p38 MAP kinase inhibitors. Part 3: SAR on 3,4-dihydropyrimido[4,5-d]pyrimidin-2-ones and 3,4-dihydropyrido[4,3-d]pyrimidin-2-ones

Article abstract of DOI:10.1016/j.bmcl.2006.05.045

p38 inhibitors based on 3,4-dihydropyrimido[4,5-d]pyrimidin-2-one and 3,4-dihydropyrido[4,3-d]pyrimidin-2-one platforms were synthesized and preliminary SAR explored. Among the pyrimido-pyrimidones the emergence of two sub-types of analogs-C7-amino-pyrimi

Full text of DOI:10.1016/j.bmcl.2006.05.045

Bioorganic & Medicinal Chemistry Letters 16 (2006) 4400–4404
p38 MAP kinase inhibitors. Part 3: SAR on 3,4-dihydropyrimido-
[4,5-d]pyrimidin-2-ones and 3,4-dihydropyrido[4,3-d]-
pyrimidin-2-ones
Swaminathan R. Natarajan,a,* David D. Wisnoski,a James E. Thompson,b
Edward A. O’Neillb and Stephen J. O’Keefeb
aDepartment of Medicinal Chemistry, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065, USA
bDepartment of Inflammation Research, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065, USA
Received 10 April 2006; revised 13 May 2006; accepted 15 May 2006
Available online 5 June 2006
Abstract—p38 inhibitors based on 3,4-dihydropyrimido[4,5-d]pyrimidin-2-one and 3,4-dihydropyrido[4,3-d]pyrimidin-2-one plat-
forms were synthesized and preliminary SAR explored. Among the pyrimido-pyrimidones the emergence of two sub-types of ana-
logs—C7-amino-pyrimidines such as 24 and C7-amino-piperidines such as 42—characterized with good p38 inhibition and better
off-target profiles in terms of ion channel activities was significant. Representative compound 54 in the pyrido-pyrimidone class
was found to be equipotent with corresponding analog in the quinazolinone series.
ꢀ 2006 Elsevier Ltd. All rights reserved.
VX-7451 (1, Fig. 1), harbinger of a new class of exqui-
sitely selective2 p38 MAP kinase inhibitors, inspired
the development of two novel series3 of inhibitors typi-
fied by dihydro-pyridopyrimidones 3 and 4, and dihy-
dro-quinazolinones 5 and 6 (Fig. 1).
Cl
Cl
N
Cl
Cl
N
O
F
O
N
R
S
F
X
N
HN
S
Dihydro-quinazolinone analogs such as
IC50 = 1 nM), (p38a IC50 = 6 nM),
5
7
(p38a
(p38a
F
F
6
1 : X = N (VX-745)
2 : X = NH
pyrimido-pyridazine core
3 : R = H
IC50 = 2 nM), and 8 (p38a IC50 = 2 nM) that are devoid
of a C7-piperidine substituent were found to lack signif-
icant cellular and functional activity. Introduction of a
solubilizing amine at the C7-position imparted function-
4 : R =
N
N
N
pyrido-pyrimidine core
al activities to this scaffold. Benchmark compounds3b
9
(0.2 nM) and 10 (0.2 nM) were not only very potent
in vitro, they also showed good functional activity (hu-
man whole blood IC50s were 10 nM and 20 nM, respec-
tively) and oral bioavailability. Compounds in this class
bearing the basic amine appendage such as the piperi-
dine moeity in 9 and 10 also displayed potent ion chan-
nel activities which correlate with increased
cardiovascular risk. Compounds in chronic usage neces-
sitate greater cardiovascular safety profiles. VX-745 (1)
R
Cl
Cl
Cl
Cl
N
O
N
R
O
N
7
7
HN
HN
Cl
F
F
: R = H
10 : R = iPr
quinazolinone core
Figure 1. Novel p38 inhibitors, VX-745 and homologs.
5 : R = OMe 7 : R = COOMe
8 : R = OH
9
6 : R = NH2
quinazolinone core
Keywords: VX-745; Cellular and functional activity; Pyrimido-
pyrimidone.
