Potent and selective proline derived dipeptidyl peptidase IV inhibitors

Article abstract of DOI:10.1016/j.bmcl.2004.07.056

In-house screening of the Merck sample collection identified proline derived homophenylalanine 3 as a DPP-IV inhibitor with modest potency (DPP-IV IC₅₀ = 1.9 μM). Optimization of 3 led to compound 37, which is among the most potent and selectiv

Full text of DOI:10.1016/j.bmcl.2004.07.056

Bioorganic & Medicinal Chemistry Letters 14 (2004) 5151–5155
Potent and selective proline derived dipeptidyl
peptidase IV inhibitors
Scott D. Edmondson,a,* Anthony Mastracchio,a Maria Beconi,b Lawrence F. Colwell, Jr.a
Bahanu Habulihaz,a Huaibing He,a Sanjeev Kumar,b Barbara Leiting,c
Kathryn A. Lyons,a Ann Mao,b Frank Marsilio,c Reshma A. Patel,c Joseph K. Wu,c
Lan Zhu,c Nancy A. Thornberry,c Ann E. Webera and Emma R. Parmeea
aDepartment of Medicinal Chemistry, Merck & Co. Inc., PO Box 2000, Rahway, NJ 07065, USA
bDepartment of Drug Metabolism, Merck & Co. Inc., PO Box 2000, Rahway, NJ 07065, USA
cDepartment of Metabolic Disorders, Merck & Co. Inc., PO Box 2000, Rahway, NJ 07065, USA
Received 28 June 2004; revised 26 July 2004; accepted 27 July 2004
Available online 23 August 2004
Abstract—In-house screening of the Merck sample collection identified proline derived homophenylalanine 3 as a DPP-IV inhibitor
with modest potency (DPP-IV IC50 = 1.9lM). Optimization of 3 led to compound 37, which is among the most potent and selective
DPP-IV inhibitors discovered to date.
Ó 2004 Elsevier Ltd. All rights reserved.
Inhibition of dipeptidyl peptidase IV (DPP-IV) has
recently emerged as a promising new approach for the
treatment of type 2 diabetes mellitus.1 DPP-IV is the
enzyme responsible for inactivation of the incretins gluc-
agon-like peptide 1 (GLP-1) and glucose-dependent
insulinotropic polypeptide (GIP). These two hormones
are secreted in response to nutrient ingestion, and each
enhances the glucose-dependent secretion of insulin.
Furthermore, GLP-1 has been shown in mammals to
stimulate insulin biosynthesis, inhibit glucagon secre-
tion, slow gastric emptying, reduce appetite, and stimu-
late the regeneration and differentiation of islet b-cells.1,2
More recently, the GLP-1 receptor has also been impli-
cated in learning and neuroprotection.3
diabetes therapies including a lowered risk of hypogly-
cemia, the potential for weight loss, and the potential
for the regeneration and differentiation of pancreatic
b-cells.1,2,4
DPP-IV is a dipeptidase that selectively binds substrates
with proline at the P1 position. Consequently, many
known DPP-IV inhibitors possess substituted pyrrol-
idines or thiazolidines linked to a-amino acid derivatives
(Fig. 1).5 Screening of the Merck sample collection iden-
tified proline derivative 3 as a moderately active DPP-IV
inhibitor (IC50 = 1.9lM).6 This inhibitor is unique in
OH
O
O
CN
Due to rapid processing of GLP-1 and GIP by DPP-IV,
the half-lives of the active peptides in blood are extre-
mely short. Inhibition of DPP-IV in humans has been
shown to increase circulating GLP-1 and GIP levels,
and leads to decreased blood glucose levels, hemoglobin
A1c levels, and glucagon levels.1,4 DPP-IV inhibitors
offer a number of potential advantages over existing
H
N
N
N
NH2
X
1
2, (X = S or CH2)
Cl
H
N
F
O
N
NH2
F
O
NH2
O
O
Ph
N
Keywords: Dipeptidyl peptidase IV; Selective; Proline amide; Thiazol-
idine; Type 2 diabetes.
