J. Defaux et al. / Bioorg. Med. Chem. Lett. 24 (2014) 3748–3752
3751
Table 4
In vitro ADME properties and safety data of compounds 12, 13 and 18
N
O
R
N
H
N
H
N
Ar
Compounds
R
Ar
Solubilitya
M)
Mouse liver microsomes
(% rem. after 1 h)
Mouse plasma (%
rem. after 6 h)
Caco-2 permeability ab/
bab (10À6 cm sÀ1
CYP3A4
inhibition
hERG
binding
(
l
)
(lM)
(lM)
O
12
Et
5.7
45.4
n.d.
n.d.
n.d.
n.d.
O
O
13
18
tBu
8.5
3.9
70.4
81.1
n.d.
n.d.
n.d.
n.d.
>60
O
N
N
CH2CH2Ph
96.7
64/29
>100
a
Measured at pH 7.4, PBS medium + 1% BSA.
In the presence of 1% BSA. n.d.: not determined. rem.: remaining.
b
30 nM), with an approximate 10-fold increase in potency com-
pared to other compounds. In comparison to the previously
described 4,7-disubstituted analogues which were inactive against
ERK2,16 their potency is maintained against Aurora B. This last
observation is unclear since Aurora kinases contain a DFG motif
able to adopt complex and unusual conformations and its possible
involvement is still studying.24
aqueous solubility (3.9–8.5 lM) was very weak and constituted a
deleterious parameter in terms of suitable PK profile. Stability tests
disclosed that the three analogues showed moderate (45.4%, 12) to
good (70.4%, 13 and 81.1%, 18) metabolic stability in liver
microsomes. In addition, urea 18 exhibited an excellent metabolic
stability of 96.7% in mouse plasma. Moreover, cell permeability in
Caco-2 cells27 (Table 4) indicated that compound 18 had a very
good cell permeability associated with a reduced efflux (0.45)
(ratio = 0.45, [b?a/a?b]). Safety testing showed no inhibition of
cytochrome P450, CYP3A428 and no undesired activity on hERG29
channel for compound 18.
In summary, a new series of dual ERK2 and Aurora B kinases
inhibitors has been developed based on (7-aryl-1,5-naphthyridin-
2-yl)urea scaffold. The molecules displayed antiproliferative
activity in various cancer cells in the low micromolar and submi-
cromolar range, a result that brings a significant improvement over
previously published 4,7-disubstituted isomers which remained
inactive in this assay.16 The notion of inhibiting multiple kinases
involved in tumor progression with a single small molecule is well
established.30 In this context, additional ERK2 inhibition led us to
design very promising antitumor agents and this SAR study pro-
vides a useful starting point for lead optimization. In particular,
efforts will be made to enhance the solubility of these poorly
water-soluble compounds. To validate the concept of synergistic
effect of the observed dual inhibition, additional pharmacological
experiments are in progress to prove if the observed antiprolifera-
tive effect can be assigned to key signaling events of the respective
target inhibition.
The binding pose found by the docking program GOLD (GOLD
version 4.0; CCDC, Cambridge, UK) for the most active compound
17 into the ATP pocket of human ERK2 is shown in Figure 2. A
key hydrogen bond with the ATP hinge residue Met106 is observed
while the urea moiety is directed toward the gatekeeper residue
Gln103. Even if no additional hydrogen bond was done
(d = 3.51 Å between NH and carbonyl groups), we can suppose that
side-chain flexibility could be engaged to favour this specific type
of interaction. Thus, as the gatekeeper residue is poorly conserved
among other tested kinases, and is known to play a well-estab-
lished role in determining the selectivity of kinase inhibitors, it is
likely that the lack of inhibitory activities may be explained in part
either by the hydrophobic nature of the corresponding residues
from HIPK1 (Phe), Pim1 (Leu), KDR (Val) and TrkA (Phe) or shorter
polar residues with c-Abl and Yes (Thr) that could not interact with
the urea moiety. Finally, the benzyl group should also probably
contribute to the affinity by making closer hydrophobic interac-
tions with Tyr34 and Val37 residues in the P-loop motif.
Compounds 11–18 were tested for their antiproliferative activ-
ity against five cancer cell lines (HCT116 colon, MDA-MB468
breast, PC3 prostate, A549 NSCLC and U87MG CNS: Table 2). In a
general trend, [7-(1-methyl-1H-pyrazol-4-yl)-1,5-naphthyridin-2-
yl]ureas 15–18 proved to be very active, except on U87MG cell line,
displaying micromolar IC50 values. Interestingly, benzylurea 17
was selective of HCT116 cell line with a very promising IC50 value
Acknowledgment
We thank AtlanChim Pharma Company for lipophilicity
determination.
of 0.191 lM and this compound was also the most active against
ERK2 kinase (IC50 = 30 nM). (1,5-Naphthyridin-2-yl)urea counter-
parts 11–14, without pyrazolyl moiety, remained inactive or
showed slight antiproliferative potency toward the tested tumor
cell lines.
In order to determine lipophilicity profile of the naphthyridine
ureas 12–18, their partition coefficient at pH 7.4, expressed as
LogD, was directly derived from measurements of their chromato-
graphic hydrophobicity index (CHI) (Table 3).16,26 LogD ranged
from 2.2 to 3.2 and were found acceptable for cell permeability.
Compounds 12, 13 and 18 were profiled for PBS aqueous
solubility and mouse liver microsomal stability (Table 4).16 Their
Supplementary data
Supplementary data associated with this article can be found, in
References and notes