*
Corresponding author.
ravi_natarajan@merck.com
Tel.:
+1
732
5940939; e-mail:
0960-894X/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.05.045

S. R. Natarajan et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4400–4404
4401
O
O
O
a
OH
OEt
Cl
Cl
N
Cl
Cl
Cl
Cl
O
N
R
X
O
N
R
X
11
12
HN
N
HN
N
b
HO
N
S
O
O
N
c
Figure 2. Pyrimido[4,5-d]pyrimidone and pyrido[4,3-d]pyrimidones as
heterocyclic versions of quinazolinone p38 inhibitors.
OEt
Cl
Me
Cl
14
13
did not display significant binding affinity to either the
potassium or calcium ion channels. The use of heterocy-
clic templates was explored in the hopes of dialing out
ion channel activities while maintaining p38 inhibitory
properties. This communication describes synthetic
routes to the pyrimido-pyrimidinone- and pyrido-pyri-
midinone-based p38 inhibitors along with a preliminary
assessment of their drug profiles ( Fig. 2).
d
Cl
N
S
Cl
N
S
H
N
N
N
e
PMB
Cl
Cl
15
16
The synthesis of pyrimido-pyrimidone scaffold is shown
in Scheme 1. Masamune homologation of 2-chloroben-
zoic acid gave b-keto ester 12, which was monoalkylated
with methyl iodide to give 13 in meager yield. Com-
pound 13 could also be synthesized by carbo-ethoxyla-
tion of 2-chlorophenyl ethyl ketone. Pyrimidine 14 was
assembled by the condensation of 13 with thiourea fol-
lowed by selective S-alkylation. The cyclization precur-
sor 17 was obtained in three steps from 14 by
chlorination with POCl3, followed by benzylic bromina-
tion, displacement of bromide with PMB amine, and
urea formation by treatment with 2,6-dichloro phenyl
isocyanate. The key cyclization step was efficient, yield-
ing 18 in 85% yield. Removal of PMB group followed by
oxidation of the methyl sulfide with MMPP gave desired
sulfone 20.
f
Cl
Cl
Cl
N
Cl
NH
Cl
N
S
O
N
S
g
O
N
N
N
N
PMB
R
Cl
Cl
18 : R = PMB
19 : R = H
h
17
i
For profiling the pyrimido-pyrimidine-templated p38
inhibitors the Ar-substituent chosen was either a 2-
chlorophenyl or a 2,4-dichlorophenyl group, both of
which contributed equally to potency. The analogs in
Table 1, generated by displacement of sulfone in 20,
can be loosely rank-ordered based on the impact of
C7-substituent on either overall polarity (logD’s) and/
or aqueous solubility of resulting p38 inhibitors. Com-
pound 19 with a C7 methyl sulfide substituent inhibits
p38 with good potency. However, in cell-based assays
(THP-1 cells) a 70-fold shift was observed and there
was no discernable inhibition of p38 in human whole
blood. This apparent loss in functional activity arises
due to some combination of poor aqueous solubility
and the lipophilic nature of this analog. Compound 20
with a methyl sulfone substituent at the C7-position be-
haved in a similar manner. Since both 19 and 20 were
weak blockers of the calcium (DLZ) and potassium
(iKr) ion channels. It was gratifying to show that the
pyrimido-pyrimidine scaffold in itself did not contribute
to increased ion channel binding affinities.
