Discovery of a novel series of CXCR3 antagonists

Article abstract of DOI:10.1016/j.bmcl.2010.04.113

The discovery of a novel series of CXCR3 antagonists is described. Starting from an HTS positive, iterative optimization gave potent compounds (IC ₅₀ 15 nM in a chemotaxis assay). The strategy employed to improve the metabolic stability of these derivatives is described.

Full text of DOI:10.1016/j.bmcl.2010.04.113

Bioorganic & Medicinal Chemistry Letters 20 (2010) 3614–3617
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of a novel series of CXCR3 antagonists
a
a
a
a
Stefano Crosignani a, , Marc Missotten , Christophe Cleva , Ruggero Dondi , Yann Ratinaud ,
Yves Humbert a, Ashis Baran Mandal b, Agnès Bombrun a, Christine Power a,
André Chollet a, Amanda Proudfoot a
*
a Merck Serono SA, 9 Chem. des Mines, 1202 Geneva, Switzerland
b Syngene International Ltd, Biocon Park, Plot No. 2&3, Bommasandra IV Phase, Bommasandra—Jigani Link Road, Bangalore 560 099, India
a r t i c l e i n f o
a b s t r a c t
Article history:
The discovery of a novel series of CXCR3 antagonists is described. Starting from an HTS positive, iterative
optimization gave potent compounds (IC50 15 nM in a chemotaxis assay). The strategy employed to
improve the metabolic stability of these derivatives is described.
Received 20 March 2010
Revised 23 April 2010
Accepted 26 April 2010
Available online 28 April 2010
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
CXCR3
Chemokines
GPCR antagonists
Rheumatoid arthritis
Multiple sclerosis
Chemokines (chemoattractant cytokines) are small (8–10 KDa)
proteins responsible for directing leukocytes towards sites of
inflammation.1 They exert their signaling by interaction with che-
mokine receptors, which are members of the GPCR family. A num-
ber of chemokines/chemokine receptors have been associated with
various autoimmune and inflammatory diseases.1 In particular, the
receptor CXCR3, which is mainly expressed on CD4+ Th1 T cells,
and its ligands CXCL9/Mig, CXCL10/IP-10, and CXCL11/I-TAC have
been linked to diseases such as rheumatoid arthritis (RA),2 multi-
ple sclerosis (MS)3 as well as allograft rejection,4 and in general
with pathologies that are associated with a major involvement of
reduction in cAMP levels, as well as an increase in Ca2+ mobiliza-
tion and actin polymerization, finally leading to cytoskeletal rear-
rangement and chemotaxis). One positive was found with sub-
micromolar potency (1, Fig. 1). Characterization of compound 1
(Table 1) confirmed the potency in different assays, and the com-
pound was found to be selective against other GPCRs. However,
the in vitro profiling of the compound showed several severe liabil-
ities: low solubility, low permeability in a Caco-2 assay, instability
in acidic media and in particular high clearance in the presence of
human and rat microsomes. Some of these liabilities could be
linked to the presence of the acyl hydrazone moiety, which we
decided to replace. A search was run for all structures present in
our chemical collection or at commercial suppliers possessing var-
ious key features found in 1 but lacking the acylhydrazone moiety,
and these compounds were screened. We were then delighted to
find that the acylhydrazone moiety could be simply replaced with
an amide group as in compound 2, with a modest loss of potency.
Th1 cells. For example, CXCR3 expression is increased in CD4+
T
cells of MS patients, and CXCR3+ T cells accumulate in active le-
sions. Moreover the levels of CXCL10 are increased in the CSF of
MS patients, and these levels correlate with the severity of the dis-
ease. Likewise, studies conducted on RA patients have showed that
the majority of the infiltrating cells in synovial fluid express CXCR3
and the levels of IP-10 and Mig are elevated relative to normal indi-
viduals. Recently, the efficacy of the anti-IP10 antibody MDX1100
reported in a Phase 2 clinical trial for RA5 reinforced the crucial role
of the CXCL10-CXCR3 axis in this disease, and the potential for
treatment that a small molecule CXCR3 antagonist could have.6
With the purpose to discover new CXCR3 antagonists, a high-
throughput screening was run on 90,000 compounds, using a cAMP
assay (CXCR3 being a Gai-coupled receptor, its activation leads to a
* Corresponding author. Tel.: +41 224149618; fax: +41 227312179.
E-mail address: stefano.crosignani@merckserono.net (S. Crosignani).
Figure 1. HTS positive 1 and Hit 2.
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.04.113

