Discovery of aminoheterocycles as potent and brain penetrant prolylcarboxypeptidase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2011.12.098

Efforts were dedicated to develop potent and brain penetrant prolylcarboxypeptidase (PrCP) inhibitors by replacing the amide group of original leads 1 and 2 with heterocycles. Aminopyrimidines including compound 32a were identified to display good PrCP inhibitory activity (32a, IC ₅₀ = 43 nM) and impressive ability to penetrate brain in mice (brain/plasma ratio: 1.4).

Full text of DOI:10.1016/j.bmcl.2011.12.098

Bioorganic & Medicinal Chemistry Letters 22 (2012) 1727–1730
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of aminoheterocycles as potent and brain penetrant
prolylcarboxypeptidase inhibitors
Zhicai Wu a, , Cangming Yang a, Thomas H. Graham a, Andreas Verras a, Renee M. Chabin b, Suoyu Xu c,
⇑
Xinchun Tong c, Dan Xie d, Mike E. Lassman d, Urmi R. Bhatt d, Margarita M. Garcia-Calvo d, Zhu Shen d,
Qing Chen d, Kelly Bleasby c, Ranabir Sinharoy d, Jeffrey J. Hale a, James R. Tata a, Shirly Pinto d,
Steven L. Colletti a, Dong-Ming Shen a
a Department of Discovery Chemistry, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065-0900, USA
b Department of Research Operations, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065-0900, USA
c Department of Drug Metabolism, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065-0900, USA
d Department of Diabetes, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065-0900, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Efforts were dedicated to develop potent and brain penetrant prolylcarboxypeptidase (PrCP) inhibitors by
replacing the amide group of original leads 1 and 2 with heterocycles. Aminopyrimidines including com-
pound 32a were identified to display good PrCP inhibitory activity (32a, IC50 = 43 nM) and impressive
ability to penetrate brain in mice (brain/plasma ratio: 1.4).
Received 26 October 2011
Revised 16 December 2011
Accepted 20 December 2011
Available online 8 January 2012
Ó 2012 Elsevier Ltd. All rights reserved.
Keywords:
Prolylcarboxypeptidase
PrCP
Inhibitors
Aminoheterocycles
Prolylcarboxypeptidase (PrCP) is a member of serine proteases
that specifically cleaves C-terminal amino acid residue connected
to proline in peptides such as angiotensin II (AngII), angiotensin
III (AngIII) and des-Arg9-bradykinin.1,2 PrCP is widely expressed
in many tissues such as lung, liver, pancreas, kidney and brain.3
PrCP was reported to play an important role in regulating melano-
cortin signaling and thus believed to be an important component
of weight maintenance.4–7 Indeed it was also reported that
PrCPꢀ/ꢀ mice demonstrated reduced body weight and food intake,
as well as fat mass reduction versus the control WT mice.4 These
findings prompted us to validate PrCP as a potential target for
obesity treatment with small molecule PrCP inhibitors.8 Several
series of potent PrCP inhibitors have been reported. Prolinyl com-
pound 1 (Chart 1) demonstrated only a moderate body weight loss
in WT mice versus PrCP KO mice.8a Another compound 2 showed
no mechanism-based body weight loss in mice.8b Since both 1
and 2 are peripheral PrCP inhibitors with limited brain penetration
(brain/plasma concentration ratio 0.138a and 0.18,8b respectively),
it is hypothesized that central PrCP inhibition might be needed
to achieve greater efficacy. We communicate here the initial effort
towards developing potent and brain penetrant PrCP inhibitors.8c
It is well known that an amide group often increases the sus-
ceptibility of a compound to be a P-gp substrate.9 To discover
potent PrCP inhibitors with ability to penetrate the brain, our ef-
forts focused on replacing the amide bond in compound 1 and 2
with a heterocycle. First, when a pyridine ring was introduced into
the scaffold represented by compound 1, it is encouraging that a
simple pyrrolidinyl pyridine 3 (Chart 2) showed 45% inhibition
against PrCP at 5 l
M.8a However, modifications around the pyri-
dine ring (compounds 4–6) resulted in only moderate increases
of PrCP inhibitory activity. Additional incorporation of an alpha
amino isobutyrate (AIB) (7 and 8) did not boost PrCP potency as
it did in 1. On the other hand, replacement of the amide in 2
with a pyrimidine ring gave more promising results. A very easily
Cl
Cl
Cl
Cl
Me
Me
N
O
N
N
O
N
N
NH
NH
N
HN
N
O
Me
H2N
N
2
1
Me
h-PrCP IC50 = 1.1 nM
m-AngIII IC50 = 6.7 nM
h-PrCP IC50 = 1 nM
m-AngIII IC50 = 25 nM
⇑
Corresponding author.