Cl
Cl
Cl
Cl
O2
O
N
N
S
O
N
N
R
j
HN
N
HN
N
Cl
Cl
20
Scheme 1. Synthesis of pyrimido[4,5-d]pyrimidone-based p38 inhibi-
tors. Reagents and conditions: (a) 1.1 equiv of imidazole, 2 h then
0.75 equiv of Mg(COOEtCH2COO)2, THF, rt, 10 h, 75%; (b) 1.2 equiv
of NaO-t-Bu, 2 equiv of MeI, THF, 45%; (c) i—1.1 equiv of thiourea,
2.2 equiv NaOEt, EtOH, 85 ꢁC, 2 h, 90%; ii—1.1 equiv of KOH,
1.0 equiv of MeI, H2O, 95%; (d) POCl3, 120 ꢁC, 85%; (e) i—1.1 equiv
of NBS, 0.1 equiv of Bz2O2, CCl4,95 ꢁC, 95%; (ii) 2.2 equiv of PMB-
NH2, CH2Cl2, rt, 70%; (f) 1.1 equiv of 2,6-dichloro phenyl isocyanate,
95%; (g) 1.5 equiv of K2CO3, 1.3 equiv of CuI, py, 150 ꢁC, 0.5 h, 85%;
(h) TFA, CH2Cl2, 100 ꢁC, 78%; (i) 1.5 equiv of MMPP, THF, 84%; (j)
1.5 equiv of amine, DMF, 150 ꢁC, 0.5 h.
prove, the resulting amino-pyrimidine derivatives were
more polar than 19 and 20 as judged by TLC and HPLC
retention times as well as calculated logD values. Com-
A series of amino-pyrimidines 24–29 were subsequently
investigated. Although aqueous solubility did not im-

4402
S. R. Natarajan et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4400–4404
Table 1. p38 inhibitory properties for pyrimido-pyrimidines
Cl
Cl
N
O
N
R
HN
N
Ar
Compound
R
p38b (nM)
THP-1b (nM)
WBb (nM)
Ca2+ (lM)
iKr (lM)
19
20
24
25
26
SMe
SO2Me
NH2
10.0
4.2
5.3
2.0
3.0
700
500
81
NA
NA
2000
850
6.0
>10.0
9.2
6.8
9.0
7.2
6.4
6.0
NH–Me
NMe2
22
NA
10.0
>10.0
NA
27
28
1.1
0.8
38
21
NA
10.0
>30
4.5
HN
HN
1000
10.0
H
N
29
1.9
28
500
>30
10
N
N
30
OH
4.0
2.0
170.0
97.0
NA
7.0
4.6
3.0
OMe
NMe2
NMe2
31a
270.0
>10.0
HN
HN
32a
33
5.6
4.2
20.0
21.0
2.3
32.0
250.0
100.0
42.0
1.0
8.8
1.7
0.22
0.7
7.0
0.5
10.0
3.8
O
N
N
N
34
0.5
NMe2
35a
36a
37
1.0
22.0
330.0
99.0
2.3
N
NMe2
20.0
4.0
1940.0
1000.0
0.2
N
N
6.9
HN
N
O
38
2.0
2.0
500.0
500.0
NA
NA
6.0
6.0
5.0
5.0
N
N
O
O
39a
N
40
41
42
43
1.2
0.5
0.6
0.7
34.0
2.6
1.0
2.6
360.0
17.0
15.0
38.0
1.3
2.2
3.3
1.2
5.8
8.9
6.4
4.5
HN
HN
N
HN
HN
N
NH
For all analogs Ar is 2-chlorophenyl. Those marked with ‘a’ Ar is 2,4-di-chlorophenyl.
NA, not active below 15 lM.
b Refs. 3a and 5.
pound 24 displayed excellent in vitro potency, was active
in cell-based assays losing potency only 20-fold, and
inhibited TNF-a release in LPS-stimulated human
whole blood with an IC50 of 2000 nM. It was encourag-
ing to note that the increased functional activities
among the amino-pyrimidines were not accompanied

S. R. Natarajan et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4400–4404
4403
NO2
NH2
N
with a concomitant increase in their ion-channel affini-
NO2
Br
Br
Cl
ties. In rats 24 was 100% orally bioavailable with a
half-life of 1.1 h, low clearance, and an AUC (po/
1 mpk) of 3.5 lM. Amino-pyrimidines 25–28 were simi-
larly active in vitro with excellent cellular transport
properties and appreciable functional activity in whole
blood. Compound 29 was particularly interesting on ac-
count of its good activity in human whole blood and
lack of activity against ion channels. It was disappoint-
ing to find pyridone 30 to be functionally inactive unlike
the amino-pyrimidines.