F
O
N
H
S
4
3
*
Corresponding author. Tel.: +1 732 5940287; fax: +1 732 5945350;
e-mail: scott_edmondson@merck.com
Figure 1. Some published DPP-IV inhibitors.
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.07.056

5152
S. D. Edmondson et al. / Bioorg. Med. Chem. Lett. 14 (2004) 5151–5155
Table 1. Inhibitory properties of selected phenylacetic acid derived
DPP-IV inhibitors
that it possesses a b-amino amide moiety rather than an
a-amino amide linked to the proline nitrogen. A recent
report from these laboratories describes initial optimiza-
tion of the homophenylalanine moiety in the context of
thiazolidine amides such as 4.6a This report describes
further optimization of screening hit 3 and the discovery
of one of the most potent and selective proline derived
DPP-IV inhibitors to date.7
CO2H
CO2H
H
H
O
N
N
O
N
N
+O
- NH3
O
TFA
Ar
+
- NH3
TFA
13-15
16-20
Compd
ortho/meta/para-
Phenylacetic acid
Ar =
DPP-IV IC50 (nM)
Inhibitors were initially synthesized from acid 7, which
was prepared by coupling acids 5 with L-proline methyl
ester 6 (or R-thiazolidine carboxylic acid methyl
ester, X = S) followed by ester hydrolysis (Scheme 1).
Coupling of acid 7 with various amines followed by
deprotection afforded the final compounds 16–26 and
47–50.8 a-Amino amides 13–15 and 51 were synthesized
the same way using N-Boc-S-cyclohexylglycine instead
of 5.
13
14
15
16
17
18
19
20
ortho
meta
para
ortho
meta
para
para
para
––
––
––
Ph
Ph
Ph
3600
3300
4200
1200
1200
510
3,4-F2Ph
2-FPh
75
54
The synthesis of a series of a-substituted phenoxyacetic
acids began with the coupling of phenol 8 with a-bromo-
esters 9 followed by N-deprotection (Scheme 2). In some
cases, the racemic Boc-protected intermediates were
resolved using chiral chromatography to give optically
pure compounds. Amines 10 were then coupled to
N-Boc-proline and then deprotected to give 11, which
in turn, were coupled with 5. Ester hydrolysis followed
by deprotection then afforded analogs 27–46.
this trend further, different positional isomers of phenyl-
acetic acids were screened (Table 1, 16–18).9 Another
potency enhancement was observed with para-substi-
tuted acid 18 compared to the ortho- and meta-isomers
16 and 17. Since the corresponding S-cyclohexylglycine
series 13–15 shows reduced potency against DPP-IV
compared to 16–18, the a-amino amide series was aban-
doned in favor of the R-homophenylalanines.
Initial SAR surrounding screening hit 3 revealed that a
phenoxyacetic acid derivative was more potent against
DPP-IV than the corresponding cyclopropyl amide or
methyl ester (data not shown). In order to investigate
Previous work from these laboratories indicated that
introduction of fluorine atoms on the pendant phenyl
ring of homophenylalanines improves potency against
DPP-IV.6 This trend was also observed in the proline
amide series (Table 1) with 3,4-difluoro derivative 19
(IC50 = 75nM) and 2-fluoro derivative 20 (IC50
=
O
54nM). The latter compound was nearly 10 times more
potent against DPP-IV than the parent phenyl analog
18, and we hoped to further improve DPP-IV potency
by investigating different right-hand side amides while
maintaining the promising 2-fluorophenyl moiety.
OMe
a, b
O
N
OH
BocHN
Ar
O
BocHN
Ar
O
+
HN
OH
7
X
X
5
6
O
N
NHR
TFA-
+O
NH3
c, d
A series of nonacid bearing proline amides was next
investigated (Table 2). Acids and acid bioisosteres such
as 24 and 26 displayed superior potency against DPP-
IV compared to nonacids. Particularly illustrative are
tetrazole 24 (IC50 = 39nM) and imidazole 25
(IC50 = 1640nM), which are similarly tethered to the
proline amide. The acid preference is also evident by
Ar
X
16-26, 47-50
Scheme 1. Synthesis of various proline amides. Reagents: (a) EDC,
HOBt, DIEA, DCM; (b) LiOHaq, THF, MeOH; (c) EDC, HOBt,
DIEA, RNH2, DCM; (d) TFA, DCM.
