A novel series of p38 MAP kinase inhibitors for the potential treatment of rheumatoid arthritis

Article abstract of DOI:10.1016/j.bmcl.2004.08.006

The discovery, rational analogue design, synthesis and SAR of a novel bisamide class of p38 MAP kinase inhibitor are reported. The activity in vitro is described for the series. The activity in vivo and pharmacokinetic properties are exemplified for the more potent analogues, such as 18. A novel p38 MAP kinase inhibitor structural class was discovered through selectivity screening. The rational analogue design, synthesis and structure-activity relationship of this series of bis-amide inhibitors is reported. The inhibition in vitro of human p38α enzyme activity and lipopolysaccharide-induced tumour necrosis factor-α release is described for the series. The activity in vivo and pharmacokinetic properties are exemplified for the more potent analogues.

Full text of DOI:10.1016/j.bmcl.2004.08.006

Bioorganic & Medicinal Chemistry Letters 14 (2004) 5383–5387
A novel series of p38 MAP kinase inhibitors for the
potential treatment of rheumatoid arthritis
Dearg S. Brown,* Andrew J. Belfield, George R. Brown, Douglas Campbell,
Alan Foubister, David J. Masters, Kurt G. Pike, Wendy L. Snelson and Stuart L. Wells
AstraZeneca, Alderley Park, Macclesfield, Cheshire SK10 4TG, UK
Received 18 May 2004; revised 30 July 2004; accepted 5 August 2004
Available online 28 August 2004
Abstract—A novel p38 MAP kinase inhibitor structural class was discovered through selectivity screening. The rational analogue
design, synthesis and structure–activity relationship of this series of bisamide inhibitors is reported. The inhibition in vitro of human
p38a enzyme activity and lipopolysaccharide-induced tumour necrosis factor-a release is described for the series. The activity in vivo
and pharmacokinetic properties are exemplified for the more potent analogues.
Ó 2004 Elsevier Ltd. All rights reserved.
Rheumatoid arthritis (RA) is a chronic disease causing
joint pain, swelling and stiffness. Progression of RA
leads to joint damage and serious disability, making
treatment an important objective. Drug treatment to
date has primarily focused on the use of nonsteroidal
anti-inflammatory drugs (NSAIDS)1,2 and disease-modi-
fying anti-rheumatic drugs (DMARDS) such as metho-
trexate,2–4 sulfasalazine,3 leflunomide,2,4,5 cortico-
steroids,2 and various combination protocols, 6 in spite
of these tending to be poorly tolerated in long-term clin-
ical use. More recently novel biological products that
modify pro-inflammatory cytokines have gained clinical
approval as DMARDS. These new biological treat-
ments act through the sequestration of either tumour
necrosis factor-a (TNF-a) (e.g. the neutralising antibody
infliximab7–9 and the soluble TNF receptor etaner-
cept8,9) or interleukin-1 (IL-1) (e.g. the receptor antago-
nist anakinra9). As these biological drugs have shown
that the lowering of pro-inflammatory cytokine levels
is a valid treatment for RA patients our aim was to iden-
tify small molecule drugs with a similar cytokine-inhibi-
ting profile. Modern small-molecule approaches to the
treatment of RA have targeted cell signalling systems,
in particular the inhibition of the mitogen-activated pro-
induction of cyclooxygenase-2 (COX-2) in lipopolysac-
charide (LPS) induced monocytes.11 The inducible
COX-2 enzyme is a key enzyme in the arachidonic acid
enzyme cascade that leads to the production of prosta-
glandins, which are associated with inflammation and
pain. Inhibition of p38 kinase is thus an attractive ap-
proach to the treatment of both pain and inflammation
in RA patients. Pyridylimidazoles such as the prototypic
SB203580 are the most studied class of p38 inhibi-
tors,12,13 but other structural series have been re-
ported.13,14 In seeking an alternative series of p38
inhibitors we elected not to pursue the extensively devel-
oped pyridylimidazole series but instead sought a more
novel starting point.
