Optimization of 2,4-diaminopyrimidines as GHS-R antagonists: Side chain exploration

Article abstract of DOI:10.1016/j.bmcl.2006.01.012

The synthesis and structure-activity relationships of the 4- and 6-substituents of 2,4-diaminopyrimidine-based growth hormone secretagogue receptor (GHS-R) antagonists are described. Diaminopyrimidines with 6-norbornenyl (4n) and 6-tetrahydrofuranyl (4p) substitutents were found to exhibit potent GHS-R antagonism and good selectivity (～1000-fold) against dihydrofolate reductase.

Full text of DOI:10.1016/j.bmcl.2006.01.012

Bioorganic & Medicinal Chemistry Letters 16 (2006) 1864–1868
Optimization of 2,4-diaminopyrimidines as GHS-R
antagonists: Side chain exploration
Bo Liu, Mei Liu, Zhili Xin, Hongyu Zhao, Michael D. Serby, Christi Kosogof,
Lissa T. J. Nelson, Bruce G. Szczepankiewicz, Wiweka Kaszubska, Verlyn G. Schaefer,
H. Douglas Falls, Chun Wel Lin, Christine A. Collins, Hing L. Sham and Gang Liu*
Metabolic Disease Research, Global Pharmaceutical Research and Development, Abbott Laboratories,
Abbott Park, IL 60064-6098, USA
Received 16 November 2005; revised 27 December 2005; accepted 4 January 2006
Available online 25 January 2006
Abstract—The synthesis and structure–activity relationships of the 4- and 6-substituents of 2,4-diaminopyrimidine-based growth
hormone secretagogue receptor (GHS-R) antagonists are described. Diaminopyrimidines with 6-norbornenyl (4n) and 6-tetrahy-
drofuranyl (4p) substitutents were found to exhibit potent GHS-R antagonism and good selectivity (ꢀ1000-fold) against dihydro-
folate reductase.
Ó 2006 Elsevier Ltd. All rights reserved.
R'
Obesity has become an epidemic disease in developed
countries and is also increasing at an alarming rate in
1 (R = -CH2OBn, R' = Cl)
IC50s = 12 nM (Binding), 170 nM (FLIPR)
Ki = 42 nM (FLIPR)
developing nations.1 Existing FDA approved anti-obes-
NH
F (i.p.) = 68% (rats)
NH2
ity drugs have only moderate efficacy.2 Ghrelin, a 28
N
2 (R = Et, R' = -SO2Me)
IC50s = 9.7 nM (Binding), 37 nM (FLIPR)
F (p.o.) = 52% (rats)
amino acid peptide with a unique n-octanoyl modifica-
tion on Ser 3, has attracted tremendous attention since
its discovery as an endogenous ligand for the growth
hormone secretagogue receptor (GHS-R).3 Besides
being a potent growth hormone secretagogue and regu-
lator of other endocrine functions, ghrelin is also impli-
cated in the short and long-term regulation of energy
balance.4 A small molecule GHS-R antagonist could
potentially be a novel anti-obesity therapy.
H2N
R
N
Figure 1. Representative first-generation 2,4-diaminopyrimidine-based
GHS-R antagonists.
ture–activity relationship (SAR) studies on the amin-
obenzyl group of 1 revealed that certain electron-
withdrawing groups could provide activity improvement,
and good oral bioavailability, as exemplified by 2
(Fig. 1).7
Recently, we reported the discovery of novel isoxazole
carboxamides5 and tetralin carboxamides6 as potent
and selective human GHS-R antagonists. Continuing
the pursuit of additional pharmacologically active
GHS-R antagonists, we identified 2,4-diaminopyrimi-
dine derivative 1 as a potent, competitive, intraperitone-
ally bioavailable small molecule GHS-R antagonist with
in vivo efficacy for decreasing food intake and body
weight in several rat studies (Fig. 1).7 Extensive struc-
Unfortunately, compounds 1 and 2 were also found to
be potent human dihydrofolate reductase (DHFR)
inhibitors,7 consistent with the fact that 2,4-diaminopyr-
imidine is a privileged pharmacophore for DHFR inhib-
itors, such as trimethoprim and pyrimethamine.8 So our
SAR efforts focused on achieving greater DHFR selec-
tivity via pyrimidine core replacement with other hetero-
cycles. Unfortunately, any deviation from the
pyrimidine template, such as 2,6-diaminopyridine, 3-
aminopyrazole, 3,5-diamino[1,2,4]triazine, etc., led to
complete abolishment of GHS-R activity (data not
shown). In the meantime, modification of the 2-amino
Keywords: Ghrelin; GHS-R antagonist; Diaminopyrimidine; DHFR
selectivity.