S. Crosignani et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3614–3617
Correlation pIC50 chemotaxis vs. binding
3615
Table 1
Characterization of compound 1
y = 0.9684x
2 = 0.755
R
Ki (IP10/ITAC, nM)
348/211
99/128
538/1810
92/219
0.1
8
7.5
7
GTP
L1.2 Chemotaxis (IP10/ITAC, nM)
Microsome Clint (human/rat) ( L/min/mg)
Caco-2 permeability (Papp, 10À6 cm/s)
cS (IP10/ITAC, nM)
l
Solubility (PBS,
l
M)a
11.9
4.8
Log P
Stability at pH 1
MW
Unstable
589
6.5
6
a
Nephelometric kinetic solubility in PBS (1% DMSO).
5.5
5
Having found a more suitable starting point, we embarked in an
optimization program. The nature of the scaffold made the chem-
istry strategy quite straightforward: depending on the point of the
molecule that we wanted to diversify, one of the routes depicted in
Scheme 1 could be applied: the sulfonamide 3 or 4 could be built
first, followed by an alkylation in basic media, or the secondary
amine 5 could be formed first, followed by reaction with a sulfonyl
chloride, enabling the introduction of the diversity at a late stage of
the synthesis. From carboxylic esters 6a (X = COOMe), via the
intermediate of the corresponding carboxylic acids 7, amides and
acylsulfonamides could be prepared, whereas starting from the
corresponding nitriles 6b (X = CN) a series of tetrazole compounds
could be obtained.
To establish the SAR of this series, it was decided to use an assay
that is physiologically relevant: a chemotaxis assay using CXCR3-
overexpressing L1.2 cells and CXCL10 as chemoattractant was
developed as previously described,7 and was used throughout the
project. The assay had a sufficient throughput to conduct the
SAR, while being a direct measure of the ability of the compounds
to inhibit the main pathophysiological result of CXCR3 activation,
with the caveat that an overexpressing cell line was used instead
of human T cells. The correlation of this assay with a more tradi-
tional binding assay was found to be very good (Fig. 2), giving at
least an indirect proof that the effect of our compounds was on tar-
get. As negative control, a chemotaxis experiment on selected com-
pounds was run using a different stimulus on appropriate cells
(using CXCL13 on CXCR5-overexpressing L1.2 cells) and no effect
on the resulting chemotaxis was observed (up to a concentration
5
5.5
6
6.5
7
7.5
8
pIC50 SPA binding
Figure 2. Correlation between activity in the L1.2 chemotaxis and binding assays.
lene (13) and 1-phenyl-cycloprop-1-yl group groups (see 14 vs 11–
12, 26 vs 20–21, 27 vs 23) could be discovered.
On the other hand, substitution on the sulfonamide ring bearing
R1 proved to be limited: any kind of substitution in the ortho and
meta positions of this aromatic group, as well as replacing the
sulfonamide with an amide group, led to a severe decrease in
potency (data not shown). The presence of the para-chlorine atom
was important for potency, and good results could also be obtained
substituting this chlorine atom with an alkoxy (16–18, 22 but not
19) or a nitrile group (23, 27). Finally, for the second benzyl substi-
tuent on the sulfonamide nitrogen R2 the SAR was found again to
be more tolerant, as long as an aromatic or heteroaromatic group
was present with an unsubstituted methylene linker. In particular
2-pyridyl (see 20 vs 11, 21 vs 12) was the preferred substituents.
On the other hand, aliphatic or cycloaliphatic groups were not tol-
erated (data not shown).
The most potent compound identified in this round of optimiza-
tion was 27, with an IC50 of 13 nM in the chemotaxis assay. How-
ever, it soon appeared that all these compounds shared a very
serious liability, in that their stability in human and rat micro-
somes was very limited. The first strategy that was attempted to
correct this issue was to block the most likely sites of metabolism.
Some evidence from the microsome stability experiments allowed
of 10 lM).
A quick optimization of amide 2 was performed on three diver-
sity points (selected examples in Table 2). While some diversity
was allowed for the amide N-substituent R3, a preference for the
more lipophilic groups and in particular for the thiophene methy-
us to identify the methylene
site of metabolism. However, derivatives 14, 26, and 27, in which
a to the amide bond as the principal
Scheme 1. Reagents and conditions: (a) or (b) R2CH2Cl, K2CO3, KI, DMF, 100 °C; (c) ArSO2Cl, TEA, DCM; (d) NaOH, H2O/THF; (e) Bu2SnO, TMSN3, toluene, 90 °C; (f) R3NH2, EDC,
HOBt; (g) R4SO2NH2, EDC, DMAP.