During the screening of the company compound library
at AstraZeneca the bisamide 1 was identified as a potent
inhibitor of the protein kinase c-raf (IC50 17nM).15 Sub-
sequent selectivity testing amongst 19 typical kinases in
vitro showed that 1 only inhibited the two isoforms p38a
and p38b (p38a IC50 210nM)16 and, in addition, it was
shown that compound 1 had no effect on the c-raf path-
way in cells.17 Thus compound 1 was a novel p38a kin-
ase inhibitor with considerable selectivity over other
protein kinases, with the exception of the closely related
protein kinase c-raf, and it had suitable lead-like physi-
cal properties (logD 2.7 at pH7.4, solubility 10lM in
pH7.4 buffer, human protein binding 99.7% bound).
Herein we describe novel p38a inhibitor analogues of
the bisamide 1 as orally active, small molecule cytokine
inhibitors for the potential treatment of RA.
tein (MAP) p38 kinase cell-signalling pathway.10
A
potential advantage of this approach over the biological
agents is that p38 kinase has been implicated in the
Keyword: p38 MAP kinase inhibitors.
*
Corresponding author. Tel.: +44 1625 516321; fax: +44 1625
510823; e-mail: dearg.brown@astrazeneca.com
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.08.006

5384
D. S. Brown et al. / Bioorg. Med. Chem. Lett. 14 (2004) 5383–5387
F
O-alkylation of 1 gave 2 with comparable potency versus
N
the kinase but neither had significant inhibition of TNF-a
production in PBMCs. Addition of a 2-methoxy (3) had
little effect but a 3-methoxy (4) resulted in modest cellu-
lar potency and the 3,4,5-trimethoxy analogue 5 had a
similar profile although these analogues were signifi-
cantly weaker p38a inhibitors than SB203580. Attempts
to improve inhibitory potency by varying the methyl
group at R1, 6–9, confirmed the requirement for a sub-
stituent (e.g. methyl or chloro) at that position. Simi-
larly, transposing the methyl group to the other side of
the central ring in 10 and amide N-methylation (11) both
resulted in loss of potency (p38a IC50 > 20lM). Chang-
ing 3,4-dimethoxy to 3,4-diethoxy (12) increased activity
slightly but it was found that a single para-substituent
larger than methoxy (13–17) was sufficient to provide
micromolar activity in the PBMC assay. However, none
of the compounds in Table 1 (1–17) were inhibitors in
human whole blood below 50lM. Although these early
simple alkoxy compounds gave little or no exposure in
rat pharmacokinetic studies, the inclusion of a 2-pyridyl-
methoxy 17 resulted in modest blood levels following
oral dosing and further work therefore focused on these
systems.
OH
N
O
H
N
H
N
S
N
H
O
O
N
Compounds were tested in vitro for inhibition of human
p38a enzyme activity and for the inhibition of TNF-a
production in LPS-stimulated human peripheral blood
mononuclear cells (PBMCs) and human whole blood
using the literature p38 inhibitor, SB203580, as a posi-
tive control.18 In this study analogue synthesis was
restricted to alkoxy-substitution in place of the hydroxyl
in 1 in order to avoid the potential for phenolic b-glucuro-
nidation in vivo, should new analogues later proceed to
animal tests.
The bisamides were prepared as exemplified by the syn-
thesis of 18 in Scheme 1. The 3-morpholinobenzoic acid
31 was prepared either via palladium-catalysed amina-
tion chemistry from ethyl 3-bromobenzoate19 or by
direct nucleophilic aromatic substitution of 3-fluoro-
benzonitrile20 with subsequent hydrolysis of the result-
ing intermediates. Conversion to the acid chloride,
reaction with 4-methyl-3-nitroaniline, and reduction,
all under standard conditions, provided the key benzamido-
aniline intermediate 32. Subsequent acylation and
deprotection afforded the phenol 33, which was then
alkylated under mild basic conditions. This product
1821 and other analogues made in similar manner
were characterised by LCMS, 1H NMR and mass
spectrometry.
