized and halogenated to give
compound 54. Displacement of the halogen with an
appropriate amine followed by reduction and cycliza-
tion in formic acid gave the benzimidazole 55. Com-
pound 45 (R2 = Br) was treated with benzeneboronic
acid in the presence of copper(II)acetate to provide
the N-phenyl derivative,20 which was treated with
trimethylaluminum and ammonium chloride in reflux-
ing toluene to give compound 13. Arylation occurs
regioselectively at the less hindered nitrogen of the
benzimidazole.
Reduction of the nitro group followed by heating in for-
mic acid provided the benzimidazole 52.
Benzimidazole 52 (R2 = Br) was converted to the corre-
sponding amidine and BOC-protected to provide a com-
mon scaffold that was alkylated with the respective
alkylhalide and deprotected to provide compounds 15–
20 and 23–30. Benzimidazole 52 (R2 = H) was alkylated
with benzylbromide and converted to the amidine
directly to give compounds 31 and 32.

N. L. Subasinghe et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2200–2204
2203
O
S
O
O
O
O
O
SMe
S
tBuONO,
CuCl2 or
CuBr2
SMe
3
R2
R2
R4
SMe
R2
S
1) R NH2,
3
DIEA 90-100%
2) Fe, NH4Cl
R
S
S
S
N
H2N
N
NO2
NO2
3) HCOOH
OMe
O
CH3CN
OMe
O
OMe
O
68-100%
82% steps 2&3
51 R2 = Br, H
54 R4 = Cl, Br
55
Scheme 4.
Table 2 lists the Ki values for the inhibition of C1s.21 N-
methylation at either nitrogen (11,12) has minimal effect
on affinity, suggesting that these nitrogens are not in-
volved in a hydrogen bond. Interestingly, when R2 is
bromine the effect of N-alkylmethyl and N-arylmethyl
substitution on affinity varies depending on the nitrogen
that is substituted.
in activity, while the corresponding regioisomers 15,
17, 19, 23, 25, 27, and 29 show only marginal enhance-
ment in activity. Unfortunately, some of the key pieces
of information required to establish a single plausible
binding model that explains the observed SAR is lack-
ing. Most significantly, a structure of an inhibitor or
substrate complex is not available, and key specificity
determining regions of the available apo structure are
disordered.11 Our modeling studies suggest that bound
inhibitors could be making hydrophobic interactions
with the region of loop 322 that is disordered in the
apo structure.11 Nevertheless, our results suggest that
there is a hydrophobic binding site for the N-alkylmeth-
yl and N-arylmethyl substituents that is proximal to the
bromine. Even though further substitution on the aryl
When the N-benzyl substituent is on the nitrogen prox-
imal to the bromine (20), there is a 13-fold improvement
in activity, whereas similar substitution at the distal
nitrogen (19) provides no enhancement in activity. This
pattern holds true for all the N-alkylmethyl and N-
arylmethyl substituents, where compounds 16, 18, 20,
24, 26, 28, and 30 provide 6- to 40-fold improvement
Table 2. SAR of the benzimidazole series
A
B
O
O
O
O
R2
N
S
R2
N
S
SMe
SMe
R1
S
S
N
N
R1
HN
HN
NH2
NH2
Compound
R2
R1
C1s inhibition Ki (lM)
A
B
11
12
13
14
15
16
Br
Br
Br
Br
Br
Br
Me
Me
Ph
0.33
0.40
0.31
0.07
0.18
0.21
Ph
Allyl
Allyl
17
18
Br
Br
0.20
0.4
0.05
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
H
Benzyl
Benzyl
0.03
0.38
0.33
(R)-a-Me-Benzyl
(S)-a-Me-Benzyl
2,6-Dichlorobenzyl
2,6-Dichlorobenzyl
2,6-Difluorobenzyl
2,6-Difluorobenzyl
2,5-Difluorobenzyl
2,5-Difluorobenzyl
2-Fluoro-5-nitrobenzyl
2-Fluoro-5-nitrobenzyl
Benzyl
0.16
0.17
0.24
0.17
0.42
0.02
0.04
0.01
0.03
H
Benzyl
0.40
0.58
H
2-Pyridylmethyl
2,6-Difluorobenzyl
2,6-Difluorobenzyl
Cl
Cl
0.10
0.10

2204
N. L. Subasinghe et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2200–2204
Table 3. Protease selectivity for compound 28
11. Gaboriaud, C.; Rossi, V.; Bally, I.; Arlaud, G. J.;
Fontecilla-Camps, J. C. EMBO J. 2000, 19, 1755.