OH
Br
CO2Et
CO2Et
O
c, b
a, b
+
BocHN
R
R
9
10
8
NH2
CO2Et
O
CO2H
O
H
TFA-
+H3N
H
O
N
R
O
N
N
d,e,b
R
O
Ar
HN
11
27-46
Scheme 2. Synthesis of phenoxyacetic acids. Reagents: (a) K2CO3, MeOH; (b) TFA, DCM; (c) Boc-S-Pro, EDC, HOBt, DIEA, DCM; (d) 5, EDC,
HOBt, DIEA, DCM; (e) LiOHaq, THF, MeOH.

S. D. Edmondson et al. / Bioorg. Med. Chem. Lett. 14 (2004) 5151–5155
5153
Table 2. Inhibitory properties of selected DPP-IV inhibitors
Table 4. Inhibitory properties of selected DPP-IV inhibitors
CO2H
H
O
TFA-
NH3
O
N
N
H
+O
TFA-
R
O
N
+O
NH3
Ar
N
F
Compd
R
DPP-IV IC50 (nM)
Compd
Ar =
Ph
DPP-IV IC50 (nM)
21
Bn
539
36
37
38
39
40
41
42
43
44
45
46
18
0.48
1.6
8.2
0.83
9.5
33
F3C
2,5-F2Ph
2,4,5-F3Ph
F5Ph
22
23
24
721
4-CF3–2-FPh
4-CF3–2,5-F2Ph
5-F–2-MePh
2-Cl–4-FPh
5-Cl–2-FPh
5-CF3–2-FPh
3-CF3OPh
MeO2C
170
39
1.7
1.6
12
N
N
N
HN
295
N
25
26
1640
118
HN
HO2C
tions proved optimal.10 Indeed, S-isopropyl derivative
35 is 65-fold more potent than the unsubstituted acid
26. We hoped to further improve potency by combining
noncommercially available homophenylalanines 511
with the optimized a-S-isopropyl substituent of 35.
O
As expected based on previously reported results,6
incorporation of a 2,5-difluoro substitution pattern into
the homophenylalanine moiety resulted in another boost
in potency against DPP-IV (Table 4). Analog 37 is
among the most potent proline derivatives yet reported,
exhibiting sub-nanomolar potency against DPP-IV
(IC50 = 0.48nM). In contrast to past observations,6
addition of a third fluoride at the para-position of 37 re-
comparing methyl ester 23 to acid 20, where a 3-fold
potency boost was observed for the acid. On the other
hand, the tether length of the acid appeared to be only
marginally important for DPP-IV potency, with a 2-fold
potency drop observed in going from phenylacetic acid
20 to phenoxyacetic acid 26. Since an acid binding
pocket in the enzyme might be responsible for the
increased affinities of the acids, we sought to further
improve potency against DPP-IV by optimizing the
tether properties of the phenoxyacetic acid series.
sulted in
a potency drop against DPP-IV (38,
IC50 = 1.6nM). Not surprisingly, this aryl group was
particularly sensitive to different substitution patterns.
Small substituent changes led to analogs with relatively
large drops in DPP-IV activity, such as the 5-fluoro-2-
methyl analog 42 (IC50 = 33nM). Nevertheless, many
of the compounds still display excellent potency against
DPP-IV.