E-mail address: zhicai_wu@merck.com (Z. Wu).
Chart 1.
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.12.098

1728
Z. Wu et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1727–1730
Cl
the purported S3 site (compound 9) caused an unexpected significant
R
Cl
Cl
loss of PrCP activity. We therefore concentrated on optimization
compound 10 for potent and brain penetrant PrCP inhibitors.
Optimization of left hand side amine was carried out in a sys-
tematic fashion. Shown in Table 1 are the results of selected com-
pounds. Piperidine (15), cyclohexylamine (14) and pyrrolidine (20)
Cl
N
N
N
N
NH
NH
N
H2N
N
CF3
3
4
, R = o-Ph, h-PrCP IC50 = 8264 nM
5, R = m-Ph, h-PrCP IC50 = 1524 nM
6, R = p-Ph, h-PrCP IC50 = 828 nM
h-PrCP 46% inh. @ 5 µM
provided h-PrCP IC50 of less than 1 lM. Additional carbon (13) or
replacement of the carbocycle with phenyl group (11 and 12) led
to decreases of their PrCP activities. Aromatic substitution at posi-
tion 3 of piperidine ring (17) increased its PrCP potency threefold.
Further modification of the phenyl group quickly showed that para
trifluoromethyl substitution generated a compound (19) with h-
PrCP IC50 of 28 nM. The corresponding meta- or ortho-substituted
18 and 17 were inferior with regard to their PrCP activity. Unfortu-
nately, theses compounds including 19 were not active in the
mouse whole plasma AngIII assay,10 most likely due to their large
serum shifts.
To reduce serum shift, polar groups were introduced into the
left hand amine and the results of selected compounds are shown
in Table 2. Replacement of piperidine with morpholine (21 and 22)
did not improve the activity in mouse AngIII assay. The comparable
PrCP activity of 21 and 22 indicated the chirality of the stereogenic
center of the left side amine had little impact on its binding to PrCP
enzyme. Incorporation of a para-methoxy group onto the phenyl
ring (23) enhanced h-PrCP activity 10-fold. Also, 23 displayed a
m-PrCP activity (IC50 46 nM). It was disappointing to find that this
Cl
Cl
Cl
Cl
F3C
O
N
N
NH
NH
N
N
N
N
N
N
N
R
NH
Bn
R
7
, R = H, h-PrCP IC50 = 3623 nM
9
, R = NH2, h-PrCP IC50 = 3882 nM
O
10, R = H, h-PrCP IC50 = 575 nM
m-PrCP IC50 = 413 nM
Me
NH2
h-PrCP IC50 = 2553 nM
8, R =
Me
m-AngIII IC50 >20,000 nM
P-gp efflux: h-BA/AB =0.3
Chart 2.
accessible compound 10 displayed an IC50 around 500 nM against
both human and mouse PrCP.8a More importantly, 10 is not a human
MDR1 P-gp substrate as shown by its BA/AB efflux ratio across LLC-
MDR1 cell monolayer of 0.3, compared to 8 and 41.5 for 1 and 2,
respectively. Again, further introduction of an amino group into
Table 2
Table 1
SAR of left side amines
SAR on human PrCP inhibitory activity of left side amines
Cl
Cl
Cl
Cl
N
N
NH
N
N
NH
N
N
R
N
R
N
Compound R
h-PrCP
m-PrCP
m-AngIII
h-
Compound
R
h-PrCP IC50 (nM)
3449
IC50 (nM) IC50 (nM) IC50 (nM) BA/
AB
N
H
11
21
22
286
241
—
—
>20,000
>20,000
1.1
0.2
N
N
12
13
14
15
2290
>10,000
839
N
H
O
O
N
H
MeO
N
H
23
21
46
9000
1
N
N
O
509
N
N
N
N
24
25
32
91
206
—
>20,000
>20,000
1.6
—
(dr =1:1)
16
537
N
CF3
N
diastereometric
ratio (dr)
(dr = 1:1)
17
18
19
20
190
91
N
(1:1)
N
26
27
58
42
—
—
>20,000
>20,000
—
—
CF3
HO
(dr = 1:1)
Me
HO
Me
F3C
F3C
N
N
(dr = 1:1)
(dr = 1:1)
N
Diastereomer
28
N
28a:
(dr = 1:1)
28
28a: 48
28b: 54
10
18
345
355
3.1
—
N
N
28b:
N
333
29

Z. Wu et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1727–1730
1729
compound only gave an IC50 of 9
l
M in mouse whole plasma AngIII
(33, IC50 1704 nM) decreases the activity significantly. Methylation
of the nitrogen (36, IC50 87 nM) slightly increases PrCP activity.
