Design and optimization of aniline-substituted tetrahydroquinoline C5a receptor antagonists

Article abstract of DOI:10.1016/j.bmcl.2008.06.059

A series of aniline-substituted tetrahydroquinoline C5a receptor antagonists were discovered. A functionality requirement of ortho substitution on the aniline was revealed. Secondary anilines, in general, outperformed tertiary analogs in inhibition of C5a-induced calcium mobilization. Further enhancement of activity was realized in the presence of an ortho hydroxyalkyl side chain. The functional IC₅₀ of selected analogs was optimized to the single-digit nanomolar level.

Full text of DOI:10.1016/j.bmcl.2008.06.059

Bioorganic & Medicinal Chemistry Letters 18 (2008) 3852–3855
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Design and optimization of aniline-substituted tetrahydroquinoline C5a
receptor antagonists
a
b
a
b
b
Yong Gong a, , J. Kent Barbay , Mieke Buntinx , Jian Li , Jean Van Wauwe , Concha Claes ,
*
Guy Van Lommen b, Pamela J. Hornby a, Wei He a
a Johnson & Johnson Pharmaceutical Research and Development, Spring House, PA 19477-0776, USA
b Johnson & Johnson Pharmaceutical Research and Development, Turnhoutseweg 30, B-2340 Beerse, Belgium
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of aniline-substituted tetrahydroquinoline C5a receptor antagonists were discovered. A function-
ality requirement of ortho substitution on the aniline was revealed. Secondary anilines, in general, out-
performed tertiary analogs in inhibition of C5a-induced calcium mobilization. Further enhancement of
activity was realized in the presence of an ortho hydroxyalkyl side chain. The functional IC50 of selected
analogs was optimized to the single-digit nanomolar level.
Received 20 March 2008
Revised 13 June 2008
Accepted 17 June 2008
Available online 20 June 2008
Ó 2008 Elsevier Ltd. All rights reserved.
Keywords:
Complement C5a
C5aR
Antagonist
Aniline
Tetrahydroquinoline
Complement component C5a, a pro-inflammatory serum pro-
tein, is generated during activation of the complement cascade
by proteolysis of C5. Its interaction with the membrane bound G-
protein coupled C5a receptor (C5aR) mediates anaphylatoxic and
chemotactic effects. Excessive complement activation may contrib-
ute to many inflammatory and autoimmune conditions.1,2 Prevent-
ing the formation of C5a (as well as C5b) using antibodies that bind
to and block proteolysis of complement component C5 has been
clinically validated as a therapeutic strategy with the recent mar-
keting approval of SolirisTM (eculizumab) by the U.S. Food and Drug
Administration as the first therapy approved for paroxysmal noc-
turnal hemoglobinuria.3 C5aR has also been an important thera-
peutic target for complement modulation.2 The search for C5aR
antagonists has resulted in the discovery of different types of li-
gands, including peptides and peptidomimetics, as well as non-
peptidic small molecules.2,4
Recently, we have disclosed a tetrahydroquinoline series of
small molecule C5aR antagonists.5 A key feature of the series was
the existence/requirement of a tertiary amine substituent at the
5-position of the tetrahydroquinoline scaffold, as illustrated by
the activities of tertiary aminonaphthalenes 1 and 2 as contrasted
to the less active secondary amine 3 (Fig. 1). Optimization of the
aminonaphthalene portion was challenging, as only a limited
number of aminonaphthalenes were accessible and activity was
Figure 1. Representative aminonaphthalene-substituted tetrahydroquinoline C5aR
antagonists.5
sensitive to naphthalene ring modifications. We considered ani-
lines as truncated aminonaphthalenes, with the potential advan-
tage that their wide accessibility would offer more opportunity
for optimization. Here, we report our exploration of anilines as 5-
position substituents within the series. This subseries has revealed
distinct structure-activity relationships (SAR) as compared to the
aminonaphthalenes and has yielded analogs with excellent po-
tency as C5aR antagonists.
