Discovery of small molecule human C5a receptor antagonists

Article abstract of DOI:10.1016/j.bmcl.2008.12.104

A novel series of small molecule C5a antagonists is reported. In particular, in vitro metabolic studies and solution based combinatorial synthesis are demonstrated as useful tools for the rapid identification of antagonists with low in vitro clearance. Me

Full text of DOI:10.1016/j.bmcl.2008.12.104

Bioorganic & Medicinal Chemistry Letters 19 (2009) 1143–1147
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of small molecule human C5a receptor antagonists
a
a
b
a
a
Hitesh J. Sanganee a, , Andrew Baxter , Simon Barber , Alastair J. H. Brown , Denise Grice , Fraser Hunt ,
*
Sarah King a, David Laughton a, Garry Pairaudeau c, Bob Thong a, Richard Weaver a, John Unitt a
a AstraZeneca R&D Charnwood, Medicinal Chemistry, Bakewell Road, Loughborough, LE11 5RH, UK
b AstraZeneca R&D Alderley Park, Mereside, Macclesfield, SK10 4TG, UK
c AstraZeneca R&D Mölndal, Pepparedsleden 1 431 83 Mölndal, Sweden
a r t i c l e i n f o
a b s t r a c t
Article history:
A novel series of small molecule C5a antagonists is reported. In particular, in vitro metabolic studies and
solution based combinatorial synthesis are demonstrated as useful tools for the rapid identification of
antagonists with low in vitro clearance. Members of this series specifically inhibited the binding of
125I-labeled C5a to human recombinant C5a receptor (C5aR). In functional cell assays these compounds
displayed surmountable antagonism against C5a and did not demonstrate any detectable agonist activity.
Ó 2009 Elsevier Ltd. All rights reserved.
Received 10 October 2008
Revised 23 December 2008
Accepted 24 December 2008
Available online 31 December 2008
Keywords:
C5a antagonist
Inverse agonists
Small molecule
Ugi multicomponent
The complement system is comprised of a cascade of interre-
lated proteases that are activated in response to immunoglobins
binding to a foreign antigen. Activation of the complement systems
leads to a stepwise hierarchy of proteolytic cleavage events ulti-
mately leading to the release of bioactive fragments (C3a, C4a
and C5a) known as anaphylatoxins. C5a is recognized as a promi-
nent mediator of inflammation through recruitment of inflamma-
tory cells to the site of injury or infection.
The functional effects of C5a are mediated by specific interac-
tion with the G-protein coupled receptor C5aR, expressed on a
variety of cells including mast cells and neutrophils.1 Consistent
with its proinflammatory properties, C5a has been implicated in
the pathology of a wide range of disorders, including rheumatoid
arthritis, asthma, respiratory distress syndrome, sepsis and inflam-
matory bowel disease.2 Significant progress has been made in re-
cent years on the understanding of the complement system and
several non-peptidic C5a antagonists have been recently reported
in the literature (Fig. 1).3
The most promising of these was the bisamide series, as exempli-
fied by compounds 1 and 2.
The bisamide series was identified following an HTS using a
whole cell binding Fluorescent Microvolume Assay Technology
(FMATTM) assay,6,10 which measured the inhibition of Alexa-647 la-
beled C5a binding to human recombinant C5aR (hrC5aR) expressed
in HEK-293 cells with G-protein Ga
16.10 The compounds were
subsequently confirmed as potent functional C5aR antagonists by
determining their inhibition of C5a-stimulated intracellular cal-
cium mobilization in the hrC5aR HEK cells (see Table 1). Activity
against the native receptor was also confirmed by measuring com-
pound mediated inhibition of C5a-stimulated calcium mobilization
in human neutrophils.16 For example, Compound (2) had a pA2 va-
lue of 7.4 in this neutrophil assay. Activity against the related C5a
binding receptor C5L2 was determined using an FMAT binding as-
say10 and the series was found to be selective. For example com-
pound (2) tested up to 30 lM displayed no inhibition of Alexa-
647 labeled C5a binding to C5L2. In additional selectivity profiling,
Interest in this area is driven by the potential therapeutic appli-
cation of a C5aR antagonist in the treatment of inflammatory dis-
eases such as arthritis, asthma and COPD. We report in this
paper the discovery of a novel series of small molecule C5aR
antagonists.