More potent compounds 32 and 36 were both separated to give
two diastereomers. With the newly found right side amines, the
chirality of the left side piperidine now affects the binding to PrCP.
For both 32 and 36, one diastereomer is about 8-fold more active
than the other one for h-PrCP and about threefold more active
for m-PrCP. Clearly, methylation of the nitrogen increases signifi-
cantly the serum shift by comparing the mouse AngIII activity of
32 and 36. Compounds 32a and 32b exhibited IC50’s of around
400 nM in the AngIII assay. But the methylated 36a and 36b gave
IC50’s of more than 2000 nM.
Next we combined the optimized right side 4-chlorophenyle-
thanamine (37) as shown in compound 32 and the potent left side
pieces. Generally the SAR (Table 4) with 37 on the right side is sim-
ilar to that with the pyrrolidine benzimidazole on the right side as
shown in Tables 1 and 2. The PrCP activities of these two series are
also comparable as represented by 28 (h-PrCP IC50 ꢁ50 nM) and 38
(h-PrCP IC50 44 nM) which has the same amine (29) on the right
side. It seems the h-BA/AB value is slightly reduced with the new
37 on the right (38: 1.6 vs 28: 3.1) and the resulting compounds
might be more brain penetrant. With fixed 37 on the right side,
we further investigated other amines on the left side with similar
space mapping to 29 but lower molecular weight. Trans-4-phen-
ylcyclohexyl amine gave compound 39 with balanced h-PrCP
(IC50 52 nM) and m-PrCP (IC50 75 nM) activity. Replacement of
the phenyl with polar methylcarbamate (40) and methanesulfon-
amide (41) increases the activity significantly. Compound 40 dis-
played an excellent potency against h-PrCP (IC50 6 nM) and m-
PrCP (IC50 26 nM). In the whole plasma m-AngIII assay, it is also
very active (IC50 159 nM).
assay. Introduction of other small polar groups including pyrazole
(24), pyridine (25) and phenylhydroxy (26) gave h-PrCP activity
with IC50 <100 nM but failed to achieve comparable activity
against m-PrCP in the presence of whole plasma. It is worth noting
that a very small polar amine hydroxyisopropyl piperidine pro-
duced a compound (27) with h-PrCP IC50 about 40 nM albeit with-
out much activity in mouse AngIII assay. Also, substituted
isoquinolinylpiperidines were found to provide potent PrCP inhib-
itors. Amine 29 (pyridylisoquinolinylpiperidine) afforded isomeric
compounds 28a and 28b with h-PrCP IC50 about 50 nM, m-PrCP
IC50 less than 20 nM and also comparable IC50 in mouse AngIII as-
say (IC50 ꢁ350 nM). Again the stereogenic center of the amine 29 is
not important for the compounds’ PrCP activities. Moreover, it is
very encouraging that all compounds tested in this series, includ-
ing 21, 22, 23, 24, and 28, are not substrates of human P-gp.
We then turned our attention to pyrrolidinyl benzimidazole
portion. Optimization of the right side amine was carried out with
fixed hydroxypropylpiperidine on the left side as in compound 27.