A general synthetic approach to the secondary (2°) and tertiary
(3°) aniline analogs is outlined in Scheme 1. The well-optimized
2,6-diethyl/dimethyl phenyl group was selectively installed in 5
via Suzuki coupling of dichloroquinolinone 4 with the correspond-
ing phenylboronic acids. Substitution of the 4-chloro group in 5
with a sodium alkoxide, reduction of the ketone with NaBH4, and
treatment of the alcohol with SOCl2 resulted in chloride 6. Alkyl-
ation of primary anilines with chloride 6 yielded 2° anilines.6 Tar-
geted 3° anilines were generated by reductive methylation of the
* Corresponding author. Tel.: +1 908 927 7923; fax: +1 908 526 6469.
E-mail address: ygong@prdus.jnj.com (Y. Gong).
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.06.059

Y. Gong et al. / Bioorg. Med. Chem. Lett. 18 (2008) 3852–3855
3853
Scheme 1. Reagents and conditions: (a) 2,6-di-R1-PhB(OH)2 (R1: Me or Et), Pd(PPh3)4, Na2CO3, H2O/toluene, reflux; (b) R2ONa/R2OH, reflux (R2: Me or CHMe2); (c) NaBH4,
MeOH; (d) SOCl2, CH2Cl2; (e) A-PhNH2, K2CO3, MeCN or DMF; (f) paraformaldehyde, NaBH(OAc)3, ClCH2CH2Cl, W, 120 °C, 10 min.
l
corresponding 2° anilines. Phenylether analogs 49 and 50 shown in
Fig. 3 were prepared similarly by alkylation of phenols with chlo-
ride 6.
Secondary anilines were originally prepared as precursors for
the targeted tertiary anilines. Some of these precursors were se-
lected and tested in the functional assay in comparison to tertiary
analogs. Surprisingly, those 2° anilines were not only active, but in
most cases were even more potent than their 3° counterparts (see
Table 2). For example, 2° aniline 25 was 4-fold more potent than
the corresponding 3° aniline 12. SAR trends in the 2° anilines were
similar to those in the 3° anilines. For instance, the synergistic
Previously demonstrated SAR within the aminonaphthalene
series suggested the importance of a tertiary amine substituent
at the 5-position of the tetrahydroquinoline core.5 Accordingly, a
set of tertiary anilines, listed in Table 1, were prepared initially
and tested in a C5a-induced calcium mobilization functional as-
say.7 Unsubstituted aniline analog
activity (IC50: 7.9
7
only offered weak
l
M), and was much (46-fold) less potent than
aminonaphthalene analog 2. However, certain substituted anilines
offered improved activities. For instance, the ortho-methyl-substi-
Table 2
C5a-induced calcium mobilization results for secondary anilines
tuted compound 8 improved potency over 7 by 10-fold to 0.79
lM
and an additional methyl or methoxy group at the m0-position was
favorable (11: IC50 0.69 lM and 12: IC50 0.56 lM, respectively).
Combining the preferred o-methyl, m0-methoxy aniline group with
a 2,6-dimethylphenyl substituent at the tetrahydroquinoline 2-po-
sition resulted in a modest decrease in activity relative to 12 (16:
IC50 3.2 lM); therefore the 2,6-diethylphenyl group was generally
retained in subsequent analogs.