Substantial improvements in potency accompanied
increasing steric bulk a to the acid in a series of substi-
tuted phenoxyacetic acids (27–31, Table 3). Diastereo-
merically pure R- and S-isomers of the a-methyl and
a-isopropyl analogs were screened and the S-configura-
Since thiazolidine amides are a common motif in many
reported DPP-IV inhibitors, we replaced the pyrrolidine
moiety with a thiazolidine in selected compounds (Table
5). In each case, a potency boost was observed with the
Table 3. Effect of a-substitution on right-hand side phenoxyacetic
acids
CO2H
O
H
TFA-
O
N
N
R
+O
NH3
Table 5. Thiazolidines are more potent than their pyrrolidine coun-
terparts
F
R
Compd
R =
Stereochemistry
a to acid
DPP-IV
IC50 (nM)
H
TFA-
NH3
Ar
N
O
N
+O
27
28
29
30
31
32
33
34
35
Me
Et
R/S
R/S
R/S
R/S
R/S
R
18
7.1
10
S
n-Pr
i-Pr
n-Bu
Me
Me
i-Pr
i-Pr
3.7
5.1
82
Compd Ar =
R =
–CH2CO2H
DPP-IV IC50 (nM)
47
48
49
50
Ph
161
87
S
12
13
2-FPh –CH2CO2Me
2-FPh –CH2CO2H
2,5-F2Ph –OCH(i-Pr)CO2H (S)
R
37
S
1.8
0.30

5154
S. D. Edmondson et al. / Bioorg. Med. Chem. Lett. 14 (2004) 5151–5155
Table 6. Selectivity profiles of selected DPP-IV inhibitors
Compd DPP-IV IC50
(lM)
QPP IC50
(lM)
DPP8 IC50
(lM)
DPP9 IC50
(lM)
FAP IC50
(lM)
PEP IC50
(lM)
APP IC50
(lM)
Prolidase
hERG IC50
(lM)
19
22
24
37
39
48
49
51
0.075
0.72
>100
55
>100
5.3
>100
>100
>100
86
>100
>100
>100
21
––
––
––
––
79
0.87
––
>100
>100
>100
>100
>100
––
>100
>100
>100
>100
>100
––
0.039
0.00048
0.0082
0.087
0.037
7.2
>100
>100
>100
59
>100
>100
>100
>100
>50
––
39
76
>100
>100
71
48
>100
100
––
>100
9.7
––
>100
>100
>100
>100
>100
4.1
0.45
>100
––
––
>90
thiazolidine derivatives compared to the corresponding
proline derivative, leading to 50 (IC50 = 0.30nM), one
of the most potent DPP-IV inhibitors yet reported.
of steady state velocities observed with 37 indicate that
this inhibitor is not a substrate for DPP-IV. Taken to-
gether, these data indicate that 37 is a reversible inhibi-
tor of DPP-IV.
In order to ensure that these inhibitors are neither sub-
strates nor irreversible inhibitors of DPP-IV, the inhibi-
tion kinetics and off-rate of potent inhibitor 37 were
measured as a representative example. The resulting
progress curves (Fig. 2) are indicative of a typical
slow binding reversible inhibitor,12 and the formation
A representative series of the above inhibitors was
assessed for their selectivity profiles against other dip-
eptidyl peptidase homologs and proline-specific enzymes
(Table 6) including quiescent prolyl peptidase (QPP/
DPP-II), dipeptidyl peptidases 8 and 9 (DPP8 and
DPP9), seprase (FAP), prolidase, amino peptidase P
(APP), and prolyl endopeptidase (PEP).9,13 The same
series of inhibitors was also tested against the hERG
potassium channel as a general test of off-target activ-
ity.14 For comparison purposes, cyclohexylglycine deriv-
ative 51 (Fig. 3) was also screened against this same
series of enzymes. All of the homophenylalanine derived
acids (19, 37, 39, and 49) and tetrazole 24 displayed
excellent selectivity against all of the counterscreens.
On the other hand, the nonacids all possess increased
activity against hERG, with the trifluoromethylbenzyl
analog 22 being the most potent (hERG IC50 = 870nM).
Cyclohexylglycine analog 51 was only moderately
potent against DPP-IV and inactive against hERG,
but it was relatively potent against DPP8 and DPP9.
The above data indicates that the acid functionality
in this series is important for selectivity over hERG
while the b-amino amide functionality is important for
selectivity over DPP8 and DPP9. Selectivity over these
homologs is important because inhibition of these
enzymes has been associated with multi-organ toxicity
in pre-clinical species.15
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
(A)
0
200
400
time, s
600
800
3.0
(B)
2.5
2.0
1.5
1.0
0.5
0.0
The rat pharmacokinetic profiles of the potent
amino acids examined were uniformly poor (Table 7),
although thiazolidine analog 49 showed improved oral
0
500
1000
1500
2000
2500
Table 7. Pharmacokinetic properties of selected DPP-IV inhibitors in
the rat (1/2mpk iv/po)
time, s
Figure 2. (A) Human recombinant DPP-IV was studied by liberation
of AMC from substrate Gly-Pro-AMC in the absence (dots) and
presence (triangles) of 37. The data were fit to the integrated rate
equation for slow binding inhibition by nonlinear regression analysis.
