Indolin-2-one p38α inhibitors III: Bioisosteric amide replacement

Article abstract of DOI:10.1016/j.bmcl.2011.09.006

Crystallographic structural information was used in the design and synthesis of a number of bioisosteric derivatives to replace the amide moiety in a lead series of p38a inhibitors which showed general hydrolytic instability in human liver preparations. Triazole derivative 13 was found to have moderate bioavailability in the rat and demonstrated potent in-vivo activity in an acute model of inflammation.

Full text of DOI:10.1016/j.bmcl.2011.09.006

Bioorganic & Medicinal Chemistry Letters 21 (2011) 6253–6257
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Indolin-2-one p38a inhibitors III: Bioisosteric amide replacement
⇑
Paul Eastwood , Jacob González, Elena Gómez, Francisco Caturla, Nuria Aguilar, Marta Mir,
Josep Aiguadé, Victor Matassa, Cristina Balagué, Adelina Orellana, María Domínguez
Almirall Research Center, Almirall Laboratories, Ctra. Laureà Miró 408, E-08980 St. Feliu de Llobregat, Barcelona, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
Crystallographic structural information was used in the design and synthesis of a number of bioisosteric
Received 2 August 2011
Revised 1 September 2011
Accepted 2 September 2011
Available online 10 September 2011
derivatives to replace the amide moiety in a lead series of p38a inhibitors which showed general hydro-
lytic instability in human liver preparations. Triazole derivative 13 was found to have moderate
bioavailability in the rat and demonstrated potent in-vivo activity in an acute model of inflammation.
Ó 2011 Elsevier Ltd. All rights reserved.
Keywords:
p38a Inhibitors
Structure-based design
Bioisosterism
Indolin-2-ones
In recent communications,1 the design, synthesis and
Glu71
X
preliminary optimization of a potent series of indolin-2-one p38
a
Me
H
N
inhibitors, as exemplified by lead 1 (Fig. 1), was described.
Glu71
H
R3
O
R1
Y
Z
Me
Me
R2
H
N
H
N
H
N
Asp168
Glu71
N
A
R3
O
R1
O
O
O
Me
R2
Asp168
H
H
X
N
N
α
1: IC50 p38 = 1.3 nM
O
R3
Y
N
R1
R2
Figure 1. Lead molecule of the indolin-2-one p38
a
inhibitor series.
Asp168
B
Figure 2. Design of amide bioisosteres.
The preliminary compounds from this series were found to be
metabolically labile when subjected to both rat and human hepatic
microsomes. Of particular note is that amide bond hydrolysis oc-
curred to a significant extent under non-oxidative metabolism con-
ditions, particularly upon incubation with human liver microsome
preparations.1a Subsequently it was demonstrated that both
oxidative and non-oxidative metabolic pathways could be attenu-
ated by the use of metabolic blocking groups and by tuning the
physiochemical properties of the series.1b
Herein are described additional studies that were aimed at the
definitive elimination of the hydrolytic metabolic pathway via
replacement of the amide moiety by stable bioisosteres. Two
distinct approaches (Fig. 2) were envisioned to achieve this goal:
(a) Bioisosteric replacement of the amide by a heterocyclic moiety:2
it was surmised that derivatives of type A (Fig. 2), in which a
heterocyclic moiety could maintain, in whole or in part, the hydro-
gen bonding interactions seen between the amide of 1 and residues
Glu71/Asp168 in the p38
a
co-crystal structure,3 could serve as
hydrolytically stable amide replacements.4
(b) Preparation of bicyclic derivatives of type B (Fig. 2): in this case
a heterocyclic amine could potentially fulfill the same functions as
the amide moiety given that (i) the crystal structure of 1 bound in
p38a
3 shows that only one lone pair of the oxygen of the carbonyl
group is used to form the hydrogen bond with Asp168 and (ii) the
dihedral angle between the tolyl ring of 1 and the carbonyl
group of the amide was reasonably acute (31°) suggesting that
conformation restriction onto to the proximal aromatic ring may
be tolerated.5
Synthetic routes to compounds of type
Schemes 1–5. Key heterocyclic intermediates (2–7) were prepared
A are given in
⇑
Corresponding author.