12. In our earlier study,10 we described the structural basis for
the development of our binding model for the thiophene
amidines. Here we extend that model to the arylsulfone
series. We note that the caveats described previously,10
due to the limited structural information available regard-
ing the enzyme conformation in the inhibitor-bound state,
also apply to the present work.
13. Guanti, G.; Dell’Erba, C.; Spinelli, D. J. Heterocycl.
Chem. 1970, 7, 1333.
14. (a) Garigipati, R. S. Tetrahedron Lett. 1990, 31, 1969; (b)
Sidler, D. R.; Lovelace, T. C.; McNamara, J. M.; Reider,
P. J. J. Org. Chem. 1994, 59, 1231.
uPA
tPA
FXa
11.4
Thrombin
>15a
Plasmin
>13a
Trypsin
1.4
Ki (lM)
>10a
>10a
a No inhibition observed at this screening concentration.
group of the arylmethyl substituents does not provide
useful SAR information, the diminished activity of com-
pounds 21 and 22 suggests that substituents on the
methylene group can prevent the benzyl residue from
achieving a conformation that favors a positive interac-
tion with the hydrophobic binding pocket. The observed
loss of activity when the bromine was substituted with
hydrogen or chlorine (cf. 20–32 and 26–35) suggests that
this residue is contributing to affinity through intermo-
lecular hydrophobic contacts.
15. Illig, C. R.; Subasinghe, N. L.; Hoffman, J. B.; Wilson, K.
J.; Rudolph, M. J.; Marugan, J. J. U.S. Patent 6291514,
2001; Chem. Abstr. 2001, 135, 242131.
16. Ramsay, G. C. et al. J. Am. Chem. Soc. 1971, 93,
1166European Patent EP 983982.
17. Field, L.; Clark, R. D. In Organic Synthesis Collective;
Wiley: New York, 1963; Vol. IV, pp 674–677.
18. Hazelton, C. J. et al. Tetrahedron 1995, 51, 5597.
19. Mrozik, H. H. U.S. Patent 4005199, 1977.
20. Lam, P. Y. S.; Vincent, G.; Clark, C. G.; Deudon, D.;
Jadhav, P. K. Tetrahedron Lett. 2001, 42, 3415.
In conclusion, lead optimization studies around the aryl-
sulfonylthiophene-2-carboxamidine template has result-
ed in a series of N-benzylbenzimidazoles with good C1s
inhibitory potency and >1000-fold selectivity over uPA
(Table 3). Compound 28 also has good selectivity over
tPA, FXa, thrombin, and plasmin.21
21. Dissociation constants were determined at 37 °C in a 96-
well format using a Molecular Devices plate reader.
Varied concentrations of inhibitors in 10 lL dimethyl
sulfoxide were added to wells with 280 lL of assay buffer
(pH 7.5) which contained 50 mM HEPES, 0.2 M NaCl,
1% dimethyl sulfoxide, 0.05% n-octyl b-D-glucopyrano-
side, and substrate, and incubated for 15 min at 37 °C.