High throughput screening of the AstraZeneca compound col-
lection followed by a hit to lead (HtL) program identified a number
of chemical series with clear structure activity relationships (SAR).
compound (2) was inactive at GPCRs FPR, CXCR2, CXCR4 and CCR3,
when tested up to 10 lM (data not shown). Species selectivity of
the bisamide series was examined in recombinant receptor cell cal-
cium mobilization assays and was demonstrated to have apparent
selectivity for the human C5aR (see Table 2), similar to previous re-
ports on small molecule C5aR antagonists.7
In order to further characterize the functional antagonism of the
bisamide series we measured the ability of exemplars from this
series to inhibit C5a stimulated GTPcS binding in HEK cell mem-
branes expressing the hrC5aR.10 Concentration response curves
to C5a were performed following pre-treatment of membranes
* Corresponding author. Tel.: +44 01509 644167.
E-mail address: Hitesh.sanganee@astrazeneca.com (H.J. Sanganee).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.12.104

1144
H. J. Sanganee et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1143–1147
F
N
R''
N
O
N
N
N
N
O
O
R'
Figure 1. A selection of literature C5a receptor antagonists.
says have demonstrated a role for Na+ in stabilizing the inactive
form of the receptor.9 Consistent with these studies we found that
by removing Na+ from the buffer we were able to detect significant
levels of constitutive activity of the hrC5aR.
Table 1
Table of bisamides
O
O
N
O
O
N
O
O
O
O
Analysis of 2 in this system demonstrated a clear reduction in
basal [35S]GTP
c
S binding by 63% with a pEC50 of 6.91 (Fig. 3). In
H
N
O
H
N
contrast rhC5a, a full agonist, increased basal [35S]GTP
cS binding
by 50% with a pEC50 of 8.7. These data are consistent with 2 having
inverse agonist activity. Furthermore, while whole cell binding and
functional (intracellular calcium mobilization) potencies were
broadly equivalent for this series, the bisamide compounds dis-
played only partial inhibition of hrC5a in hrC5aR membrane-bind-
ing assays,10 with relatively low potencies in comparison to whole
cell binding. Full inhibition and increased potencies were subse-
quently obtained when the membrane-binding assay was supple-
F
F
(2)
(1)
Compound
MW
Log D (pH 7.4)
C5aa R pIC50
Human micsb
Rat hepsc
1
2
467
477
3.1
3.9
6.9
7.6
>150
>150
>150
>150
a
Human C5a receptor HEK calcium mobilization antagonist assay. Values are
mented with 10
consistent with the two-state model of GPCR activation whereby
the inclusion of GTP S in the system would be expected to favor
lM GTPcS. (see Table 3). These data are
means of at least two experiments.
b
Human microsome metabolism intrinsic clearance (
Rat Sprague–Dawley hepatocyte metabolism (
l
L/min/mg).4
l
L/min/106 cells).5
c
c
the dissociation of G-protein from the receptor complex (i.e., form
the inactive, R state) thereby increasing the apparent affinity of
rhC5a receptor for inverse agonists.
Table 2
Compound species selectivity
The bisamide compounds 7 were obtained using a choice of two
synthetic routes, dependent upon the availability of commercial
starting reagents. Scheme 1 involved a one-pot synthesis via an
Ugi multicomponent reaction, where a mixture of a carboxylic acid
3, aniline 4, isonitrile 5 and aldehyde 6 were stirred together in
methanol.11 Where the isonitrile component was unavailable
(Scheme 2) the amide bond was constructed using a conventional
coupling reaction with an appropriate amine and the carboxylic
acid intermediate 8. The acid 812 was in turn prepared from a 4
component Ugi reaction using the convertible isonitrile 5.13
Although a number of potent bisamides were identified, issues
relating to their metabolic instability had to be addressed. It is
likely that these DMPK issues arise from high lipophilicity, leading
to high measured intrinsic clearances in rat hepatocytes and hu-
man liver microsomes. The main objective of this HtL programme
Compound Human rC5aa
pIC50
R
Mouse rC5ab
pIC50
R
Rat rC5ac
pIC50
R
Dog rC5ad
pIC50
R
1
2
6.9
7.5
NA
NA
NA
NA
NA
NA
Inhibition of C5a-stimulated intracellular calcium mobilization in HEK cells trans-
fected with recombinant bmouse, crat and ddog C5a receptors. Values are means of
at least two experiments. NA: not active at 30 lM.
with increasing concentrations of the antagonists (see Fig. 2 for
example). The results demonstrated that all C5a antagonists tested
inhibited C5a in an apparent surmountable fashion.