N
N
OMe
OMe
OMe
OMe
H
N
H
N
H
N
N
O
O
O
O
10
11
Replacement of the dimethylamino with a morpholino
substituent (18) resulted in a modest increase in kinase
inhibition with similar cellular potency (PBMC IC50
1.2lM), but encouraging activity in human whole
blood. Whole blood activity was observed for a number
Significant inhibition of p38a was found for a variety of
substituted alkoxy-compounds 1–17 (Table 1). Simple
O
O
O
N
N
NO2
H2N
Cl
H
OBn
a
i) b
i) d
NH2
OH
N
ii) c
ii) e
O
O
31
32
O
N
O
N
OH
O
N
f
H
N
H
N
H
N
H
N
O
O
O
O
33
18
Scheme 1. Preparation of 18. Reagents and conditions: (a) (COCl)2, DMF(cat), CH2Cl2, rt (quantitative); (b) Et3N, CH2Cl2, rt (80%); (c) 10% Pd on
carbon, H2, MeOH, rt (95%); (d) Et3N, CH2Cl2, rt (61%); (e) 10% Pd on carbon, H2, MeOH, rt (84%); (f) 2-(chloromethyl)pyridine, K2CO3, DMA,
60°C (55%).

D. S. Brown et al. / Bioorg. Med. Chem. Lett. 14 (2004) 5383–5387
5385
Table 1. Inhibition of human p38a enzyme activity and LPS-induced TNF-a release in human PBMCs by dimethylamino-substituted bisamides in
vitro
N
R3
H
N
H
N
R2
O
O
R1
No.
R1
R2
R3
p38a IC50 (lM)
PBMC IC50 (lM)
1
Me
Me
Me
Me
Me
H
H
HO
0.18
0.18
0.10
0.18
0.23
>20
>50
>50
>50
6.4
3.9
—
2
H
MeO
MeO
MeO
MeO
MeO
MeO
MeO
MeO
EtO
3
2-MeO
3-MeO
3,5-(MeO)2
3-MeO
3-MeO
3-MeO
3-MeO
3-EtO
H
4
5
6
7
F
Cl
1.90
0.22
0.34
0.05
0.06
0.17
0.27
0.31
0.051
0.026
22.0
6.4
9.5
4.5
7.0
4.6
3.4
9.5
1.4
0.19
8
9
12
Br
Me
Me
Me
Me
Me
Me
13
14
EtO
i-PrO
H
15
16
H
MeO(CH2)2O
EtO(CH2)2O
2-Pyridyl-CH2O
H
17
SB203580
H
Table 2. Inhibition of human p38a and LPS induced TNF-a release in human whole blood by para-(pyrid-2-ylmethoxy)-substituted bisamides in
vitro
NR2R3
O
N
X
H
N
H
N
R1
O
O
No.
X
R1
NR2R3
NMe2
p38a IC50 (lM)
HWB IC50 (lM)
17
18
HC
HC
N
H
H
H
F
0.051
0.030
0.010
0.064
0.020
0.025
0.026
>40
3.7
2.8
Morpholinyl
Morpholinyl
Morpholinyl
Morpholinyl
Pyrrolidinyl
19
20
HC
HC
HC
11.2
10.2
14.1
4.2
21
22
CF3
F
SB203580
Table 3. In vitro activity and rat PKAUCs for a normalised 1mg/kg dose for a series of morpholino-substituted bisamides
O
N
R2
O
H
N
H
N
R1
O
O
No.
R1
R2
p38a IC50 (lM)
HWB IC50 (lM)
AUC (lMh)
23
18
24
25
26
27
28
29
30
H
H
H
H
F
Phenyl
0.037
0.030
0.095
0.022
0.017
0.018
0.016
0.028
0.097
40.0
3.7
—
1.38
—
2-Pyridyl
3-Pyridyl
4-Pyridyl
4-Pyridyl
4-Thiazolyl
4-Thiazolyl
14.7
3.3
0.13
<0.02
<0.12
0.71
2.03
0.73
3.4
1.5
H
F
2.5
H
F
2-Methyl-4-thiazolyl
2-Methyl-4-thiazolyl
14.6
2.8

5386
D. S. Brown et al. / Bioorg. Med. Chem. Lett. 14 (2004) 5383–5387
Table 4. Inhibition of TNF-a production in SEB-stimulated Balb/C
10000
1000
100
10
mice in vivo24
No.