b,c
a
N
N
Cl
Cl
44
45
46
d
Br
Br
Cl
Br
Br
Cl
H
N
e,f
N
N
PMB
Compounds 31–39 represented various 1,2-diamine or
1,2-oxo-amine derivatives. Although most of these com-
pounds were potent in vitro and functionally active,
compounds 32 and 35 inhibited p38 in human whole
blood with an IC50 of 32 and 42 nM, respectively, they
were potent blockers of ion channels. The pharmacoki-
netic profile of 35 in rats was unremarkable, high clear-
ance coupled with poor absorption resulted in very low
oral bioavailability of 3%. The piperazine and morpho-
line series of analogs did not evoke enough curiosity to
warrant further work.
48
47
g
Cl
Cl
Cl
Cl
O
NH
Br
Br
Cl
O
N
Br
Cl
h
N
N
N
N
PMB
R
50 : R = PMB
51 : R = H
i
49
A series of 4-amino piperidines were also profiled. Com-
pound 42, a 4-amino piperidine capped with an iso-pro-
pyl group, showed excellent potency in whole blood at
the same time offering greater than 100-fold safety win-
SnMe3
j
N
Boc
55
`
dow vis-a-vis affinity toward ion channels. Compound
Boc
42 exhibited modest pharmacokinetic properties when
dosed in rats with a t1/2 of 2 h, moderate clearance,
and 24% oral bioavailability. The synthesis of 40, 41,
and 43 was motivated by the need to attenuate piperi-
dine metabolism and further improve PK profiles. How-
ever, 40 did not display adequate functional potency,
while 41 and 43 though equally active did not withstand
metabolism anymore than 42.
Cl
Cl
Cl
Cl
N
NH
O
N
O
N
k,l
HN
N
HN
N
Cl
Cl
53
52
m
The comparator to benchmark quinazolinone com-
pound 10 in the pyrido-pyrimidone series was accessed
via a synthetic route outlined in Scheme 2. Suzuki
coupling on 444 with 2-chlorophenyl boronic acid gave
biaryl 45. Reduction of the nitro group in 45 followed
by ortho dibromination and deamination gave 47. The
scaffold core 51 was then accomplished in a parallel
fashion to the one described for the pyrimido-pyrimi-
done core in Scheme 1. That the bromide 51 was not
functionally active came as no surprise. Compound 51
was coupled to stannane derivative 55 in a Stille reac-
tion, followed by reduction of the double bond,
deprotection of the Boc group, and reductive amina-
tion on the liberated piperidine with acetone gave
comparator compound 54. While 10 inhibited p38
more potently than 54 in vitro, they were equipotent
in inhibiting TNF-a in human whole blood. 54 dis-
played a remarkably similar pharmacokinetic profile
as 10 in rats. Thus, 54 was 18% orally bioavailable
with a t1/2 of 2.1 h, AUC (po/1 mpk) of 365 nM,
and relatively high clearances. Compound 54 dis-
played significantly better ion channel activity profile
and it is thought that this advantage accrued over qui-
nazolinone 10 is attributed to the heterocyclic tem-
plate and hence generally useful.
Cl
O
Cl
N
N
HN
N
Cl
54
Scheme 2. Synthesis of pyrido[4,3-d]pyrimidone-based p38 inhibitors.