The two-state model of GPCR activation assumes that the recep-
tors exist in an equilibrium between inactive, R and active, R*
states.8 Previous studies using functional [35S]GTP
c
S binding as-
180
160
140
120
100
80
hrC5a
Compound 2
120
Control
10-5M
100
10-6M
80
10-7M
60
40
20
0
60
40
20
0
-11
-10
-9
-8
-7
-6
-5
-11
-10
-9
-8
-7
-6
-5
Log10 [Ligand] (M)
Log10 [hrC5a] (M)
Figure 3. Compound 2 and rhC5a effects on basal [35S]GTP
cS binding activity in
Figure 2. Inhibition of rhC5a-stimulated [35S]GTP
c
S
binding by compound 2.
rhC5a-HEK cell membranes. Values are the means of at least two independent
Values are the means of at least two independent experiments.
experiments.

H. J. Sanganee et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1143–1147
1145
Table 3
Table 5
Compound biological activities
SAR of compounds marked A and B
Compound C5aa R whole
cell binding
C5ab
Ca2+
pIC50
R
C5ac
membrane-
binding pIC50
R
C5ad R membrane-
binding + GTP
cS
pIC50
pIC50
O
O
O
R
1
2
7.1
7.4
6.9
7.5
5.5
6.1
6.6
7.5
N
N
H
F
Values are means of at least two experiments.
O
a
Human C5a receptor HEK whole cell [Alexa647]C5a competition binding assay.
Human C5a receptor HEK calcium mobilization antagonist assay.
b
c
Human C5a receptor HEK membrane 125I-labeled C5a competition binding
Compound
R
C5a R
pIC50
Human
mics
Rat
Log D (pH
assay.
d
heps
7.4)
Human C5a receptor HEK membrane 125I-labeled C5a competition binding
assay in presence of 10
l
M GTP
c
S.10
13 (A)
14
7.2
<7
104
—
>150
—
2.55
OH
OH
O
ArCO2H
+
Ar'NH2
—
—
Ar'
4
Ar
N
3
(a)
O
Ar''
15 (B)
6.0
135
33
N+
N
Ar''
O
NH
R
C
R
5
6
7
Scheme 1. (a) Methanol, 40 °C, 4 h.
O
ArCO2H
+
Ar'NH2
O
Ar'
Ar
N
4
3
(a-c)
O
O
O
Ar''
N+
Ar''
O
C
OH
8
(d)
6
O
5
Ar'
Ar
N
O
Ar''
NH
7
R
Scheme 2. (a) Methanol, 40 °C, 4 h (b) potassium t-butoxide, THF, (c) Lithium
hydroxide/water, (d) HATUTM, Hunigs base, RNH2.
was to address the key issues of maintaining potency while reduc-
ing lipophilicity and thereby hopefully increasing metabolic stabil-
ity. Metabolite identification was also used to find the key sites of
Figure 4. Plot of % Maximal inhibition of C5aR in the calcium mobilization assay
using 10
l
M compound versus calculated ACD log D (pH 7.4).
Table 4
Compound biological activities
R3
O
R4
R1
N
O
R2
Compound
R1
R2
R3
R4
C5a R Ca2+ pIC50
NH
9
Ph
CH3
o-Fluoro Phenyl
<5
NH
10
Ph
2,4 dimethoxy phenyl
CH3
<5
NH
OH
11
12
CH3
Ph
2,4 dimethoxy phenyl
2,4 dimethoxy phenyl
o-Fluoro Phenyl
o-Fluoro Phenyl
6.1
<5

1146
H. J. Sanganee et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1143–1147
metabolism of the bisamides. This approach is in line with our lead
criteria outlined in previous HtL publications.14
was demonstrated that the phenyl amide could be replaced by a
series of substituted heterocycles. The 2- and 3-furan analogs with
a substituent next to the point of attachment (16,17) were among
the most potent but other five membered heterocycles (thienyl,
thiazolyl, oxazolyl and isoxazolyl) also had good activity. These
compounds had slightly reduced lipophilicity but were still rapidly
metabolised in vitro. Metabolite identification on 16 showed the
major site of metabolism was oxidation of the methyl substituent
on the furan ring. Replacement with other groups either gave good
potency and poor metabolic stability (methoxymethyl 18, chloro
19, 20) or good metabolic stability and poor potency (dimethyl-
aminomethyl 21 hydroxymethyl 22). The diasteromeric mixture
of the methoxymethylfuran 18 was separated by HPLC into the
two single isomers. All activity resided in one diastereomer of un-
known absolute configuration but in vitro clearance was similar for
both isomers.
Initial efforts were focused on the SAR of the three-aryl rings
(R1–R3) (Table 4), replacing each in turn with a methyl group.