Compound
A
Ca2+ flux Inhibition IC50
IC50 Ratio (NCH3/
NH)a
(
lM)
19
20
21
22
23
24
25
26
27b
28
o-Me
0.91
7.4
1.5
1.4
2.5
0.43
0.14
0.29
>10
>10
0.87
1.7
Table 1
m0-OMe
C5a-induced calcium mobilization results for tertiary anilines
o-Me, m0-Cl
o-Me, m0-F
o-Me, m0-Ph
o-Me, m0-Me
o-Me, m0-OMe
o-Me, m0-CF3
o-Me, m0-OH
o-Me, m0-
CH2OH
>4
1.6
4.0
29
30
o-Me, p-OMe
o-OMe, m0-
OMe
2.4
0.42
2.6
6.2
Compounda
R1
A
Ca2+ flux Inhibition IC50
(lM)
31
32
33
34
35
36
37
38
39
40
o-Cl, m0-OMe
o-Ph, m0-OMe
o-Et
0.38
>10
0.73
5.0
0.41
2.5
7
8
9
Et
Et
Et
Et
Et
Et
Et
Et
H
o-Me
7.9
0.79
2.6
>10
0.69
0.56
6.2
2.6
0.95
3.2
o-Me, m0-Cl
o-Me, m0-Ph
o-Me, m0-Me
o-Me, m0-OMe
o-Me, p-OMe
o-OMe, m0-OMe
o-Cl, m0-OMe
o-Me, m0-OMe
o-Me, m-OMe
o-Me, o0-Me
10
11
12
13
14
15
16
17
18
o-CH2NH2
o-CO2H
o-CONH2
o-CO2Me
o-COMe
>10
>10
>10
>10
Et
o-CF3
o-CN
Me
Me
Me
4.9
>10
>10
a
Values are IC50 ratios between tertiary anilines in Table 1 and corresponding
secondary aniline counterparts in Table 2.
a
b
All compounds were tested as racemates.
Compound 27 was prepared by treatment of 25 with BBr3 in CH2Cl2.

3854
Y. Gong et al. / Bioorg. Med. Chem. Lett. 18 (2008) 3852–3855
effect of o-methyl and m0-methoxy substitutions on activity was
again observed (disubstituted aniline 25: IC50 0.14 M). While
and Val286, which have been proposed to be involved in triggering
activation of the receptor.9 The i-propoxy oxygen in the central
part of the molecule forms a hydrogen bond to Tyr258. The
o-hydroxyethyl-phenyl moiety is involved in two interactions: a
l
holding the ortho-position constant as methyl, in addition to the
methoxy group at the m0-position, an optionally trifluorinated
methyl group was also preferred, halo and phenyl groups were tol-
erated, and hydroxy and hydroxymethyl substituents reduced
activity (compounds 21–28). An initial screen of additional 2°
o-monosubstituted anilines failed to identify compounds offering
significant further potency increases (compounds 33–40).
The introduction of functional groups on the anilines was in-
tended to pick up additional interactions between ligand and bind-
ing site (as discussed later) to improve potency. Although initial
results with m0-hydroxy in 27 and m0-hydroxymethyl in 28 were
disappointing, a significant advancement in potency was observed
in the case of o-hydroxyalkyl anilines (Table 3). o-Hydroxymethyl
cation-p interaction between Arg206 and the phenyl ring, and a
direct hydrogen bond between Ser138 and the hydroxyl group in
the o-hydroxyethyl chain. This binding mode can serve as a start-
ing point to understand the SAR in this series of compounds.
Further SAR extension was briefly explored, as represented by
the examples in Fig. 3. Aminopyridine 48 was about 30-fold less
potent than aniline analog 41. Replacement of the N-linkage in ter-
tiary aniline 12 or secondary aniline 25 with an O-linkage in phen-
ylether 49 offered comparable submicromolar activity. For aniline
46, the phenylether linked analog 50, however, lost potency by
about 80-fold, which indicated SAR differences between anilines
and phenylether analogs.
analog 41 possessed impressive activity with an IC50 of 0.12 lM.
The potency was further improved to double-digit nanomolar with
either additional substituents on the aniline ring (compounds 42–
45), or homologation to o-hydroxyethyl analog 46. A single-digit
nanomolar IC50 was achieved when the methoxy group on the tet-
rahydroquinoline core in 46 was replaced by i-propoxy in 47.
Compound 47 was docked into a homology model of human
C5aR.8 Fig. 2 depicts the putative ligand binding site9 and the pro-
posed binding mode of 47.10 In this binding mode, the diethylphe-
nyl group sits in the hydrophobic pocket close to residues Ile116
Representative compounds were also tested in a [125I]C5a radi-
oligand binding assay.11 The results are listed in Table 4. Selected
compounds showed activity in this assay, with K0is ranging from
submicromolar (unsubstituted tertiary aniline 7) to low double-di-
git nanomolar (substituted secondary anilines 25, 45, and 46). Fur-
ther in vitro characterization, including receptor and species
selectivity, will be published separately.
