4-Arylcyclohexylalanine analogs as potent, selective, and orally active inhibitors of dipeptidyl peptidase IV

Article abstract of DOI:10.1016/j.bmcl.2007.08.049

A novel series of 4-arylcyclohexylalanine DPP-4 inhibitors was synthesized and tested for inhibitory activity as well as selectivity over the related proline-specific enzymes DPP-8 and DPP-9. Optimization of this series led to 28 (DPP-4 IC₅₀ =

Full text of DOI:10.1016/j.bmcl.2007.08.049

Bioorganic & Medicinal Chemistry Letters 17 (2007) 5806–5811
4-Arylcyclohexylalanine analogs as potent, selective,
and orally active inhibitors of dipeptidyl peptidase IV
David E. Kaelin,a,* Abigail L. Smenton,a George J. Eiermann,b Huaibing He,a
Barbara Leiting,c Kathryn A. Lyons,a Reshma A. Patel,c Sangita B. Patel,a
Alexsandr Petrov,b Giovanna Scapin,a Joseph K. Wu,c Nancy A. Thornberry,c
Ann E. Webera and Joseph L. Duffya
aDepartment of Medicinal Chemistry, Merck Research Laboratories, Rahway, NJ 07065, USA
bDepartment of Pharmacology, Merck Research Laboratories, Rahway, NJ 07065, USA
cDepartment of Metabolic Disorders, Merck Research Laboratories, Rahway, NJ 07065, USA
Received 26 June 2007; revised 22 August 2007; accepted 23 August 2007
Available online 26 August 2007
Abstract—A novel series of 4-arylcyclohexylalanine DPP-4 inhibitors was synthesized and tested for inhibitory activity as well as
selectivity over the related proline-specific enzymes DPP-8 and DPP-9. Optimization of this series led to 28 (DPP-4 IC50 = 4.8 nM),
which showed an excellent pharmacokinetic profile across several preclinical species. Evaluation of 28 in an oral glucose tolerance
test demonstrated that this compound effectively reduced glucose excursion in lean mice.
Ó 2007 Elsevier Ltd. All rights reserved.
Glucagon-like peptide 1 (GLP-1) is an incretin hormone
that has been shown to be responsible for glucose-stim-
ulated insulin secretion following nutrient ingestion.1
GLP-1 is rapidly inactivated, via proteolytic cleavage,
by the serine protease dipeptidyl peptidase IV (DPP-4).
Inhibition of DPP-4 has been shown to lead to an
increase in circulating levels of endogenous GLP-1,
and this in turn causes a reduction in blood glucose
levels. As a result, DPP-4 inhibition has emerged as an
effective method for the treatment of type 2 diabetes.2
apparent link between inhibition of DPP-8/DPP-9 and
toxicity in preclinical species.5
Two potential issues were identified within the phenylal-
anine class of compounds to which 1 and 2 belong. The
compounds generally exhibited ion channel binding
(hERG) and a high serum potency shift (>10-fold
increase in apparent IC50). We believed that both of
these properties were arising from the highly lipophilic
biaryl moiety, so an effort was undertaken to replace
Recently, JANUVIATM (sitagliptin phosphate), a potent,
selective, and orally active DPP-4 inhibitor, was ap-
proved by the FDA for the treatment of diabetes.3
Continuing work in these laboratories has led to the dis-
covery of a number of potent and structurally diverse
DPP-4 inhibitors (Fig. 1).4 One class of phenylalanine
derivatives, typified by 1 and 2, were shown to be potent
against DPP-4 and selective over the related proline-spe-
cific peptidases DPP-8 and DPP-9 (Table 1).4a,b This
selectivity was regarded as an absolute requirement for
all structural classes in our program because of the
Me
N
O
R
O
Me
O
N
N
NH2
NH2
N
N
N
F
F
3
F
1 R = CON(Me)2
2 R = Me
Me
O
R1
O
N
H
O
N
NH2
Me
N
H
NH2
H
Ar
F
F
Keyword: DPP-4 inhibitors.
*
4
5
Corresponding author. Tel.: +1 732 594 2266; fax: +1 732 594
9545; e-mail: david_kaelin@merck.com
Figure 1. Structures of DPP-4 inhibitors.
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.08.049