*
Corresponding author. Tel.: +1 847 935 1224; fax: +1 847 938 1674;
e-mail: gang.liu@abbott.com
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.01.012

B. Liu et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1864–1868
1865
NO2
group of 2 to other groups (hydroxyl, methylamino,
NO2
NO2
NO2
d
O
O
methyl, etc.) also resulted in inactive analogs (data not
shown). In this letter, we will describe the SAR studies
on the diaminopyrimidine 4- and 6-position side chain
modification to achieve high ghrelin antagonist potency
and good DHFR selectivity.
a
b
c
O
O
HO
HO
O
O
O
O
N
Br
O
Br
9
10
11
O
O
S
NO2
e,f
O
H
NO2
g,h,i
R1 R2
N
O
The diaminopyrimidine 6-position side chain modifica-
tions (4a–4t) were prepared in a similar fashion as de-
scribed previously.7 Briefly, 4-nitrophenyl-acetonitrile
was first acylated with appropriate acid chlorides
(Scheme 1). Subsequent methylation, cyclization with
guanidine, and reduction of the nitro group yielded ani-
line 3. Reductive amination of 3 with p-methylsulfonyl
benzaldehyde provided the ghrelin antagonists 4a–4t.
Carboxamides 4u and 4v were synthesized from basic
hydrolysis of ethyl ester in 4t, followed by condensing
the resulting acid with corresponding benzyl amines.
Cl
N
N
N
N
H2N
N
H2N
N
H2N
12
14a-14l
13
Scheme 3. Reagents and conditions: (a) 1—SOCl2, 2—BrCH2CH2OH;
(b) CH3CH2COCl, Et3N, 0 °C-rt, 43% over 3 steps; (c) NaH, DMF, rt;
(d) guanidine, EtONa, reflux, 70% over 2 steps; (e) TMSBr, dioxane,
150 °C; (f) POCl3, 100 °C, 85% over 2 steps; (g) HNR1R2, dioxane,
170 °C, 20 min; (h) 10% Pd/C, MeOH, rt; (i) 4-methanesulfonylbenzal-
dehyde, NaBH3CN, MeOH/AcOH/NaOAc, rt, ca. 35% over three steps.
The 4-alkylaminopyrimidine analogs 14a–14l were pre-
pared as described in Scheme 3. Condensation of
bromoethyl alcohol and an acid chloride made in situ
from p-nitrophenyl acetic acid and thionyl chloride led
to ester 9. Acylation of 9 with propionyl chloride pro-
vided enol ester 10. Treatment of 10 with NaH in
DMF produced a dienolate intermediate, which under-
went an intramolecular O-alkylation reaction followed
by a facile cleavage of the enol ester to provide ketene
acetal 11. Condensation between 11 and guanidine gave
pyrimidine 12, which was converted to 4-chloropyrimi-
dine 13 in a 2-step sequence. Reaction of chloropyrimi-
dine 13 with a variety of primary and secondary amines
gave 4-akylamino pyrimidines. Reduction of the nitro
group, followed by a reductive amination step, produced
final analogs 14a–14l.