HWB
AUC
% Inhibition
IC50 (lM) (lMh) 30mg/kg 10mg/kg 3mg/kg
18
20
3.7
11.2
10.2
14.1
3.3
1.38
0.13
59
43
55
7
—
—
4
—
—
—
i.v.
p.o.
21
22
0.04
0.07
3
25
26
0.13
75
40
56
81
71
35
50
43
24
44
37
44
27
47
14
—
28
—
33
0
3.4
1.5
<0.02
<0.12
0.71
27
28
2.5
29
30
14.6
2.8
2.03
0.73
1
0.00
5.00
10.00
15.00
20.00
25.00
SB203580
0
Time (hours)
Graph 1. PKprofile for compound 19 in rat; dosing at 3mg/kg po
(n = 2) and at 1mg/kg iv (n = 3).
of other morpholino systems including the pyridyl ana-
logue 19, the fluoro and trifluoromethyl substituted
benzamides 20 and 21, and a pyrrolidinyl analogue 22
(Table 2). There seemed to be no clear explanation for
this significant improvement in whole blood activity in
moving from the dimethylamino to the cyclic amines;
kinase activity and lipophilicity (e.g. logDÕs 2.8 for 17,
3.2 for 18) were similar. In this instance the human pro-
tein binding (0.3% and 3.8% free for 17 and 18, respec-
tively) could be a factor but the fraction free was less for
other compounds that exhibited good blood potency
(e.g. 19 has only 0.11% free). Furthermore the morpho-
lino oxygen atom did not appear to be acting as a hydro-
gen bond acceptor in view of the similar inhibitory
potencies of TNF-a in blood of 20 and 22.
A full PKprofile was carried out in rat for the morpho-
linopyridine 19, which had shown good potency in
human whole blood, excellent metabolic stability (hu-
man microsomes, rat and human hepatocytes) in vitro
and high oral exposure in rat PK( Graph 1). From the
iv profile (dosing 1mg/kg) the compound exhibited low
clearance (10mL/min/kg) and moderate volume of distri-
bution (1.6L/kg) resulting in a half-life of 2.9h. Based
on the oral arm (dosing at 3mg/kg) bioavailability was
determined to be 46%. This compound had rather high
human protein binding (0.1% free) and very low solubil-
ity (<1lM) and was not progressed further but subse-
quent studies have shown that these issues can be
addressed within the series.
A range of further heterocyclic compounds was investi-
gated all of which were clearly more potent inhibitors in
blood than the corresponding phenyl compound 23, pre-
sumably due to the reduced lipophilicity and protein
binding in whole blood (Table 3). The best substituents
included the pyridyl isomers, 4-thiazolyl, and 2-methyl-
4-thiazolyl 24–30. In order to confirm that the series still
retained the original attribute of excellent selectivity for
p38a the methylthiazolyl 29 was evaluated against a
range of kinases and exhibited no significant activity at
10lM.22 As the compounds in Table 3 generally exhibi-
ted good metabolic stability during incubation of the
compounds with rat hepatocytes in vitro the more po-
tent derivatives were dosed orally in rats (n = 2 animals)
to assess systemic exposure. Pharmacokinetic (PK) evalu-
ation was carried out using a cassette dosing protocol
and area under the curve (AUC) normalised for a dose
of 1mg/kg (compounds were dosed at ꢀ2mg/kg in a
propyleneglycol formulation).23
In conclusion, the novel bisamide series of p38a kinase
inhibitors has afforded novel compounds with micromo-
lar activity in whole blood cytokine production assays in
vitro and significant inhibition of TNF-a production in
SEB-stimulated Balb/C mice in vivo. Several analogues
have good exposure from oral dosing and a leading ana-
logue has excellent PKparameters in rat. Further struc-
tural modification in the bisamide series may afford
improved physical properties leading to compounds
suitable for the treatment of RA.