E-mail address: paul.eastwood@almirall.com (P. Eastwood).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.09.006

6254
P. Eastwood et al. / Bioorg. Med. Chem. Lett. 21 (2011) 6253–6257
R
O
O
O
O
R
N
CO2Me
N
N
(i)
(ii)
NH
NH2
(i)
(ii)
N
O
NH2
N
N
I
I
X
X
X
X
2
7
(iii)
X = Br or I
(v)
(iv)
X = I
X = B(OH)2
R
(iii)
NH
N
HN
(vi)
N
O
N
Scheme 4. Reagents and conditions: (i) NH2NH2ꢀH2O, EtOH, reflux, 86%; (ii)
MeC(OEt)3, AcOH, 150 °C, 78%; (iii) iPrMgCl, B(OiPr)3, THF, ꢁ10 °C to rt then HCl
(aq), 66%.
X
X
X
3
Scheme 1. Reagents and conditions: (i) RC(OMe)2NMe2, 120 °C; (ii) NH2OHꢀHCl,
NaOAc, AcOH, 100 °C; (iii) NH2NH2ꢀH2O, AcOH, 90 °C; (iv) (CF3CO)2O, Et3N, CH2Cl2,
0 °C to rt; (v) HCl, EtOH, 0 °C; (vi) RCONHNH2, MeOH, 75 °C then POCl3, 110 °C.
H
X
N
(i)
2-7
9-21
O
R1
22-23
24-32
R2
NH2
N
OH
N
8: X = Br or B(pin)
(ii)
O
N
Scheme 5. Reagents and conditions: (i) PdCl2dppfꢀCH2Cl2, 2 M Cs2CO3 (aq),
1,4-dioxane, 100–110 °C.
N
(i)
I
I
4
I
NH2
NH
HN
the oxadiazole heterocycles (data not shown) gave rise to less po-
tent compounds with poor and erratic whole blood potencies and
high metabolic turnover. However, compounds containing a 1,2,4-
triazole core ring system (12, 13) were, in general, found to be po-
tent enzyme inhibitors whilst possessing moderate whole blood
potencies. The corresponding imidazole derivatives 14 and 15
which would be protonated at physiological pH were considerably
less potent and no further derivatives were synthesized. As ex-
pected, the replacement of the amide furnished, in all cases, hydro-
lytically stable molecules as ascertained by incubation with both
rat and human liver microsomes in the absence of the co-oxidant
NADPH (data not shown).
From this selection of potential heterocyclic amide bioisosteres,
the specific hydrogen-bonding requirements and overall polarity
was seemingly best met by a 1,2,4-triazole and a further set of
compounds was prepared that contained this ring system (Table
2). Methyl was found to be the optimal substituent for the triazole
ring albeit from a very limited set of compounds (12, 16, 17).
Changes to the spirocyclic portion of the molecule were then inves-
tigated. Replacement of the spirocycle with a gem-dimethyl substi-
tuent (19) resulted in a drop in potency of several fold. A reduction
in ring size to give the cyclobutyl derivative 18 gave a compound
which was potent in the enzymatic assay but showed a large shift
in the whole blood assay. However, increasing the ring size to six
(iv)
I
(iii)
N
I
5
Scheme 2. Reagents and conditions: (i) NH2OHꢀHCl, NaOAc, EtOH, H2O, 65 °C, 60%;
(ii) (CH3CO)2O, 130 °C; (iii) LiHMDS, THF, 10 °C then 3 M HCl (aq), 25%; (iv)
CH3COCH2Cl, KHCO3, THF, H2O, reflux, 22%.
O
O
HN
N
(i)
(ii)
Br
Br
Br
Br
6
Scheme 3. Reagents and conditions: (i) Br2, CHCl3, 0 °C to rt, 95%; (ii)
CH3C(@NH)NH2ꢀHCl, KHCO3, THF, H2O, reflux, 5%.
using standard synthetic chemistry (Schemes 1–4).6 Suzuki
coupling of intermediates 2–7 (Scheme 5) with indolin-2-one
fragments 81a gave rise to final compounds 9–21 (Tables 1 and 2).