The 6-ether side chain modifications (7a–7h) were
achieved starting from the previously described benzyl
ether 5.7 Debenzylation and bromination of 5 generated
benzyl bromide 6 (Scheme 2). A small amount of water
in the reaction mixture prevented the formation of ani-
line acetamide during this process. Williamson ether
synthesis using 6 with various alcohols provided 6-ether
analogs 7a–7h. Analogously, 6-aminoalkyl-substituted
pyrimidines 8a–8p could be synthesized through dis-
placement of benzyl bromide 6 with different amines.
NH2
NH2
a,b,c,d
e
R
Cl
N
+
O2N
O
R
H2N
N
N
3
O
O
S
S
Speculating that the 6-ethyl group of diaminopyrimidine
2 might occupy the same binding site as the octanoyl
group of ghrelin, we replaced the ethyl group of 2 with
different lipophilic chains and found that this position
is quite tolerant of different substitution. As shown in
Table 1, both straight and branched aliphatic chain
(3–8 carbons) substituents led to analogs (4a–4h) with
low nanomolar binding IC50 values comparable to those
of lead compound 2.9 Unfortunately, the longer aliphat-
ic chains (4f–4h) did not provide any FLIPR potency
boost as expected. Only 6-isopropyl diaminopyrimidine
4a was 2-fold more potent in both binding and FLIPR
assays than 2. Consistent with the beneficial effect of
a-branching, all the 6-cycloalkyl-substituted diamino-
pyrimidines (4i–4l) exhibited equal or a slightly better
FLIPR activity. In contrast, 6-bicycloalkyl substituents
led to compounds (4m–4o) with 2- to 8-fold-loss in
FLIPR potency. Further functionalization of the ali-
phatic chain provided some highly potent analogs, such
as propionyl benzyl amides 4u and 4v, both of which
have single-digit nanomolar potency in FLIPR assay
and sub-nanomolar binding IC50 for 4u.
O
O
H
N
H
N
f,g
NH2
N
NH2
N
4t as
starting
material
N
N
NHR
R
H2N
H2N
4a-4t
4u-4v
O
Scheme 1. Reagents and conditions: (a) Et3N, DMAP, CH2Cl2, rt; (b)
TMSCHN2, CH2Cl2, rt; (c) guanidine, EtOH, reflux, 45–80% over
three steps; (d) H2, Pd(OH)2, MeOH, 95%; (e) p-methanesulfonyl-
benzaldehyde, NaBH3CN, MeOH/HOAc/NaOAc, rt, 70%; (f) LiOH,
THF/MeOH, rt; (g) H2NR, TBTU, i-Pr2NEt, DMF, rt, 82% over two
steps.
O
S
O
O
O
O
O
S
S
H
H
N
H
N
N
NH2
N
NH2
N
NH2
N
a
b
N
N
N
OBn
7a-7h
Br
O
H2N
H2N
H2N
R
5
6
O
O
c
S
H
N
NH2
N
N
NR1R2
8a-8p
H2N
Previous SAR studies have illustrated the importance
of the 6-benzyloxymethyl ether in 1 for maintaining
ghrelin activity.7 Not surprisingly, this ether in
combination with 4-methylsulfonyl benzyl amine
Scheme 2. Reagents and conditions: (a) 30% HBr/AcOH with 5%
H2O, reflux, 4 h, 72–93%; (b) ROH, NaH, THF, DMPU, rt, 12–25%;
(c) HNR1R2, i-Pr2NEt, THF, rt, 25–32%.