Replacement at centers R2 or R3 with methyl was not tolerated
and led to complete loss of activity (9, 10). However replacement
of R1 (11) did retain some potency (pIC50 6.1). Replacement of
the cyclopentylamide by its parent carboxylic acid (12) also re-
sulted in complete loss of activity.
Metabolite analysis highlighted the cyclopentyl motif as the
main area of metabolic liability. To address this finding a focused
library of cyclopentyl replacements was synthesised. The data from
this library is illustrated (Fig. 4) as plot of % maximal inhibition of
the C5a receptor in the recombinant HEK calcium mobilization as-
say10 against calculated ACDLog D7.4 for compounds spot tested
15
at 10 lM. Two key compounds emerged as having lower log D
and good inhibition (marked as A and B Fig. 4) and these com-
pounds were chosen for further biological analysis (Table 5, 13,
15). The diastereomer 13 (A) prepared from the chiral alcohol
was found to have a pIC50 7.2 and lower measured log D (relative
to cyclopentyl amide) although still had a high intrinsic clearance.
Furthermore, the diastereomer 14 prepared from the chiral alcohol
of opposite stereochemistry was lower potency.
Maintaining the alcohol side chain but addressing the log D fur-
ther by varying the aryl amide, proved rewarding. Key compounds
synthesized following this strategy are highlighted in Table 6. It
We discovered a series of C5aR antagonists through an FMAT-
based C5a binding HTS and characterized them as human C5aR
selective inverse agonists. SAR at the four variable positions of
the bisamides 1 and 2 was determined rapidly using library syn-
thesis. A large number of compounds with ‘lead-like’ potency were
prepared but these all had poor metabolic stability. Metabolic sta-
bility is possible in this series as shown by compounds 20 and 21
but these had poor C5a potency. Other factors that were taken into
account when considering this series as potential leads were the
high molecular weight and low solubility, as well as lack of species
Table 6
SAR table for aryl amide replacements
R
O
O
O
H
N
N
F
OH
O
Compound
R
C5a R Ca2+ pIC50
7.2
Human mics
104
Rat heps
> 150
Log D (pH 7.4)
2.6
ACD Log D (pH 7.4)a
1.9
13
16
17
7.7
7.4
126
134
87
75
2.0
—
1.5
1.5
O
O
MeO
Cl
18
19
7.0
6.7
76
35
59
2.0
2.2
0.8
1.2
O
148
O
Cl
O
20
21
6.5
5.5
48
12
12
33
<3
8
2.4
—
1.3
0.8
Me2N
HO
O
22
NA
1.4
À0.2
O
a
ACD log D7.4 calculations15 are approximately 1 log unit lower than measured.

H. J. Sanganee et al. / Bioorg. Med. Chem. Lett. 19 (2009) 1143–1147
1147
H.; Naka, Y. J. Biol. Chem. 2002, 277, 49403; (b) Waters, S. M.; Brodbeck, R. M.;
Steflik, J.; Yu, J.; Baltazar, C.; Peck, A. E.; Severance, D.; Zhang, L. Y.; Currie, K.;
Chenard, B. L.; Hutchison, A. J.; Maynard, G.; Krause, J. E. J. Biol. Chem. 2005, 280,
40617.
cross over. On the evidence available this series was not progressed
into Lead Optimisation.
8. Leff, P. Trends. Pharmacol. Sci. 1995, 16, 89.
Acknowledgments
9. Wenzel-Seifert, K.; Hurt, C. M.; Seifert, R. J. Biol. Chem. 1998, 273, 24181.
10. See online publication supplementary material for detailed methods.
11. Rikimaru, K.; Yanagisawa, A.; Kan, T.; Fukuyama, T. Synlett 2004, 1, 41.
We thank Andrew Wright and Michael Bernstein and their team
for analytical support, and John Steele for helpful discussions.