Biological results are presented in Tables 1 and 2.7 Of the three
isomeric oxadiazoles prepared (9–11), only the 1,2,4-isomer 9
showed respectable potency, albeit substantially less than the
amide derivative 1. All changes made to the alkyl substituent on
Table 1
Biological activities of heterocyclic derivativesa,7
Me
H
N
X
R3
O
R1
Y
Z
R2
Compd
R1, R2
R3
X
Y
Z
p38
a
IC50 (nM)
Whole blood IC50 (nM)
18
172 10
ND
ND
127 13
1
9
10
11
12
–(CH2)4–
–(CH2)4–
–(CH2)4–
–(CH2)4–
–(CH2)4–
—
—
O
N
N
NH
—
N
N
O
N
—
N
O
N
N
1.3 0.05
32 10
583 149
741 318
7.4 0.7
2
Me
Me
Me
Me
13
14
15
Me
Me
Me
NH
NH
NH
N
N
C
N
C
7.6 2.7
93
6
O
O
O
173
8
ND
ND
N
880 190
a
IC50 values are reported as the mean of at least two experiments. ND = Not Determined.

P. Eastwood et al. / Bioorg. Med. Chem. Lett. 21 (2011) 6253–6257
6255
Table 2
Biological activities of triazole derivativesa,7
Me
H
N
H
N
R3
O
R1
N
N
R2
Compd
R1, R2
R3
p38a
IC50 (nM)
Whole Blood IC50 (nM)
12
16
17
18
19
–(CH2)4–
–(CH2)4–
–(CH2)4–
–(CH2)3–
Me, Me
Me
H
cyc-Pr
Me
Me
7.4 0.7
9.0 2.0
135 12
127 13
253 64
ND
624 150
ND
18
1
87 16
13
20
21
Me
Me
Me
7.6 2.7
93
6
O
F
2
0.5
98 43
F
4.7 0.4
50
6
N SO2Me
a
IC50 values are reported as the mean of at least two experiments unless otherwise stated. ND = Not Determined.
members with concomitant introduction of heteroatoms, with the
aim of attenuating oxidative metabolism, furnished a number of
interesting compounds (13, 20, 21) with good potency in the enzy-
matic assay and that were reasonably potent in the whole-blood
assay.
Following the second strategy to replace the amide, a number of
bicylic compounds of type B (Fig. 2) were then prepared—represen-
tative synthetic routes to key intermediates are given in
Scheme 6.6a Benzisoxazole intermediates 22 and 23 were accessed
in several synthetic steps from commercially available benzonitr-
iles, key steps being ortho lithiation and trapping with iodine to
prepare the key iodides and microwave-assisted cyclization to give
the benzisoxazole nuclei 22. As before, palladium-catalyzed
coupling of 22 or 23 with the appropriate indolin-2-ones 8
(Scheme 5) gave compounds 24–32. Biological activities and struc-
tures of final compounds are presented in Table 3.
R4
R4
I
R4
I
(i)
(ii-v)
(vi-vii)
OH
N
N
F
F
N
F
F
R4
I
R4
R4
(HO)2B
H
N
(viii)
R3
(ix)
H
H
N
N
I
R3
R3
O
O
N
N
N
OH
22
23
Scheme 6. Reagents and conditions: (i) LiTMP, THF, I2, ꢁ78 °C to rt; (ii) H2SO4,
dioxane, 115 °C; (iii) BH3ꢀMe2S, B(OMe)3, THF, 0 °C to rt; (iv) MnO2, CH2Cl2, 50 °C;
(v) NH2OHꢀHCl, EtOH, rt; (vi) NCS, DMF, 55 °C; (vii) R3NH2, THF, 0 °C to rt; (viii) DBU,
THF, 150 °C, microwaves; (ix) PrMgCl, B(OiPr)3, THF, ꢁ10 °C to rt then HCl (aq).