Inspired by our in-house effort of optimizing the pyrroloidinyl
benzimidazole part in compound 1 series,8c different phenylethyl-
amines were put on the right side of the pyrimidine core. Shown in
Table 3 are results of selected compounds. The PrCP activity is sen-
sitive to the substitution at the benzylic position. The unsubstitut-
ed benzylic compound 34 and gem-methyl substituted compound
35 are not very active while the monomethyl substituted 32 is
much more active. The chirality of this benzylic stereogenic center
is also critical. Compound 30 with S configuration was around 50-
fold more potent than compound 31 with R configuration. It seems
the size of the substitution group on the phenyl ring plays an
important role on the binding to PrCP. The smaller group fluorine
Potent PrCP inhibitors with encouraging P-gp properties were
tested in CF1 and wild type mice to assess their ability to penetrate
the brain. Data for selected compounds are shown in Table 5. It
seemed that the brain penetration correlated qualitatively with
the m-BA/AB values. Compound 28a unexpectedly turned out to
be a mouse Pgp substrate (m-BA/AB: 5.9) and the brain/plasma
concentration ratio was only 0.29. Compounds 21 and 32a were
Table 3
SAR of right side amines
Me
N
N
HO
Me
N
R
Compound
R
h-PrCP
IC50 (nM)
m-PrCP
IC50 (nM)
m-AngIII
IC50 (nM)
Table 4
Data of selected compounds
Me
Me
Me
(dr = 1:1)
Me
N
H
30
31
130
N
N
Me
R
N
H
(dr = 1:1)
Me
N
H
7922
181
Cl
Compound
R
h-PrCP
IC50
(nM)
m-PrCP
IC50 (nM) IC50 (nM)
m-AngIII
h-
BA/
AB
N
H
32
32a:
Cl
32b:
N
32a: 43
32b: 350
176
570
350
407
(dr = 1:1)
38
39
44
52
—
—
—
1.6
—
Me
N
N
N
H
33
34
35
1704
5541
1024
F
N
H
75
Cl
N
H
Me Me
H
N
N
MeO
Me
H
Cl
40
41
6
26
159
—
—
—
O
N
H
Me
N
36
87
36a:
36b:
H
Me
N
Cl
S
15
105
O
O
36a: 31
36b: 223
146
797
2210
5000
N
H

1730
Z. Wu et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1727–1730
Table 5
with a pyrimidine ring. This modification reduced the potential
Brain penetration of selected compoundsa
of these compounds as P-gp substrates and enhanced their brain
penetration. Compound 32a with a reasonable molecular weight
of 375 displayed very good human PrCP inhibitory activity
(IC50 = 43 nM) and impressive ability to penetrate the brain in mice
(brain/plasma ratio: 1.4). Further efforts to maintain the brain pen-
etration and improve h-PrCP and m-PrCP activities as well as over-
all profile to identify a proof of concept molecule will be reported
in due course.
Compound
m-BA/ABb
CF-1 WT
BrainðnMÞ
Ratio
0.79
PlasmaðnMÞ
Cl
Cl
N
242ꢂ53
304ꢂ34
NH
1.9
N
N
N
N
O
References and Notes
21
Cl
N
Cl
1. (a) Yang, H. Y. T.; Erdos, E. G.; Chiang, T. S. Nature 1968, 218, 1224; (b) Yang, H.
Y. T.; Erdos, E. G.; Chiang, T. S.; Jenssen, T. A.; Rodgers, J. G. Biochem. Pharmcol.
1970, 19, 1201; (c) Odya, C. E.; Marinkovic, D. V.; Hammon, K. J.; Stewart, T. A.;
Erdos, E. G. J. Biol. Chem. 1978, 253, 5927; (d) Skidgel, R. A.; Erdos, E. G. Immun.
Rev. 1998, 161, 129.
2. (a) Shariat-Madar, Z.; Mhdi, F.; Schmaier, A. H. J. Biol. Chem. 2002, 277, 17962;
(b) Moreira, C. R.; Schmaier, A. H.; Mahdi, F.; da Motta, G.; Nader, H. B.; Shariat-
Madar, Z. FEBS Lett. 2002, 523, 167; (c) Shariat-Madar, Z.; Mhdi, F.; Schmaier, A.
H. Blood 2004, 103, 4554.
3. (a) Kumamoto, K.; Stewart, T. A.; Johnson, A. R.; Erdos, E. G. J. Clin. Invest. 1981,
67, 210; (b) Skidgel, R. A.; Wickstrom, E.; Kumamoto, K.; Erdos, E. G. Anal.
Biochem. 1981, 13, 196.
4. Wallingford, N.; Perround, B.; Gao, Q.; Coppola, A.; Gyengesi, E.; Liu, Z. W.; Gao,
X. B.; Diament, A.; Haus, K. A.; Shariat-Madar, Z.; Hahdi, F.; Wardlaw, S. L.;
Schmaier, A. H.; Warden, C. H.; Diano, S. J. Clin. Invest. 2009, 119, 2291.
5. (a) Mallela, J.; Yang, J.; Shariat-Madar, Z. Int. J. Biochem. Cell Biol. 2009, 41, 477;
(b) Wang, L.; Feny, Y.; Zhang, Y.; Zhou, H.; Jiang, S.; Niu, T.; Wei, J. L.; Xu, X.; Xu,
X. P.; Wang, X. Am. J. Obstet. Gynecol. 2006, 195, 162.
N
NH
108ꢂ16
367ꢂ28
N
N
N
5.9
0.29
1.4
N
N
N
28a
Me
N
Me
HO
Me
N
N
H
2.8c
400ꢂ35
291ꢂ19
Cl
32a
a
b
c
1 mpk IV dose. Matrix concentration measured at 4 h post-dose.
Mouse LLC-MDR1a cell line.
24 h following 30 mpk SC mini pump administration.
6. Shariat-Madar, B.; Kolte, D.; Verlangiert, A.; Shariat-Madar, Z. Diabetes, Metab.
Syndr. Obes.: Targets Ther. 2010, 3, 67.