D. E. Kaelin et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5806–5811
5807
Table 1. Potency and selectivity over DPP-8 and DPP-9, selectivity over hERG, and rat PK parameters
Compound
IC50 (lM)
Rat PK (1/2 mpk iv/po)
DPP-4
DPP-8
DPP-9
hERG
AUCn (lM h/mpk)
Clp (mL/min/kg)
F (%)
1
2
3
4
0.012
0.064
0.004
0.016
>100
88
69
86
4.6
1.1
6.2
4.8
2.4
0.68
4.8
8.6
7.0
42
67
85
43
56
>100
25
>100
>100
86
>100
the 4-fluorophenyl group present in 1 and 2 with more
polar substituents. This led to the discovery of heterocy-
clic derivatives like 3, which had an improved hERG
profile.4c Shortly thereafter, it was discovered that the
phenyl ring in the phenylalanine series could be replaced
with a cyclohexyl ring to give compounds like 4 that had
good intrinsic potency, high selectivity against DPP-8
and DPP-9, and low ion channel binding.6 Unfortu-
nately, compound 4 showed a lower than anticipated ef-
fect in a murine OGTT (oral glucose tolerance test), and
this was ultimately attributed to low oral exposure (as
evidenced by AUCn) in rodents. Based on these discov-
eries, efforts were initiated to combine the structural fea-
tures present in 3 and 4 to give hybrid structures like 5 in
the hopes of generating potent and selective inhibitors
with improved pharmacokinetic profiles.
mina gave a mixture of diasteromeric alcohols that was
subsequently oxidized to give the cyclohexanone 9
(Scheme 1).
One of two approaches was then used to install the 4-
substituent. In the first approach (Scheme 2), the cyclo-
hexanone 9 was first converted to its corresponding enol
triflate 10 by treatment with 2 equivalents of LiHMDS
followed by PhNTf2. The enol triflate was converted
to vinyl boronate 11 by employing conditions similar
to those used in the preparation of 7. Palladium-cata-
lyzed coupling of 11 with an appropriate aryl or hetero-
aryl halide afforded cyclohexenes of the general
structure 12.
The second approach that was used to prepare the com-
mon intermediate 12 commenced with the addition of an
appropriately substituted aryllithium or arylmagnesium
bromide reagent to cyclohexanone 9 (Scheme 3). The
resultant diastereomeric mixture of benzylic alcohols
13 was then dehydrated by the action of Burgess re-
agent. This method was employed in instances where
the requisite organometallic reagent was commercially
available (e.g., 4-fluorophenylmagnesium bromide).
The preparation of a representative hybrid scaffold is de-
scribed in Schemes 1–4. The synthesis commenced with
a
palladium-catalyzed coupling reaction between
bis(pinacolato)diboron and the aryl bromide 6 to give
7 as was previously described.4c The boronate ester in
7 was then cleaved oxidatively to furnish phenol 8.
Hydrogenation of 8 in the presence of rhodium on alu-
O
B
Br
NHBoc
F
NHBoc
F
O
F
a
N
F
b
N
F
O
O
Me2N
O
Me2N
O
7
6
HO
O
NHBoc
N
NHBoc
N
F
F
F
c, d
O
O
8
O
O
Me2N
Me2N
9
Scheme 1. (a) Bis(pinacolato)diboron, Pd(dppf)Cl2, KOAc, DMSO, 90 °C, 12 h, 90–98%; (b) H2O2 (30% aq), aq NaOH (3N), THF, 86–97%; (c) 5%
Rh/Al2O3, H2 (50 psi), MeOH, 12 h; (d) Dess-Martin periodinane, CH2Cl2, 56–81% (2 steps).
XO
Ar
NHBoc
N
NHBoc
N
F
F
F
F
a
c
9
O
O
O
O
Me2N
Me2N
12
10: X = Tf
b
O
O
B
11: X =
Scheme 2. (a) LiHMDS, THF, À78 °C, 1.5 h, then PhNTf2, À78 to 5 °C, 48–93%; (b) bis(pinacolato)diboron, Pd(dppf)Cl2, dppf, KOAc, p-dioxane,
80 °C, 46–57%; (c) Ar–Br, Pd(dppf)Cl2, K2CO3, DMF, 80 °C, 56–87%.