Reactions were initiated by the addition of 10 lL of
enzyme in assay buffer without dimethyl sulfoxide and
substrate, and the change in absorbance at 405 nm was
monitored for 15 min. IC50s were obtained from the
inverse slope of plots of the ratio of initial velocity in the
absence of inhibitor to initial velocity in the presence of
inhibitor as a function of inhibitor concentration. All
velocities were corrected for background substrate con-
version (no enzyme addition). The dissociation constant
(Ki) was calculated using the equation Ki = IC50/(1 + S/
Km) where S is the substrate concentration (Cheng, Y.;
Prusoff, W. H. Biochem Pharmacol. 1973, 22, 3099). Km
values for each enzyme–substrate pair were determined
from Hanes–Wolf plots using the same final buffer
conditions as Ki determinations. Respective substrate,
substrate concentration, and Km values for each enzyme
were: human C1s: benzyloxycarbonyl-Gly-Arg-S-benzyl,
45 lM in 200 lM DTNB (5,50-dithio-bis-[2-nitrobenzoic
acid]); 190 lM; human a-thrombin:succinyl-Ala-Ala-Pro-
Arg-p-nitroanilide, 100, 320 lM; human factor Xa:
References and notes
1. (a) Walport, M. J. N. Engl. J. Med. 2001, 344, 1058; (b)
Walport, M. J. N. Engl. J. Med. 2001, 344, 1140.
2. (a) Kirschfink, M. Immunol. Rev. 2001, 180, 177; (b)
Caliezi, C.; Wuillemin, W. A.; Zeerleder, S.; Redondo, M.;
Eisele, B.; Hack, C. E. Pharmacol. Rev. 2000, 52, 91.
3. Pascual, M. A.; Crespo, M.; Tolkoff-Rubin, N. Nefrologia
2001, 21, 327.
4. D’Ambrosio, A. L.; Pinsky, D. J.; Connolly, E. S. Mol.
Med. 2001, 7, 367.
5. Carugati, A.; Pappalardo, E.; Zingale, L. C.; Cicardi, M.
Mol. Immunol. 2001, 38, 161.
6. Lissoni, P.; Barni, S.; Cattaneo, G.; Archili, C.; Crispino,
S.; Tancini, G.; D’Angelo, L.; Magni, S.; Fiorelli, G. Int.
J. Biol. Markers 1990, 5, 195.
7. Langlois, P. F.; Gawryl, M. S. Clin. Immunol. Immunopa-
thol. 1988, 47, 152.
8. Cooper, N. R. Adv. Immunol. 1985, 37, 151; Arlaud, G. J.;
Gaboriaud, C.; Thielens, N. M.; Rossi, V. Biochem. Soc.
Trans. 2002, 30, 1001; Arlaud, G. J.; Gaboriaud, C.;
Thielens, N. M.; Budayova-Spano, M.; Rossi, V.; Fonte-
cilla-Camps, J. C. Mol. Immunol. 2002, 39, 383; Gal, P.;
Ambrus, G.; Zavodszky, P. Immunobiology 2002, 205, 383.
9. Buerke, M.; Schwertz, H.; Seitz, W.; Meyer, J.; Darius, H.
J. Immunol. 2001, 167, 5375.
10. Subasinghe, N. L.; Ali, F.; Illig, C. R.; Jonathan Rudolph,
M.; Klein, S.; Khalil, E.; Soll, R. M.; Bone, R. F.;
Spurlino, J. C.; DesJarlais, R. L.; Crysler, C. S.; Cum-
mings, M. D.; Morris, P. E., Jr.; Kilpatrick, J. M.;
Sudhakara Babu, Y. Bioorg. Med. Chem. Lett. 2004, 14,
3043.
benzyloxycarbonyl-D-Arg-Gly-Arg-p-nitroanilide,
100,
260 lM; human urokinase: pyro-Glu-Gly-Arg-p-nitroani-
lide, 86, 87 lM; human plasmin: H-D-Val-Leu-Lys-p-
nitroanilide, 150, 300 lM; human trypsin: benzyloxycar-
bonyl-D-Arg-Gly-Arg-p-nitroanilide, 60, 61 lM; human
tPA (2-chain): methlysulfonyl-D-cyclohexyl-Try-Gly-Arg-
p-nitroanilide, 200, 200 lM. Within-run assay coefficient
of variation (CV) was <10%; between-run CV was <20%.
22. Perona, J. J.; Craik, C. S. J. Biol. Chem. 1997, 272, 29987.