3616
S. Crosignani et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3614–3617
Figure 3. Compounds for Table 2.
Table 2
SAR of selected amides
Figure 4. Compounds for Table 3.
Compounds
R1
R2
R3
L1.2
Microsome
Chemotaxis
IC50 (nM)
Clint (human/rat)
(lL/min/mg)
a
2
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Cl
Cl
Cl
Cl
Cl
Cl
OMe
OMe
OEt
OCF3
Cl
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ab
B
D
E
G
I
B
C
B
B
B
D
B
B
B
B
G
G
2275
1985
1295
192
238
2450
697
347
673
5580
200
85
1152
360
510
940
25
490/1701
—
—
—
211/507
—
—
—
Table 3
SAR of selected carboxylic acid, oxadiazolone, tetrazole and acylsulfonamide
compounds
Compounds R1
R2
X
Y
L1.2
Microsome
Chemotaxis Clint (human/rat)
a
IC50 (nM)
(lL/min/mg)
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
Cl
Ph
OMe Ph
Jb
J
J
J
J
K
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
L
M1
M2
M3
M4
H
H
F
F
F
2940
6440
1132
513
1080
695
496
470
1295
882
233
540
885
1460
2560
911
254
440
158
192
157
236
183
339
216
207
<10/20
—
—
ND/<10
—
—
<10/25
—
—
—
—
—
—
—
—
Ph
Ph
—
—
Cl
Cl
CN
Cl
Cl
Cl
Cl
Cl
CN
OEt
Cl
Cl
Cl
Cl
Cl
Cl
OEt
CN
Cl
CN
Cl
Cl
Cl
Cl
2-Py
2-F-Ph
2-Py
2-Py
2-Py
2-F-Ph
3-Cl-Ph
4-Cl-Ph
2-Py
2-Py
2-Py
3-Py
2-Py
2-F-Ph
3-Cl-Ph
2-Py
2-Py
265/538
583/1191
—
—
—
—
Cl
OMe
CN
Cl
Cl
Cl
H
H
H
H
H
H
H
H
H
H
H
H
F
F
F
F
H
H
H
H
H
257/1256
435/474
CN
13
a
Average of two experiments. SD were generally within 50% of average.
See Figure 3.
2-Py
b
4-(OMe)Ph
3-(OMe)Ph
4-F-Ph
P
Q
2-Py
—
—
—
this position was blocked, proved to still be unstable, with metab-
olism in this case happening on the left part of the molecule.
Another strategy was identified when we investigated com-
pound 28 (Fig. 4 and Table 3). This intermediate in the synthesis
of the corresponding amides was found to still be active on the
receptor, albeit with reduced potency. More importantly, this com-
pound was found to be stable in the presence of both human and
rat microsomes. Since several metabolically labile positions still
exist in this molecule, a possible explanation of this improvement
is linked to the decrease in Log D of the molecule, which could re-
duce the affinity for the Cytochrome P450 isoforms responsible for
the metabolism of the amide derivatives.
2-Py
109/70
ND/17c
<10/71
ND/67
—
—
10/23
27/122
2-F-Ph
2-F-Ph
2-F-Ph
2-Py
2-Py
2-Py
2-Py
a
b
c
Average of two experiments. SD were generally within 50% of average.
See Figure 3.
Clint in mouse microsomes.
Several other carboxylic acid derivatives were tested, but only
limited improvements in potency were obtained, the best com-
pounds being those with a fluorine atom in ortho to the carboxylic
acid (e.g., 31).
In order to restore sufficient levels of potency on the receptor, a
number of possible bioisosters of the carboxylic acid were also ex-
plored. Both tetrazole (34–49) and oxadiazolone (33) were found
to be possible replacements, with a moderate gain in potency.
SAR around these structures allowed the identification of several
compounds with potency in the 100–200 nM range in the chemo-
taxis assay. The SAR was found to follow the same patterns recog-
nized for the amide subseries, with a preference for a chlorine,
alkoxy or nitrile substituent as R1. For R2, the 2-pyridyl was again
found as a preferred substituent, but also the 2-fluorophenyl group
was found to be equally potent. Although it was not possible to re-
cover the low nanomolar IC50 obtained with the best amide com-
pound, the tetrazoles were found to be generally more stable in
microsomes than the corresponding amides, in keeping with the
theory that the affinity of the compounds for the CYP isoform(s)
is generally governed by their lipophilicity (with the possible
exception of 46, for which an alternative metabolic pathway
involving O-dealkylation can be envisioned and which could be