Compound 2512 was selected for a pharmacokinetic (PK) study
in rats and results are presented in Table 5. Upon intravenous dos-
ing at 2 mg/kg, compound 25 had a fairly long terminal elimination
half-life of over 6 h. Volume of distribution and systemic clearance
were relatively high. When compound 25 was dosed orally at
10 mg/kg, systemic exposure was achieved with t1/2 over 5 h and
bioavailability of 21%.13
Table 3
C5a-induced calcium mobilization results for o-hydroxyalkyl anilines
In summary, a novel aniline-substituted tetrahydroquinoline
series of C5aR antagonists were discovered. Distinct SAR was
revealed as compared with those of the previously disclosed
aminonaphthalene-substituted antagonists. Micromolar to submi-
cromolar IC50s of initial tertiary anilines were improved with sec-
ondary aniline analogs. Further enhancement of activity was
realized with the introduction of an ortho hydroxyalkyl side chain.
The receptor binding affinity, putative binding mode, and PK prop-
Compound
R2
n
A
Ca2+ flux Inhibition IC50
(lM)
41
42
43
44
45
46
47
Me
Me
Me
Me
Me
Me
CHMe2
1
1
1
1
1
2
2
H
0.12
m0-Cl
m0-CF3
p-F
p-Cl
H
0.026
0.022
0.053
0.071
0.030
0.007
H
Figure 3. Aminopyridine and phenylether analogs.
Table 4
[
125I]-C5a binding assay results for selected compounds
Compound
Ki ( M)
7
25
45
46
l
0.68
0.010
0.013
0.021
Table 5
Rat PK resultsa for compound 25
IV (2 mg/kg)
PO (10 mg/kg)
t1/2: 6.3 ( 1.2) h
t1/2: 5.3 ( 1.8) h
Vss: 4.5 ( 0.2) L/kg
Cl: 21 ( 9) mL/min/kg
Cmax: 0.29 ( 0.3) mg/mL
AUC: 42 ( 12) min ꢀ mg/mL
F: 21 ( 6)%
a
Values are means of four rats, standard deviation is given in parentheses.
Figure 2. Binding mode of 47 in human C5aR homology model.

Y. Gong et al. / Bioorg. Med. Chem. Lett. 18 (2008) 3852–3855
3855
An attempt to make an o-CH2CO2H analog was complicated by the cyclization
of 2-aminophenyl acetic acid to the oxindole.
7. The calcium mobilization assay was performed using the human monocytic
cell line U937 (ATCC CRL-1593). For details, see Ref. 5.
8. The human C5aR homology model was constructed using bovine rhodopsin X-
ray crystal structure (PDB code 1L9H) as template. Both model construction
and molecular docking were done using Maestro 6.0, module Prime and Glide
(Schrodinger LLC, Portland, OR).
erties of representative compounds were also explored. We hope
that these studies may aid in the eventual development of orally
bioavailable C5aR antagonists capable of modulation of the com-
plement response.
Acknowledgments
9. (a) Gerber, B. O.; Meng, E. C.; Dotsch, V.; Baranski, T. J.; Bourne, H. R. J. Biol.
Chem. 2001, 276, 3394; (b) Higginbottom, A.; Cain, S. A.; Woodruff, T. M.;
Proctor, L. M.; Madala, P. K.; Tyndall, J. D. A.; Taylor, S. M.; Fairlie, D. P.; Monk, P.
N. J. Biol. Chem. 2005, 280, 17831; (c) Nikiforovich, G. V.; Marshall, G. R.;
Baranski, T. J. Biochemistry 2008, 47, 3117.
We are grateful to the ADME team at Johnson & Johnson Phar-
maceutical Research and Development, Spring House, PA, for phar-
macokinetic analysis, and to Dr. Christopher Teleha for providing
intermediate 4.
10. Compound 47 was a racemate. Both (R)- and (S)-enantiomers were docked to
the model. The binding mode of (R)-enantiomer was favored over (S)-
enantiomer and was illustrated in Fig. 2. Chiral separation/synthesis was
planned for further profiling.