12. Synthesis of compound (12)
(in Table 4). To
a solution of benzoic acid
(0.6 mmol), Carbonic acid, 2-isocyano-2-methylpropyl methyl ester
(0.8 mmol) and 2,4-dimethoxyaniline (0.6 mmol) in methanol (10 ml) was
added 2-fluorobenzaldehyde (1.0 mmol) and the mixture stirred at 40 deg for
4 h. The mixture was evaporated to dryness and the residue purified by silica
chromatography (1:1 ethyl acetate/isohexane). The above intermediate
(0.372 mmol) was dissolved in tetrahydrofuran (10 ml) and to this solution
was added potassium t-butoxide 1.0 M THF (0.5 mmol). Upon completion of
reaction the mixture was acidified with glacial acetic acid and evaporated to
dryness. The residue was taken up into a mixture of dioxan/water (1:1) and to
this solution was added lithium hydroxide (2 equivalents) and stirred
overnight. The mixture was acidified and mixture was evaporated to dryness
and purified (HPLC, SymmetryÒ, 0.1% aqueous ammonium acetate:acetonitrile,
gradient elution 75:25 to 5–95 over 15 min) to afford 12 0.073 g (46%) as a
colourless solid. MS (APCI) 410 (M + H, 100%). 1H NMR (90 °C) (399.826 MHz,
DMSO) d 7.28–7.10 (m, 6H), 7.05–6.85 (m, 4H), 6.41 (s, 1H), 6.27 (d, J = 7.4 Hz,
1H), 6.02 (s, 1H), 3.59 (s, 3H), 3.29 (s, 3H).
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2008.12.104.
References and notes
1. Gerard, N. P.; Gerard, C. Nature 1991, 349, 614.
2. Guo, R. F.; Ward, P. A. Annu. Rev. Immunol. 2005, 23, 821.
3. For example, see: Neurogen Corp., WO Patent 0251414, 2002.; Neurogen Corp.,
WO Patent 2003082826, 2003.; Mitsubishi Pharma Co., WO 2006128670, 2006.;
Jerini A. G. WO Patent 06128670, 2006.; Taylor, S. M.; Fairlie, D. P. Expert Opin.
Ther. Pat. 2000, 10(4), 449; Hutchison, A. J.; Krause, J. E. Annu. Rep. Med. Chem.
2004, 39, 139; Proctor, L. M.; Woodruff, T. M.; Taylor, S. M. Expert Opin. Ther. Pat.
2006, 16, 445; Kent Barbay, J.; Gong, Y.; Butinix, M.; Li, J.; Concha, C.; Hornby, P. J.;
Van Lommen, G.; Van Wauwe, J.; He, W. Bioorg. Med. Chem. Lett. 2008, 18, 2544;
Gong, Y.; Barbay, J. K.; Buntinx, M.; Li, J.; Van Wauwe, J.; Claes, C.; Van Lommen,
G.; Hornby, P. J.; He, W. Bioorg. Med. Chem. Lett. 2008, 18, 3852.
4. McGinnity, D. F.; Parker, A. J.; Soars, M.; Riley, R. J. Drug Metab. Dispos. 2000, 24,
1327.
5. Soars, M.; Grime, K.; Sproston, J. L.; Webborn, P. J. H.; Riley, R. J. Drug Metab.
Dispos. 2007, 35, 859.
6. Miraglia, S.; Swartzman, E. E.; Mellentin-Michelotti, J.; Evangelista, L.; Smith,
C.; Gunawan, I.; Lohman, K.; Goldberg, E. M.; Manian, B.; Yuan, P. M. J. Biomol.
Screen. 1999, 4, 193.
13. Lindhorst, T.; Bock, H.; Ugi, I. Tetrahedron 1999, 55, 7411.
14. (a) Baxter, A.; Cooper, A.; Kinchin, E.; Moakes, K.; Unitt, J.; Wallace, A. Bioorg.
Med. Chem. Lett. 2006, 16, 960; (b) Baxter, A.; Brough, S.; Cooper, A.;
Floettmann, E.; Foster, S.; Harding, C.; Kettle, J.; McInally, T.; Martin, C.;
Mobbs, M.; Needham, M.; Newham, P.; Paine, S.; St-Gallay, S.; Salter, S.; Unitt,
J.; Xue, Y. Bioorg. Med. Chem. Lett. 2004, 14, 2817; (c) Baxter, A.; Bent, J.; Bowers,
K.; Braddock, M.; Brough, S.; Fagura, M.; Lawson, M.; McInally, T.; Mortimore,
M.; Robertson, M.; Weaver, R.; Webborn, P. Bioorg. Med. Chem. Lett. 2003, 13,
4047; (d) Baxter, A.; Bennion, C.; Bent, J.; Boden, K.; Brough, S.; Cooper, A.;
Kinchin, E.; Kindon, N.; McInally, T.; Mortimore, M.; Roberts, B.; Unitt, J. Bioorg.
Med. Chem. Lett. 2003, 13, 2625.
15. Log D was calculated using the ACD/Log D (Version 10) software available from
ACD/labs.
16. Baly, D.; Gibson, U.; Allison, D.; DeForge, L. Methods Enzymol. 1997, 287, 70.
7. (a) 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,