i
substituent in either R4 or R5 (see Table 3), predisposing the two
biaryl rings in an almost orthogonal arrangement,8 is necessary
for optimal enzymatic potency (derivatives 24–26), although re-
moval of the R4 substituent led to compounds with poor
whole blood potency (25, 26); (ii) the five-membered spirocyclic
ring can be replaced by either a gem-dimethyl group or a
It was pleasing to find that the first benzisoxazole derivative
prepared (24) had low nanomolar potency in both the isolated en-
zyme and whole blood assays. From the limited number of deriva-
tives prepared the following trends were observed: (i)
a
Table 3
Biological activities of bicyclic aromatic derivativesa,7
R4
H
N
H
N
O
R1
R3
O
N
R5
R2
R4
Compd
R1, R2
R3
R5
p38a
IC50 (nM)
Whole blood IC50 (nM)
24
25
26
27
–(CH2)4–
–(CH2)4–
–(CH2)4–
Me, Me
cyc-Pr
cyc-Pr
cyc-Pr
cyc-Pr
Me
H
H
H
H
Cl
H
0.9 0.2
38 14
4.0 1.5
0.4 0.1
2.0 0.1
1257 577
438 29
50 11
Me
28
cyc-Pr
Me
Me
H
0.16 0.03
8.4 0.4
O
F
29
cyc-Pr
H
0.06 0.01
10.7 4.9
731 199
F
Me, Me
Me, Me
30
31
H
H
Me
Me
H
H
38 11
5.0 2.0
145
4
O
F
32
H
Me
H
1.2 0.2
137 (n = 1)
F
a
IC50 values are reported as the mean of at least two experiments unless otherwise stated.

6256
P. Eastwood et al. / Bioorg. Med. Chem. Lett. 21 (2011) 6253–6257
six-membered ring (vide supra) giving rise to derivatives that are
extremely potent in the enzymatic assay (27–29); (iii) the high
intrinsic potency of the six-membered ring analogues allows for
the removal of the amine substituent whilst maintaining com-
pounds of good enzymatic potency (30–32).
Other bicylic nuclei other than the benzisoxazole were briefly
investigated (Fig. 3). The indazole 339 proved to be inactive (cf.
the corresponding benzisoxazole 31, IC50 = 5.0 nM). The triazolo-
pyridine 34,10 however, was found to be a potent (IC50 = 34 nM)
inhibitor in the enzymatic assay; unfortunately this compound
was found to lose almost two orders of magnitude of potency in
the whole blood assay and no further examples were prepared.
The in-vitro metabolic turnover in hepatic microsomes, perme-
ability in a Caco-2 cell line and cytotoxicity in a Chinese hamster
ovarian (CHO) cell line were determined for a selection of the most
interesting compounds and results are given in Table 4.
from both the triazole and benzisoxazole series and the results are
presented in Tables 5 and 6.
Compound 13 was found to have a good intravenous half life in
rat, as a result of low clearance and a moderate volume of distribu-
tion, and this coupled with a moderate oral bioavailability and
good blood levels triggered further profiling of this derivative.
Bicycle 28 was also found to have a good intravenous pharmacoki-
netic profile with both low clearance and a moderate volume of
distribution contributing to a good half-life—however, upon oral
dosing a poor oral bioavailability was noted.
Compound 13 was tested in-vivo in an acute model of inflam-
mation in the rat based on LPS-induced TNF
model, compounds were dosed orally one hour prior to LPS admin-
istration and the amount of TNF in plasma was measured 1.5 h la-
ter (coinciding with the peak TNF production). Compound 13
dose-dependently inhibited TNF production with an ED50
a
production.7 In this
a
a
a
=
In general, compounds presented good to excellent apical to ba-
sal flux in the Caco-2 assay apart from compound 21 which ap-
pears to have transporter-compromised permeability issues as
determined by the significantly large BA/AB ratio. Metabolic turn-
over (NADPH dependent) was found to be low to moderate in both
rat and human microsomes and, in contrast to the amide lead, no
metabolism was seen in the absence of NADPH in the human prep-
aration (data not shown). In addition, compounds did not show any
significant cytotoxicity towards CHO cells with the exception of
compound 20.
0.69 mg/kg (95% CI: 0.4–1.0) performing better than the clinically
relevant reference compound BIRB-796 (ED50 = 7.9 mg/kg, 95%
CI:3–22).