Binding constant kon = 2.0 0.5 Æ 106 MÀ1 sÀ1 was calculated from
from individual data sets. (B) The dissociation rate was determined by
pre-incubation of DPP-IV with 37 under the same conditions and
subsequent 100-fold dilution. Eight individual data sets were fit as
described above to determine koff = 1.2 0.3 Æ 10À3 sÀ1, which corre-
sponds to t1/2 = 10min.
Compd
Clp
(mLminÀ1 kgÀ1
t1/2 (h) Oral AUCnorm
(lMhkgmgÀ1
)
F (%)
)
19
22
24
37
39
48
49
47
78
0.36
0.78
0.20
0.63
0.36
0.10
0.19
0.020
0.183
––
<1.0
38
54
<1.0
1.2
150
47
0.003
<0.001
0.031
0.151
<1.0
11
140
25
10

S. D. Edmondson et al. / Bioorg. Med. Chem. Lett. 14 (2004) 5151–5155
5155
O
CO2H
CO2H
O
F
H
H
TFA-
NH3
O
N
O
N
N
N
+O
O
+
F
TFA-
NH3
51, DPP-IV IC50 = 7200 nM
37, DPP-IV IC50 = 0.48 nM
Figure 3. Cyclohexylglycine analog 51 and potent and selective DPP-IV inhibitor 37.
4. Ahren, B.; Landin-Olsson, M.; Jansson, P.-A.; Svensson,
M.; Holmes, D.; Schweizer, A. J. Clin. Endocrinol. Metab.
2004, 89, 2078.
5. (a) Villhauer, E. B.; Brinkman, J. A.; Naderi, G. B.;
Burkey, B. F.; Dunning, B. E.; Prasad, K.; Mangold, B.
L.; Russell, M. E.; Hughes, T. E. J. Med. Chem. 2003, 46,
2774; (b) Ashworth, D. M.; Atrash, B.; Baker, G. R.;
Baxter, A. J.; Jenkins, P. D.; Jones, D. M.; Szelke, M.
Bioorg. Med. Chem. Lett. 1996, 6, 1163, (2 X = CH2) and
2745 (2, X = S).
6. (a) Xu, J.; Ok, H. O.; Gonzalez, E. J.; Colwell, L. F., Jr.;
Habulihaz, B.; He, H.; Leiting, B.; Lyons, K. A.; Marsilio,
F.; Patel, R. A.; Wu, J. K.; Thornberry, N. A.; Weber, A.
E.; Parmee, E. R. Bioorg. Med. Chem. Lett. 2004, 14, 4759;
(b) Brockunier, L. L.; He, J.; Colwell, L. F., Jr.;
Habulihaz, B.; He, H.; Leiting, B.; Lyons, K. A.; Marsilio,
F.; Patel, R. A.; Teffera, Y.; Wu, J. K.; Thornberry, N. A.;
Weber, A. E.; Parmee, E. R. Bioorg. Med. Chem. Lett.
2004, 14, 4763.
bioavailability and exposure. In the case of methyl ester
48, a rapid and complete conversion of the ester to acid 49
was observed. In order to determine the reason for the
poor oral exposures, three representative compounds
(19, 20, and 37) were administered orally to portal vein
cannulated rats. Compared to a well absorbed positive
control (theophylline), all three compounds were poorly
absorbed ([compound] < 10nM). Thus, the poor bioavail-
ability of these proline derived acids can be attributed
to poor oral absorption of this series of inhibitors. It ap-
pears that the acid functionality contributes to the poor
absorption since 22 possesses improved oral bioavaila-
bility and exposure compared to the other analogs.