S. Crosignani et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3614–3617
3617
due to a different CYP isoform). An additional advantage of the tet-
Supplementary data
razole compounds compared to the corresponding amides was a
notable gain in aqueous solubility at pH 7.4: while all amides
Supplementary data (typical conditions for the synthesis of
compounds, as well as experimental details on the in vitro assays.
Full experimental conditions and characterization of all com-
pounds can be found in patent WO2009124962) associated with
this article can be found, in the online version, at doi:10.1016/
j.bmcl.2010.04.113.
showed low solubility (<10
lM), all tested tetrazole compounds
had solubilities greater than 100
l
M.
Among the compounds profiled in PK experiments in mice com-
pound 47 was identified as having the best profile, with a medium
Clearance of 1.7 L/h/kg (ca. 30% of mouse liver blood flow) and a Vss
of 1.0 L/kg after iv dosing (0.2 mg/kg) and with a bioavailability of
83% and a half-life of 1 h after po dosing (10 mg/kg).
References and notes
In a further effort to increase the potency while keeping good
metabolic stability, we turned our attention to acylsulfonamides
50–53, which would combine the presence of a negatively charged
group at physiological pH (to increase metabolic stability and sol-
ubility) with the possibility to extend the molecule to occupy a
similar region of space as occupied by the amide compounds (to
try to restore low nanomolar potency). The compounds prepared
showed potency on the receptor, but without a significant increase
compared to the tetrazole compounds.
In summary, we have discovered a new series of CXCR3 antag-
onists. The first round of optimization yielded a potent compound
(27, IC50 13 nM), which was however unstable in microsomes.
Lowering the Log D of the molecules by incorporating a negatively
charged group (carboxylic acid, tetrazole or acylsulfonamide) led
to active compounds with an improved stability versus oxidative
metabolism. The best compound identified, 47, while still having
an insufficient potency and pharmacokinetic profile for using in a
1. Viola, A.; Luster, A. D. Annu. Rev. Pharmacol. Toxicol. 2008, 48, 171.
2. (a) Patel, D. D.; Zachariah, J. P.; Whichard, L. P. Clin. Immunol. 2001, 98, 39; (b)
Qin, S.; Rottman, J. B.; Myers, P.; Kassam, N.; Wienblatt, M.; Loetscher, M.; Koch,
A. E.; Moser, B.; Mackay, C. R. J. Clin. Invest. 1998, 101, 746.
3. (a) Sorensen, T. L.; Tani, M.; Jensen, J.; Pierce, V.; Lucchinetti, C.; Folcik, V. A.; Qin,
S.; Rottman, J.; Sellebjerg, F.; Strieter, R. M.; Frederiksen, J. L.; Ransohoff, R. M. J.
Clin. Invest. 1999, 103, 807; (b) Sindern, E.; Patzhold, T.; Ossege, L. M.; Gisevius,
A.; Malin, J.-P. J. Neuroimmunol. 2002, 131, 186; (c) Sorensen, T. L.; Trebst, C.;
Kivisakk, P.; Klaege, K. L.; Majmudar, A.; Ravid, R.; Lassmann, H.; Olsen, O. B.;
Strieter, R. M.; Ransohoff, R. M.; Sellebjerg, F. J. Neuroimmunol. 2002, 127, 59; (d)
Balashov, K. E.; Rottman, J. B.; Weiner, H. L.; Hancock, W. W. Proc. Natl. Acad. Sci.
U.S.A. 1999, 96, 6873.
4. Belperio, J. A.; Keane, M. P.; Burdick, M. D.; Lynch, J. P., III; Zisman, D. A.; Xue, Y.
Y.; Li, K.; Ardehali, A.; Ross, D. J.; Strieter, R. M. J. Immunol. 2003, 171, 4844.
5. Medarex press release, May 14th 2009.
6. For a review on CXCR3 antagonists, see: Wijtmans, M.; Verzijl, D.; Leurs, R.; de
Esch, I. J. P.; Smit, M. J. ChemMedChem 2008, 3, 861.
7. Colvin, R. A.; Campanella, G. S. V.; Manice, L. A.; Luster, A. D. Mol. Cell. Biol. 2006, 26,
5838. Briefly compounds, in presence of CXCL10, were transferred on the bottom
chambers of a ChemotX system (8
lm pore size 96 well, Neuroprobe, Gaitherburg,
MD). CXCR3-overexpressing L1.2 cells diluted in RPMI 5% FCS were then loaded on
the top of the ChemotX membrane. The plates were incubated for 4 h at 37 °C, 5%
CO2.The reaction was stopped by removing the membranes and cell migration
was quantified using the CyQuant kit (Molecular Probes, Eugene, OR)..
clinical setting, is
optimization.
a suitable lead compound for further