In order to determine general selectivity, 13 was tested at a con-
centration of 10
lM in a panel of 51 representative kinases (Up-
state) and panel of 54 proteins which included GPCRs,
a
phosphodiesterases, hormone receptors, ion channels and trans-
porters (CEREP ‘ExpresSProfile’). Overall, 13 demonstrated an
excellent selectivity profile.11 In relation to potential cardiovascu-
lar toxicity, 13 proved to be a relatively weak inhibitor of the the
On the basis of the in-vitro results, the intravenous and oral rat
pharmacokinetic profiles were determined for compounds selected
human IKr (hERG) ion channel (49 8% block @ 10
lM) as deter-
Table 6
Pharmacokinetic profiles in rat (10 mg/kg oral)a
H
N
H
N
N
Compd
Cmax (ng/mL)
AUC0–1 (ng h/mL)
F (%)
NH2
NH2
O
O
HN
N
N
N
1
13
21
28
31
367
1105
632
479
141
2644
11,683
1371
3372
961
10
29
13
6
O
33
p38α: 60 2% I @ 60μM
34
p38α IC50 = 34 11 nM
WB IC50 = 2113 260 nM
8
a
Formulations: Compds 1,13, 21, 31: 0.5% methylcellulose + 0.1% Tween 80; 28:
40% PEG + 0.4% HCl 1 N + 0.5% DMSO.
Figure 3. Miscellaneous bicyclic nuclei.
Table 4
In-vitro profiles of selected derivatives
Rat/human metabolism (+NADPH)a
Caco-2 (AB/BA) Pappb (10ꢁ6 cm/s)
Cytotoxicity CHO IC50 (lM)
c
Compd
1
44/55
15/13
23/<5
5/<5
18/10
ND
18/27
ND
2/45
26/16
10/4
28/16
26
12
13
20
21
28
29
31
53% @ 100
48% @ 200
20
100
49
l
l
M
M
<5/6
18/26
25/32
9/3
56% @ 200
16% @ 200
l
l
M
M
a
b
c
% Turnover after a 30 min incubation period at 37 °C of a 5 lM solution of test compound with hepatic microsomes (1 mg/mL).
Passive permeability through a Caco-2 monolayer determined using 12.5
lM test compound. ND = Not Determined.
For assay details see Ref. 1a.
Table 5
Pharmacokinetic profiles in rat (1 mg/kg iv)a
Compd
t½ (h)
Cmax (ng/mL)
Cl (mL/min/kg)
AUC0–1 (ng h/mL)
Vss (L/kg)
MRT (h)
1
13
21
28
31
5.3
3.1
3.8
2.9
1.9
2002
1719
1295
1475
752
6.6
4.2
17.2
3.2
2532
4063
1030
5271
1230
0.8
1.0
1.75
0.8
2.0
2.1
4
1.8
4.6
2.5
13.8
a
Formulations: Compds 1: 40% PEG; 13, 21, 31: 40% PEG + 0.4% HCl 1 N; 28: 40% PEG + 0.4% HCl 1 N + 0.5% DMSO.

P. Eastwood et al. / Bioorg. Med. Chem. Lett. 21 (2011) 6253–6257
6257
7. Full details of all biological assays and models used in this article can be found
in: Lumeras, W.; Caturla, F.; Vidal, L.; Esteve, C.; Balagué, C.; Orellana, A.;
Domínguez, M.; Roca, R.; Huerta, J. M.; Godessart, N.; Vidal, B. J. Med. Chem.
2009, 52, 5531. and Ref. 1a.
8. The torsion angle between these two rings in the bound conformation of
compound 1 is 71° as ascertained from the crystal structure (Ref. 1a).
9. Compound 33 was accessed by the following synthetic route:
mined in human embryonic kidney (HEK-293) cells. In addition,
compound 13 did not inhibit any of the most relevant human cyto-
chrome P450 (CYP) isoforms (IC50 >25
2C19 and 2D6).
lM for CYPs 1A2, 3A4, 2C9,
In summary, the replacement of the metabolically unstable
amide moiety in a chemical lead series by a variety of mono and
bicyclic heterocyclic bioisosteres has led to potent and metaboli-
cally stable compounds. A representative example (compound
13) demonstrated moderate bioavailability in the rat and potent
in-vivo activity in an acute model of inflammation.