7. Diament, A. L.; Warden, C. H. Int. J. Obes. 2004, 28, 199.
Me
N
N
Me
8. (a) Zhou, C.; Garcia-Calvo, M.; Pinto, S.; Lombardo, M.; Feng, Z.; Bender, K.;
Pryor, K. D.; Bhatt, U. R.; Chabin, R. M.; Geissler, W. M.; Shen, Z.; Tong, X.;
Zhang, Z.; Wong, K. K.; Roy, R. S.; Chapman, K. T.; Yang, L.; Xiong, Y. J. Med.
Chem. 2010, 53, 7251; (b) Shen, H. C.; Ding, F.-X.; Zhou, C.; Xiong, Y.; Verras, A.;
Chabin, R. M.; Xu, S.; Tong, X.; Xie, D.; Lassman, M. E.; Bhatt, U. R.; Garcia-Calvo,
M.; Geissler, W. M.; Shen, Z.; Chen, D.; Sinha, R. R.; Hale, J. J.; Tata, J. R.; Pinto, S.;
Shen, D.; Colletti, S. L. Bioorg. Med. Chem. Lett. 2011, 21, 1299; (c) Graham, T.;
Shen, H. C; Liu, W.; Xiong, Y.; Verras, A.; Bleasby, K.; Bhatt, U. R.; Chabin, R. M.;
Chen, D.; Chen, Q.; Garcia-Calvo, M.; Geissler, W. M.; He, H.; Lassman, M. E.;
Shen, Z.; Tong, X.; Xie, D.; Xu, S.; Colletti, S. L.; Hale, J.; Tata, J. R.; Pinto, S.; Shen,
D. Bioorg. Med. Chem. Lett. 2012, 22, 659.
N
N
a
+
H2N
Cl
N
H
Cl
Cl
Cl
Cl
44
42
43
Me
HO
Me
Me
N
N
Me
NH
HO
Me
N
N
H
45
Cl
b
32a, 32b (1:1)
9. (a) Seelig, A. Eur. J. Biochem. 1998, 251, 252; (b) Gatlik-Landwojtowicz, E.;
Aanismaa, P.; Seelig, A. Biochemistry 2006, 45, 3020; (c) Turunen, B. J.; Ge, H.;
Oyetunji, J.; Desino, K. E.; Vasandani, V.; Guthe, S.; Himes, R. H.; Audus, K. L.;
Seelig, A.; Georg, G. I. Bioorg. Med. Chem. Lett. 2008, 18, 5971; (d) Lawrence, D.
S.; Copper, J. E.; Smith, C. D. J. Med. Chem. 2001, 44, 594.
Scheme 1. Reagents and conditions: (a) Et3N, CH3CN, 75 °C, 2 h, 74%; (b) Et3N,
CH3CN, 170 °C, 5 h, 95%.
10. Mouse whole plasma AngIII assay measures PrCP activity/inhibition in mouse
plasma. The enzyme assay described in Ref. 8a utilizes a FRET substrate (Mca-
Ala-Lys-Dnp) that is incompatible with plasma due to high protein binding.
Therefore LC/MS was used to monitor the cleavage of exogenous angiotensin III
(AngIII, RVYIHP;F) by endogenous plasma PrCP. Since peptide substrates are
susceptible to cleavage by multiple proteases in plasma, proteases cocktail was
added and assay pH was optimized to pH 5.5 to minimize background
proteases activity so that the rate of AngIII cleavage is proportional to PrCP
activity. In detail, Experiments were carried out at 37 °C under standard
not mouse Pgp substrates (m-BA/AB <3) and they penetrated the
brain very effectively with brain concentration 242 and 400 nM.
Their corresponding brain/plasma ratios were 0.79 and 1.4,
respectively.
The synthesis of the representative compound 32 is illustrated
in Scheme 1. Reaction of dichloropyrimidine 42 with phenylethyl-
amine 43 provided aminopyridine 44, which reacted further with
piperidine 45 to afford diastereomers 32a and 32b after chiral
separation.
reaction conditions defined as 48
ammonium acetate, 3 L of protease inhibitor cocktail and 4
for 8 min. Reaction was stopped with 100
l
L
of DIO mouse plasma,
3
l
l
L
of 5 M
l
l
L of 50
M AngIII
l
L of 8 M urea. Peptides were
extracted and then analyzed with LC/MS. IC50 values are reported as single
determinations or as the average of at least two determinations.
In summary, a class of potent PrCP inhibitors has been discov-
ered by replacing the amide group of the original leads 1 and 2