Acknowledgements
The authors thank our colleagues Lucy C. Ashman,
Carol L. Biddulph, Karen M. Burns, and Brian Eller-
shaw for developing and carrying out the biological
assays, Dermot McGinnity for the PKstudy on com-
pound 19, and Graham Crawley for assistance in pro-
ducing the manuscript.
Bisamides analogues with IC50 values <15lM in human
whole blood were examined for inhibition of TNF-a
release in Balb/C mice challenged with staphylococcal
enterotoxin B (SEB). Compounds 18, 21, 25 and 27–29
inhibited circulating TNF-a levels by greater than 50%
at 30mg/kg, thus establishing in vivo efficacy for this ser-
ies. At this stage a PK–PD relationship was not estab-
lished as this would require mouse PKand whole
blood data (Table 4).
References and notes
1. Sundy, J. S. Mod. Ther. Rheum. Dis. 2002, 23.
2. (a) Pisetsky, D.; St. Clair, E. W. JAMA—J. Am. Med.
Assoc. 2001, 286(22), 2787; (b) Hughes, L. B.; Moreland,
L. W. BioDrugs 2001, 15(6), 379.
3. Capell, H. J. Rheumatol. Suppl. 2002, 66, 38.

D. S. Brown et al. / Bioorg. Med. Chem. Lett. 14 (2004) 5383–5387
5387
4. Laan, R. F. J. M.; Van Riel, P. L. C. M.; Van De Putte, L.
B. A. Curr. Opin. Rheumatol. 2001, 13(3), 159.
5. (a) Emery, P.; Scott, D. L.; Strand, V. Mod. Ther. Rheum.
Dis. 2002, 121; (b) Alldred, A.; Emery, P. Exp. Opin.
Pharmacother. 2001, 2(1), 125.
18. For in vitro assay experimental procedures see: Brown, D.
S.; Brown, G. R. WO 9959959 A1 19991125; Chem. Abstr.
131: 351089.
19. Brown, D. S.; Brown, G. R.; Cohen, P. WO 9915164 A1
19990401; Chem. Abstr. 130: 267218.
6. OÕDell, J. R. Mod. Ther. Rheum. Dis., 2002, 147.
7. (a) Keating, G. M.; Perry, C. M. BioDrugs 2002, 16(2),
111; (b) Markham, A.; Lamb, H. M. Drugs 2000, 59(6),
1341.
8. (a) Criscione, L. G.; St. Clair, E. W. Curr. Opin.
Rheumatol. 2002, 14(3), 204; (b) Mikuls, E. R.; Moreland,
L. W. Exp. Opin. Pharmacother. 2001, 2(1), 75.
9. (a) Taylor, P. C. Curr. Pharm. Des. 2003, 9(14), 1095; (b)
Paget, S. A. Adv. Ther. 2002, 19(1), 27.
10. Kumar, S.; Boehm, J.; Lee, J. C. Nat. Rev. (Drug Discov.)
2003, 2, 717.
11. Dean, J. L. E.; Brook, M.; Clark, A. R.; Saklatvala, J. J.
Biol. Chem. 1999, 274(1), 264.
12. (a) Jackson, P. F.; Bullington, J. L. Curr. Top. Med. Chem.
2002, 2, 1011; (b) Badger, A. M.; Bradbeer, J. N.; Votta,
B.; Lee, J. C.; Adams, J. L.; Griswold, D. E. J. Pharmacol.
Exp. Ther. 1996, 279(3), 1453; (c) Boehm, J. C.; Smietana,
J. M.; Sorenson, M. E.; Garigipati, R. S.; Gallagher, T. F.;
Sheldrake, P. L.; Bradbeer, J.; Badger, A. M.; Laydon, J.
T.; Lee, J. C.; Hillegass, L. M.; Griswold, D. E.; Breton, J.
J.; Chabot-Fletcher, M. C.; Adams, J. L. J. Med. Chem.
1996, 39(20), 3929; (d) Cuenda, A.; Rouse, J.; Doza, Y.
N.; Meier, R.; Cohen, P.; Gallagher, T. F.; Young, P. R.;
Lee, J. C. FEBS Lett. 1995, 364(2), 229; (e) Gallagher, T.