In conclusion, a series of potent and highly selective pro-
line derived DPP-IV inhibitors has been described, cul-
minating in the discovery of analog 37, a 0.48nM
DPP-IV inhibitor with >50,000-fold selectivity over all
other enzymes tested (Fig. 3). In contrast, cyclohexylgly-
cine derived a-amino amide analog 51 is a poor DPP-IV
inhibitor with no selectivity over DPP8 and DPP9. Con-
sequently, it appears that the excellent selectivity pro-
file of the above described inhibitors is derived from
the b-amino amide functionality combined with the
carboxylic acid moiety. Future efforts in this structure
class will focus on maintaining the excellent potency
and selectivity while improving the pharmacokinetic
profile of these homophenylalanine derived DPP-IV
inhibitors.
7. This work was reported in part: Edmondson, S. D.
Abstract 99, 227th National Meeting of the ACS, Ana-
heim, CA, March 2004.
1
8. All final compounds were characterized by H NMR and
LC–MS.
9. IC50 determinations for DPP-IV and other proline pep-
tidases were carried out as described in: Leiting, B.; Pryor,
K. D.; Wu, J. K.; Marsilio, F.; Patel, R. A.; Craik, C. S.;
Ellman, J. A.; Cummings, R. T.; Thornberry, N. A.
Biochem. J. 2003, 371, 525.
10. In the case of the a-methyl series the stereochemistry was
confirmed by synthesis of optically pure derivatives via
Mitsunobu reaction of 8 with R- or S-lactic acid esters. In
the case of the isopropyl substituent stereochemistry was
confirmed by X-ray crystal structure of 37.
11. (a) As described in Ref. a optically active b-amino acids 5
were prepared using Schollkopfꢀs synthesis of a-amino
acids followed by Arndt–Eistert homologation. See:
Schollkopf, U. Tetrahedron 1983, 39, 2085; (b) Podlech,
J.; Seebach, D. Liebigs Ann. 1995, 1217.
References and notes
1. (a) For lead references, See: Drucker, D. J. Exp. Opin.
Investig. Drugs 2003, 12, 87–100; (b) Wiedeman, P. E.;
Trevillyan, J. M. Curr. Opin. Investig. Drugs 2003, 4,
412–420; (c) Augustyns, K.; Van der Veken, P.; Senten,
K.; Haemers, A. Expert Opin. Ther. Patents 2003, 13,
499–510; (d) Weber, A. E. J. Med. Chem. 2004, 47, 4135.
2. (a) Reimer, M. K.; Holst, J. J.; Ahren, B. Eur. J.
Endocrinol. 2002, 146, 717; (b) Pospisilik, J. A.; Martin,
J.; Doty, T.; Ehses, J. A.; Pamir, N.; Lynn, F. C.; Piteau,
S.; Demuth, H.-U.; McIntosh, C. H. S.; Pederson, R. A.
Diabetes 2003, 52, 741.
12. Williams, J. W.; Morrison, J. F. Methods Enzymol. 1979,
6, 437.
13. (a) For lead references, see: Sedo, A.; Malik, R. Biochim.
Biophys. Acta 2001, 1550, 107; (b) Rosenblum, J. S.;
Kozarich, J. W. Curr. Opin. Chem. Biol. 2003, 7, 496.
14. Friesen, R. W.; Ducharme, Y.; Ball, R. G.; Blouin, M.;
Boulet, L.; Cote, B.; Frenette, R.; Girard, M.; Guay, D.;
Huang, Z.; Jones, T. R.; Laliberte, F.; Lynch, J. J.;
Mancini, J.; Martins, E.; Masson, P.; Muise, E.; Pon, D.
J.; Siegel, P. K. S.; Styhler, A.; Tsou, N. N.; Turner, M. J.;
Young, R. N.; Girard, Y. J. Med. Chem. 2003, 46, 2017.
15. Lankas, G.; Leiting, B.; Sinha Roy, R.; Eiermann, G.;
Biftu, T.; Kim, D.; Ok, H.; Weber, A. E.; Thornberry, N.
A. Diabetes 2004, 53(S2), A2.
3. During, M. J.; Cao, L.; Zuzga, D. S.; Francis, J. S.;
Fitzsimons, H. L.; Jiao, X.; Bland, R. J.; Klugmann, M.;
Banks, W. A.; Drucker, D. J.; Haile, C. N. Nature Med.
2003, 9, 1173.