1866
B. Liu et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1864–1868
Table 1. SAR of 6-alkyl-substituted diaminopyrimidines
Table 2. SAR of 6-cyclic ether/6-methyl ether-substituted diamino-
pyrimidines
O
S
O
O
S
O
O
O O
H
S
H
N
N
S
NH2
N
NH2
N
O
H
N
H
N
NH2
N
N
NH
N
2
N
NHR
N
H2N
N
R
H2N
4u-4v
4a-4t
7a-7h
O
O
R
4p-4s
H N
2
R
H2N
a
a
Compound
R
Binding IC50
(nM)
FLIPR IC50
(nM)
a
a
Compound
R
Binding IC50
(nM)
FLIPR IC50
(nM)
2
Ethyl
Isopropyl
n-Butyl
9.7
5.4
37
14
32
40
11
211
62
38
57
19
14
21
36
5
Not applicable
0.3
0.4
7.1
1.9
4a
4b
4c
4d
4e
4f
4g
4h
4i
11
10
NDb
7a
Isobutyl
t-Butyl
7b
7c
7d
2.8
7.5
2.7
6.8
53
n-Pentyl
n-Hexyl
19
6.0
–CH3
O
n-Heptyl
n-Octyl
6.0
9.6
5.2
Cyclopropyl
Cyclobutyl
Cyclopentyl
Cyclohexyl
NDb
NDb
NDb
NDb
O
7e
7f
3.4
38
9.7
4j
4k
4l
187
N
O
4mc
NDb
137
4p
NDb
54
(S)
(endo & exo)
O
4q
NDb
NDb
NDb
57
H
H
H
H
(R)
4nc
NDb
66
O
4rc
4sc
38
H
(exo)
O
O
180
4oc
4t
NDb
NDb
283
82
H
a Values are averages of triplicate from one assay against human
GHS-R.
b ND, not determined.
c A racemic mixture.
(endo)
O
O
4u
4v
0.4
2.0
3.0
5.4
was observed with compound 4r. The slightly larger
dihydrobenzoxine-substituted 4s fared much worse in
the functional assay.
Cl
a Values are averages of triplicate from one assay against human
GHS-R.
We also explored amino replacements for the 6-benzyl-
oxymethyl ether side chain of 5. In general, N-alkylam-
inomethyl-substituted diaminopyrimidines (8a–8g) were
substantially less potent than the ether analogs in both
binding and FLIPR, as illustrated by the direct compar-
ison between amine 8d and ether 7b (Table 3). In con-
trast, when an N-arylaminomethyl group was
introduced in the diaminopyrimidine 6-position, the
resulting compounds (8h–8k) displayed large potency
improvement. Particularly, both sec-amine 8h and tert-
amine 8i showed single-digit nanomolar IC50s in both
binding and FLIPR assays. In comparison, N-benzy-
laminomethyl substituents (8l–8m) were 5- to 10-fold
worse than the corresponding N-aryl analogs. Conform-
ationally restricted amines, such as isoindoline, indoline,
and tetrahydroisoquinoline, provided equipotent func-
tional antagonists to compound 2.
b ND, not determined.
c A racemic mixture.
moiety generated compound 5, one of the more potent
GHS-R antagonists in both binding and FLIPR.
Replacing the benzyl group of 5 with an isopentenyl
group led to 7a with sub-nanomolar binding potency
and low nanomolar functional activity (Table 2). Satu-
ration of the double bond led to 7b with a 3- to 6-fold
loss of potency, while the much shorter methyl ether 7c
was substantially weaker. Smaller furanylmethyl ethers
(7d–7e) were equipotent in the FLIPR assay to 5, but
the more basic 3-pyridylmethyl ether (7f) was much less
potent in the functional assay. In the meantime, cyclic
ethers directly linked to the 6-position of diaminopyr-
imidine core were also investigated. Non-aromatic, cyc-
lic tetrahydrofuran-substituted diaminopyrimidines 4p
and 4q exhibited comparable FLIPR activity to that
of 2, without any chirality preference. When the phenyl
ring was re-introduced to the 6-position side chain as
dihydrobenzofuran, a slightly better FLIPR activity
Previous SAR studies indicated that 2-amino group of the
diaminopyrimidine core is critical for maintaining the
intrinsic ghrelin antagonist activity. Using the chemistry
developed in Scheme 3, we were able to examine substitu-

B. Liu et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1864–1868
Table 3. SAR of 6-aminoalkyl-substituted diaminopyrimidines Table 4. SAR of diaminopyrimidine 4-amino modification
1867
O
O
O
S
S
O
H
N
H
N
R1 R2
N
NH2
N
N
N
N
NR1R2
8a-8p
H2N
H2N
14a-14l
–NR1R2
Binding IC50
(nM)
FLIPR IC50
(nM)
a
a
Compound R1
R2
Binding
FLIPR
Compound
a
IC50 (nM) IC50 (nM)
a
8a
8b
8c
–NH2
–NH(CH3)
–N(CH3)2
634
279
186
7160
1380
2580
14a
14b
14c
14d
14e
14f
14g
14h
14i
Me
Me
Et
H
Me
H
Et
H
H
H
H
H
H
H
37
192
244
450
151
276
199
354
601
700
248
3250
69
772
>10,000
380
237
Et
n-Pr
8d
8e
61
65
1420
945
N
H
n-Bu
i-Amyl
736
2480
N
H
2-Hydroxyethyl
Cyclopropyl
3080
>10,000
>10,000
354
8f
122
16
3430
316
9.1
9.0
12
14j
Ph
Bn
N
H
14k
14l
Morpholin-4-yl
>10,000
N
H
8g
8h
8i
a Values are averages of triplicate from one assay against human
GHS-R.