F.; Fier-Thompson, S. M.; Garigipati, R. S.; Sorenson, M.
E.; Smietana, J. M.; Lee, D.; Bender, P. E.; Lee, J. C.;
Laydon, J. T.; Griswold, D. E.; Chabot-Fletcher, M. C.;
Breton, J. J.; Adams, J. L. Bioorg. Med. Chem. Lett. 1995,
5(11), 1171.
20. (a) Brown, G. R.; Foubister, A. J.; Roberts, C. A.; Wells,
S. L.; Wood, R. Tetrahedron Lett. 2001, 42(23), 3917; (b)
Belfield, A. J.; Brown, G. R.; Foubister, A. J.; Ratcliffe, P.
D. Tetrahedron 1999, 55(46), 13285, Corrigendum Tetra-
hedron 2000, 56(47), 4107; (c) Brown, G. R.; Foubister, A.
J.; Ratcliffe, P. D. Tetrahedron Lett. 1999, 40(6),
1219.
21. The product gave the following data: NMR (DMSO-d6)
2.19 (s, 3H), 3.18 (t, 4H), 3.73 (t, 4H), 5.23 (s, 2H), 7.12 (d,
3H), 7.2 (d, 1H), 7.34 (m, 3H), 7.42 (s, 1H), 7.54 (m, 2H),
7.8 (m, 2H), 7.97 (d, 2H), 8.58 (d, 1H), 9.72 (s, 1H), 10.1 (s,
1H); Mass MH+ 523.
22. For further information on kinases and assay protocols
see: Davies, S. P.; Reddy, H.; Caivano, M.; Cohen, P.
Biochem. J. 2000, 351(1), 95. The kinase panel used was
the Core panel shown in Table 1.
23. For an example of this type of method see: Korfmacher,
W. A.; Cox, K. A.; Ng, K. J.; Veals, J.; Hsieh, Y.;
Wainhaus, S.; Broske, L.; Prelusky, D.; Nomeir, A.;
White, R. E. Rapid Commun. Mass Spectrom. 2001, 15(5),
335.
24. Procedure for mouse SEB in vivo model: Female Balb/C
mice (18–22g) were dosed orally with test compounds (30,
10 and 3mg/kg in 20% DMSO/60% PEG 400/20% water;
0.1mL volume; 4 animals per group), or vehicle alone,
60min prior to intraperitoneal (ip) challenge with 50g
staphylococcal enterotoxin B (SEB from Toxin Technol-
ogy BT202) in 0.2mL pyrogen-free physiological saline
(Phoenix PL 1502/0006R). Control animals were chal-
lenged (ip) with 0.2mL pyrogen-free physiological saline.
Animals were sacrificed 2h post SEB-challenge and bled
via the vena cava. Serum was obtained following incuba-
tion in serum gel tubes (Sarstedt 41-1500.005) for 30min at
room temperature and centrifugation. Samples were stored
at À80°C prior to determination of mouse TNF-a content
by ELISA (Genzyme 80-2802-05). Cyclosporin A (Neoral,
Novartis)(20mg/kg dosed orally in 0.1mL physiological
saline) and SB220025 (100, 25, 6.25 and 1.56mg/kg
in DMSO/PEG400/water formulation; 0.1mL dosed
orally) (synthesised in house) were used as positive
controls.
13. Chakravarty, S.; Dugar, S. Ann. Rep. Med. Chem. 2002,
37, 177.
14. Cirillo, P. F.; Pargellis, C.; Regan, J. Curr. Top. Med.
Chem. 2002, 2, 1021.
15. Hedge, P. J.; Boyle, F. T. WO 9822103 A1 19980528;
Chem. Abstr. 129, 45323.
16. Work carried out on behalf of AstraZeneca by Sir Philip
Cohen, Division of Signal Transduction Therapy, Univer-
sity of Dundee, Dundee, UK.
17. Hall-Jackson, C. A.; Eyers, P. A.; Cohen, P.; Goedert, M.;
Boyle, F. T.; Hewitt, N.; Plant, H.; Hedge, P. Chem. Biol.
1999, 6(8), 559.