Reagents and conditions: (a) 1.1 equiv of 2-chlorophenyl boronic acid,
2.5 equiv of Cs2CO3, toluene/MeOH/H2O (8:1:1), 0.02 equiv of
Pd(PPh3)4, reflux, 12 h, 88%; (b) Raney Ni, MeOH, rt, 7 h, 85%; (c)
2.1 equiv of Br2, THF/1 M HCl (1:2), 2 h, 85%; (d) 1.65 equiv of
iamylONO, THF, reflux, 1 h, 80%; (e) 1.1 equiv of NBS, 0.1 equiv of
Bz2O2, CCl4, 95 ꢁC, 85%; (f) 2.2 equiv of PMB-NH2, CH2Cl2, rt, 75%;
(g) 1.1 equiv of 2,6-dichloro phenyl isocyanate, 95%; (h) 1.5 equiv of
K2CO3, 1.3 equiv of CuI, py, 150 ꢁC, 0.5 h, 89%; (i) TFA, CH2Cl2,
100 ꢁC, 75%; (j) 1.1 equiv of 56, 0.04 equiv of Pd(PPh3)4, DMF,
110 ꢁC, 85%; (k) H2, Pt2O (10%), 0.15 h, 20 psi, 65%; (l) TFA, rt, 0.5 h,
95%; (m) 2.5 equiv of Me2CO, 1.5 equiv of NaCNBH3, MeOH, rt, 5 h,
76%.
In conclusion, the pyrimido-pyrimidones and pyrido-
pyrimidones were found to be viable platforms
for accessing potent p38 inhibitors. Among the

4404
S. R. Natarajan et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4400–4404
pyrimido-pyrimidone analogs profiled in Table 1 two
sub-types evoke particular interest for follow-up dis-
covery. Amino-pyrimidines such as 24 displayed a use-
ful compromise between their p38 inhibitory
properties yet poor ion channel affinities. There clearly
exists an opportunity to optimize the potency without
concomitantly increasing ion channel affinities and
thereby identify a safe clinical compound. The ami-
no-piperidine derivative such as 42, representing the
second sub-type, was very potent in human whole
blood, offered a slightly diminished safety margin
References and notes
1. Bemis, G. W.; Salituro, F. G.; Duffy, J. P.; Harrington, E.
M. US Patent 6,147,080, 2000.
2. Fitzgerald, C. E.; Patel, S. B.; Becker, J. W.; Cameron, P.
M.; Zaller, D.; Pikounis, V. B.; O’Keefe, S. J.; Scapin, G.
Nat. Struct. Biol. 2003, 10, 764.
3. (a) Natarajan, S. R.; Wisnoski, D. D.; Singh, S. B.;
Stelmach, J. E.; O’Neill, E. A.; Schwartz, C. D.; Thompson,
C. M.; Fitzgerald, C. E.; O’Keefe, S. J.; Kumar, S.; Hop, C.
E. C. A.; Zaller, D. M.; Schmatz, D. M.; Doherty, J. B.
Bioorg. Med. Chem. Lett. 2003, 13, 273, and references
within; (b) Stelmach, J. E.; Liu, L.; Patel, S. B.; Pivnichny,
J. V.; Scapin, G.; Singh, S. B.; Hop, C. E. C. A.; Wang, Z.;
Strauss, J. R.; Cameron, P. M.; Nichols, E. A.; O’Keefe, S.
J.; O’neill, E. A.; Schmatz, D. M.; Schwartz, C. D.;
Thompson, C. M.; Zaller, D. M.; Doherty, J. B. Bioorg.
Med. Chem. Lett. 2003, 13, 277.
4. Hawkins, G. F.; Roe, A. J. Org. Chem. 1949, 14, 328.
5. Anti human TNF-a was coated on immulon four plates.
THP-1 cells (density = 2.5 · 106/mL) were suspended into
96-well plates containing a PBS-based medium. Compound
was added as solution in DMSO followed by addition of
LPS. The reaction was incubated for 4 h at 37 ꢁC under
CO2. TNF-a release was measured in the supernatants by
ELISA. Reported IC50s are means from three
measurements.
`
(100-fold) vis-a-vis ion channels. It is reasoned that
modulating the basicity of the piperidine would be
one way to decrease affinities of these compounds
for ion channels, while maintaining good p38-medi-
ated therapeutic activity. Compound 54, comparator
to benchmark quinazolinone compound 9, was found
to be equipotent, orally bioavailable, and displayed
much weaker ion channel affinities. In general, the
heterocyclic templated p38 inhibitors match quinazoli-
nones in terms of potency and functional activities
while offering favorable ion channel profiles. The
impetus provided by these findings has directly
impacted the goal of finding suitable pre-clinical can-
didates and will be reported.