5.5
N
H
Table 5. DHFR selectivity of representative diaminopyrimidine
analogs
2.5
N
F
Compound FLIPR
% hDHFR % hDHFR
inhibition
8j
1.5
IC50 (nM) inhibition
at 1 lM
N
H
at 10 lM
F
NDa
96
10
2
4d
4k
4l
37
11
>99
47
>99
87
8k
8l
N
21
36
73
16
97
50
N
H
49
4m
4n
4o
4p
4q
4u
7b
7g
8i
137
66
3
21
38 (55% inh. at 30 lM)
45
NDa
NDa
NDa
NDa
82
N
8m
8n
8o
8p
4
69
283
54
25 (38% inh. at 30 lM)
NDb
NDb
NDb
34
57
84
93
97
93
84
58
82
N
N
3.0
6.8
38
9.0
41
69
88
34
13
42
29
8p
14a
41
N
41
a ND, not determined.
a Values are averages of triplicate from one assay against human
GHS-R.
b ND, not determined.
6-position variations (4d, 4k, 4u, 7b, and 8i) led to min-
imal DHFR selectivity improvement.10 However, a clear
trend for achieving DHFR selectivity emerged from
data shown in Table 5. When the diaminopyrimidine
6-position was substituted with a bulkier cyclohexyl
group (4l), moderate DHFR selectivity improvement
was observed. Further increase of the steric bulkiness
of the cyclohexane ring with bicyclic norbornane (4m)
and norbornene (4n) resulted in compounds with ca.
1000-fold separation of ghrelin and DHFR potencies.
Equally encouraging, when the diaminopyrimidine
6-position was substituted with a (S)-tetrahydrofuranyl
group, the resulting compound 4p also exhibited
ꢀ1000-fold DHFR selectivity.
tions on the 4-amino group of the diaminopyrimidine
ring. As shown in Table 4, only 4-methylaminopyrimi-
dine 14a has the same level of potency as 2 in both binding
and FLIPR. Other larger substituents (14c, 14e–14k) and
N,N-disubstitution (14b, 14d, and 14l) resulted in sub-
stantial binding and FLIPR potency decreases. An
unsubstituted amino group at 4-position of pyrimidine
ring is necessary for optimal ghrelin antagonism.
As indicated earlier, 2,4-diaminopyrimidine is a privi-
leged pharmacophore for DHFR inhibitors, and com-
pounds 1 and 2 were found to be potent DHFR
inhibitors. Therefore, the impact of diaminopyrimidine
4,6-position modification on DHFR selectivity was
examined. As shown in Table 5, many of the preferred
In summary, extensive SAR exploration has revealed
that the 2,4-diaminopyrimidine core is optimal for

1868
B. Liu et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1864–1868
6. (a) Zhao, H.; Xin, Z.; Liu, G.; Schaefer, V.; Falls, D.;
Kaszubska, W.; Collins, C.; Sham, H. J. Med. Chem.