Me
(i, ii)
33
CN
I
F
Reagents and conditions: (i) NH2NH2ꢀH2O, EtOH, 100 °C, 78%; (ii) 8 [X = B(pin)],
PdCl2dppfꢀCH2Cl2, 2 M Cs2CO3 (aq), 1,4-dioxane, 110 °C, 29%.
References and notes
10. Compound 34 was accessed by the following synthetic route:
1. (a) Eastwood, P.; González, J.; Gómez, E.; Vidal, B.; Caturla, F.; Roca, R.; Balagué,
C.; Orellana, A.; Domínguez, M. Bioorg. Med. Chem. Lett. 2011, 21, 4310; (b)
Eastwood, P.; González, J.; Gómez, E.; Caturla, F.; Balagué, C.; Orellana, A.;
Domínguez, M. Bioorg. Med. Chem. Lett. 2011, 21, 5270.
2. Recent reviews on bioisosterism include: (a) Meanwell, N. A. J. Med. Chem.
2011, 54(8), 2529; (b) Lima, L. M.; Barriero, E. J. Curr. Med. Chem. 2005, 12, 23.
3. The crystallographic binding mode of compound 1 has been published in Ref.
1a.
Me
Me
I
Me
(i,ii)
(iii)
N
N
N
O
I
H2N
PMB
HN
NHNH2
N
H
N
H
Cl
PMB = p-methoxybenzyl
4. During the course of this work it was revealed that heterocyclic replacements
Me
I
of the amide are tolerated in other series of p38
a
inhibitors: (i) Angell, R. M.;
(iv)
(v)
H
N
34
N
Bamborough, P.; Cleasby, A.; Cockerill, S. G.; Jones, K. L.; Mooney, C. J.; Somers,
D. O.; Walter, A. L. Bioorg. Med. Chem. Lett. 2008, 18, 318; (ii) Angell, R. M.;
Angell, T. D.; Bamborough, P.; Brown, D.; Brown, M.; Buckton, J. B.; Cockerill, S.
G.; Edwards, C. D.; Jones, K. L.; Longstaff, T.; Smee, P. A.; Smith, K. J.; Somers, D.
O.; Walker, A. L.; Willson, M. Bioorg. Med. Chem. Lett. 2008, 18, 324.
PMB
N
N
Reagents and conditions: (i) NaNO2, HCl, H2O, 0 °C then KI, 0 °C to rt, 58%; (ii)
NH2NH2ꢀH2O, pyridine, 115 °C, 78%; (iii) p-methoxybenzylisocyanate, CH2Cl2,
rt, 81%; (iv) POCl3, CH3CN, 75 °C, 39%; (v) 8 [X = B(pin)], PdCl2dppfꢀCH2Cl2, 2 M
Cs2CO3 (aq), 1,4-dioxane, 110 °C then TFA, CH2Cl2, rt 20%.
5.
A similar strategy was adopted independently in another series of p38a
inhibitors: Pettus, L. H.; Xu, S.; Cao, G.-Q.; Chakrabarti, P. P.; Rzasa, R. M.; Sham,
K.; Wurz, R. P.; Zhang, D.; Middleton, S.; Henkle, B.; Plant, M. H.; Saris, C. J. M.;
Sherman, L.; Wong, L. M.; Powers, D. A.; Tudor, Y.; Yu, V.; Lee, M. R.; Syed, R.;
Hsieh, F.; Tasker, A. S. J. Med. Chem. 2008, 51(20), 6280.
11. IC50 values were determined for relevant inhibitory values (>50% inhibition @
10
l
M) with the following findings: 5-HT1B: IC50 ꢂ10
lM, cRaf: IC50 = 1.6 lM,
TrkA(h): IC50 = 7.7
l
M.
6. (a) Full synthetic details can be found in the following patents: Eastwood, P. R.;
González, J.; Matassa, V. PCT Int. Appl. WO2009132774, 2009.; (b) Eastwood, P.
R.; González, J.; Vidal, B.; Aguilar, N. PCT Int. Appl. WO2009124692, 2009.