References and notes
11. The [125I]C5a radioligand binding assay was conducted using dibutyryl cAMP-
differentiated U937 cells. For details, see Ref. 5.
1. (a) Mizuno, M.; Cole, D. S. Expert Opin. Investig. Drugs 2005, 14, 807; (b)
Holland, M. C. H.; Morikis, D.; Lambris, J. D. Curr. Opin. Investig. Drugs 2004,
5, 1164.
2. Monk, P. N.; Scola, A.-M.; Madala, P.; Fairlie, D. P. Br. J. Pharmacol. 2007, 152,
429.
12. To 5-chloro-2-(2,6-diethylphenyl)-4-methoxy-5,6,7,8-tetrahydroquinoline (6,
0.13 g, 0.40 mmol) in DMF (1 mL) were added 5-methoxy-2-methylaniline
(0.22 g, 1.6 mmol) and potassium carbonate (0.22 g, 1.6 mmol). The suspension
was stirred at ambient temperature for five days, and then heated at 40 °C for
another day. The mixture was purified by HPLC (10–90% acetonitrile in water
containing 0.05% TFA). The combined pure fractions were partially
concentrated, basified with saturated sodium bicarbonate, and extracted
with dichloromethane. The organic extract was washed with water, dried
over magnesium sulfate, filtered, and concentrated, yielding aniline 25 as a
white foam (0.12 g, 70% yield). 1H NMR (CDCl3, 300 MHz) d 7.28 (t, 1H,
J = 7.5 Hz), 7.14 (d, 2H, J = 7.2 Hz), 6.98 (d, 1H, J = 8.7 Hz), 6.63 (s, 1H), 6.50 (d,
1H, J = 1.8 Hz), 6.24 (dd, 1H, J1 = 2.4 Hz, J2 = 8.1 Hz), 4.92 (s, 1H), 3.83 (s, 3H),
3.76 (s, 3H), 3.66 (br, 1H), 3.06 (dd, 1H, J1 = 5.1 Hz, J2 = 18 Hz), 2.84 (m, 1H),
2.26–2.46 (m, 5 H), 2.06 (s, 3H), 1.90–2.03 (m, 2H), 1.64 (m, 1H), 1.10 (m, 6H).
MS m/z 431.27 (MH)+. Anal. calcd for C28H34N2O2ꢀ0.5 H2O: C, 76.50; H, 8.03; N,
6.37. Found: C, 76.48; H, 7.96; N, 6.16.
3. Zareba, K. M. Drugs Today 2007, 43, 539.
4. (a) Schnatbaum, K.; Locardi, E.; Scharn, D.; Richter, U.; Hawlisch, H.; Knolle, J.;
Polakowski, T. Bioorg. Med. Chem. Lett. 2006, 16, 5088; (b) Hutchison, A. J.;
Krause, J. E. Annu. Rep. Med. Chem. 2004, 39, 139; (c) Proctor, L. M.; Woodruff, T.
M.; Taylor, S. M. Expert Opin. Ther. Patents 2006, 16, 445; (d) Sumichika, H.;
Sakata, K.; Sato, N.; Takeshita, S.; Ishibuchi, S.; Nakamura, M.; Kamahori, T.;
Ehara, S.; Itoh, K.; Ohtsuka, T.; Ohbora, T.; Mishina, T.; Komatsu, H.; Naka, Y. J.
Biol. Chem. 2002, 277, 29403; (e) Sumichika, H. Curr. Opin. Investig. Drugs 2004,
5, 505.
5. Barbay, J. K.; Gong, Y.; Buntinx, M.; Li, J.; Claes, C.; Hornby, P. J.; Van Lommen,
G.; Van Wauwe, J.; He, W. Bioorg. Med. Chem. Lett. 2008, 18, 2544.
6. For anilines containing other nucleophilic functional groups (e.g., CH2OH,
CO2H, and CH2NH2), no protections were necessary in the initial quick aniline
screening, as the desired products could be separated from undesired isomers.
13. The plasma level of 27, a potential des-methyl metabolite, was also found to be
minimal.