2004, 47, 6655; (b) Zhao, H.; Xin, Z.; Patel, J. R.; Nelson,
L. T. J.; Liu, B.; Szczepankiewicz, B. G.; Schaefer, V.;
Falls, D.; Kaszubska, W.; Collins, C.; Sham, H.; Liu, G.
Bioorg. Med. Chem. Lett. 2005, 15, 1825.
7. Serby, M.; Zhao, H.; Szczepankiewicz, B. G.; Kosogof, C.;
Xin, Z.; Liu, B.; Liu, M.; Nelson, L. T. J.; Kaszubska, W.;
Falls, D.; Schaefer, V. G.; Bush, E.; Shapiro, R.; Droz, B.;
Knourek-Segel, V.; Fey, T.; Brune, M.; Beno, D.; Turner,
T.; Collins, C.; Sham, H.; Liu, G. J. Med. Chem. 2006,
manuscript submitted.
ghrelin antagonist activity, and the diaminopyrimidine
6-position is a key region for GHS-R potency improve-
ment. Most importantly, 6-norbornenyl (4n) and 6-tet-
rahydrofuranyl (4p)-substituted diaminopyrimidines
exhibited potent GHS-R antagonism with good selectiv-
ity (ꢀ1000-fold) against DHFR. Evaluation of ADME/
PK properties of these analogs and further optimization
of them in their DHFR selectivity could provide valu-
able tool compounds for elucidating the roles ghrelin
plays in obesity.11
8. Sirotnak, F. M., Burchall, J. J., Ensminger, W. B.,
Montgomery, J. A., Eds.; Folate Antagonists as Therapeu-
tic Agents; Academic Press: New York, 1984.
References and notes
9. In vitro (human GHS-R binding and FLIPR) assay were
performed using the conditions described in Ref. 5a.
10. Human DHFR inhibition was assayed colorimetrically by
following the nonenzymatic reduction of MTS (3-[4,
5-dimethylthiazol-2-yl]-5-[3-carboxymethoxyphenyl]-2-[4-
sulfophenyl-2H-tetrazolium, inner salt), by tetrahydrofo-
late, to a soluble formazan. Methotraxate was used as a
positive control in the assay.
11. Xin, Z.; Serby, M.; Zhao, H.; Christi K.; Szczepankie-
wicz, B. G.; Liu, M.; Liu, B.; Hutchins, C.; Sarris, K.;
Hoff, E.; Falls, D.; Lin, C.; Ogiela, C.; Collins, C.;
Brune, M.; Bush, E.; Droz, B.; Fey, T.; Knourek-Segel,
V.; Shapiro, R.; Jacobson, P.; Beno, D.; Turner, T.;
Sham, H.; Liu, G. J. Med. Chem. 2006, manuscript in
preparation .
1. Zimmet, P.; Alberti, K. G. M. M.; Shaw, J. Nature 2001,
41, 782.
2. Shi, Y.; Burn, P. Nat. Rev. Drug Discov. 2004, 3, 695.
3. Kojima, M.; Hosoda, H.; Date, Y.; Nakazato, M.;
Matsuo, H.; Kangawa, K. Nature 1999, 402, 656.
4. (a) Ueno, H.; Yamaguchi, H.; Kangawa, K.; Nakazato,
M. Regul. Pept. 2005, 126, 11; (b) Cummings, D. E.;
Foster-Schubert, K. E.; Overduin, J. Curr. Drug. Targets
2005, 6, 153.
5. (a) Liu, B.; Liu, G.; Xin, Z.; Serby, M.; Zhao, H.;
Schaefer, V.; Falls, D.; Kaszubska, W.; Collins, C.; Sham,
H. Bioorg. Med. Chem. Lett. 2004, 14, 5223; (b) Xin, Z.;
Zhao, H.; Serby, M.; Liu, B.; Schaefer, V.; Falls, D.;
Kaszubska, W.; Collins, C.; Sham, H.; Liu, G. Bioorg.
Med. Chem. Lett. 2005, 15, 1201.