5808
D. E. Kaelin et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5806–5811
OH
against the related proline-specific peptidases DPP-8
Ar
NHBoc
F
F
and DPP-9. For this reason, an alternate scaffold was
investigated in an attempt to mitigate the unwanted
off-target activity. We opted not to separate the cis
and trans isomers of 16–18 at this stage, but rather to
use the determined activity of the mixtures as a guide
to which analogs should be focused on first in the alter-
nate scaffold design.
a
b
N
9
12
O
Me2N
O
13
Scheme 3. (a) Ar–Li or Ar–MgBr(Cl), THF, À78 °C to rt; (b) Burgess
reagent, THF, 49–79% (2 steps).
It had been shown previously that replacement of the
b-methyl substituent present in the phenylcyclohexylala-
nine inhibitors with a polar group often resulted in an
increase in selectivity over DPP-8 and DPP-9.4b,c Hence,
this strategy was employed for the synthesis of addi-
tional inhibitors.
TFA-
TFA-
H
H
+
Ar
+
Ar
NH3
NH3
F
F
F
F
a, b, c
H
H
12
N
N
O
O
15
Me2N
O
Me2N
O
14
Scheme 4. (a) 10% Pd/C, H2 (1 atm), MeOH; (b) HPLC (Chiracel OD,
i-PrOH/hexane), 16–32% (2 steps); (c) TFA, CH2Cl2, 88–100%.
Several analogs were prepared that incorporated the
triazolopyridine present in 18 onto the b-dimethylamido
scaffold and the data for these compounds, obtained as
single diastereomers, are summarized in Table 3. Both
the 3,3-difluoropyrrolidine and (S)-fluoropyrrolidine
amides were investigated. The more active monofluoro
diastereomer 20 exhibited good potency against DPP-4
and also showed significant selectivity over DPP-8 and
DPP-9. When dosed in rats, compound 20 showed only
moderate oral bioavailability and a short half-life.8
Alternatively, difluoropyrrolidine amide 22 possessed
similar potency and selectivity to 20 but had a substan-
tially improved pharmacokinetic profile in rats. How-
ever, due to the relatively high clearance (Clp = 40 mL/
min/kg) of 22, the AUCn (0.93 lM h/mpk) for this com-
pound was lower than we desired.
The double bond in 12 was reduced by hydrogenation in
the presence of palladium on carbon to give a mixture of
cis- and trans-cyclohexane diastereomers that, in most
cases, was separated by HPLC (Scheme 4). Cleavage
of the Boc-protecting group afforded the target com-
pounds 14 and 15 as single diastereomers. An identical
procedure was employed, starting with the known aryl
bromide,4a to prepare analogs that contained a b-methyl
substituent in place of the b-dimethylamide group.
Table 2 lists the DPP-4 inhibitory potency and selectiv-
ity over the proline-specific enzymes DPP-8 and DPP-9
for several of the b-methyl-substituted inhibitors.7a,b It
should be noted that compounds 16–18 were analyzed
for inhibitory activity as mixtures (ca. 1:1) of cis and
trans cyclohexane isomers. Compounds 16 and 17
showed high intrinsic potency against DPP-4, but the
apparent potency was decreased in the presence of
50% human serum. Compound 18 on the other hand
was highly potent both in the presence and absence of
human serum, despite a substantial serum shift. All
three inhibitors exhibited low micromolar activity
Next, we prepared several additional analogs (23–30)
that incorporated the dimethylamido side chain and
the difluoropyrrolidine amide (Table 4). Compound
28, which contained the regioisomeric triazolopyridine
present in 3,4c showed excellent potency against DPP-4
and high selectivity over DPP-8 and DPP-9. However,
this compound did exhibit a larger than expected de-
crease in potency in the presence of 50% human serum.
Finally, compound 28 was found to be inactive against
Table 2. Potency in the presence of 50% human serum (HS) and selectivity over DPP-8 and DPP-9
Me
O
N
NH2
R1
R2
F
Compounda
R1
R2
IC50 (lM)
DPP-4 (0% HS)
0.012
DPP-4 (50% HS)
DPP-8
3.6
DPP-9
3.1
16
17
18
H
F
F
0.39
0.15
0.028
F
0.0046
0.0015
6.4
4.3
11
N
N
1.7
N
N
a All compounds are a mixture (ca. 1:1) of cis- and trans-cyclohexane diastereomers.

D. E. Kaelin et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5806–5811
Table 3. Potency and selectivity over DPP-8/DPP-9 and selected rat PK data
5809
Me
N
O
Me
O
N
NH2
R
F
N
N
N
Compound
R
IC50 (lM)
Rat PK
DPP-4 (0% HS)
DPP-4 (50% HS)
DPP-8
DPP-9
t1/2 (h)
0.54
F (%)
19 (Diast A)
20 (Diast B)
H
H
0.31
0.0023
0.82
0.026
>100
53
3.7
28
31
98
21 (Diast A)
22 (Diast B)
F
F
0.26
0.0025
0.67
0.023
95
33
2.6
25
2.9
Table 4. Potency and selectivity over DPP-8/DPP-9
Me
N
O
Me
O
N
NH2
R
F
F
Compound
R
IC50 (lM)
DPP-4 (0% HS)
DPP-4 (50% HS)
DPP-8
DPP-9
MeO
N
23 (Diast A)
24 (Diast B)
0.10
0.26
2.7
1.9
>100
>100
40
>100
N
OMe
25 (Diast A)
26 (Diast B)
0.15
0.011
1.5
0.29
>100
41
15
21
F
N
27 (Diast A)
28 (Diast B)
0.032
0.0048
0.13
0.13
>100
30
38
38
N
N
N
N
29 (Diast A)
30 (Diast B)
0.22
0.038
0.45
0.12
>100
22
50
>100
N
the hERG ion channel (IC50 > 90 lM) so for this reason
it was selected for further in vivo characterization de-
spite the high serum shift.
dextrose challenge (Fig. 2). Gratifyingly, 28, when
administered orally as a single dose 60 min prior to a
dextrose challenge, reduced blood glucose levels in a
dose-dependent fashion down to a minimum effective
dose (MED) of 0.03 mpk which resulted in a 23% reduc-
tion in blood glucose excursion relative to vehicle con-
trol. The MED for both 1 and 4 was 0.1 mpk which
resulted in a 21% and 26% reduction in blood glucose,
respectively.
The pharmacokinetic profile of 28 was determined for
several species (Table 5). In general, 28 exhibited excel-
lent oral bioavailability and low clearance. The half-life
across species was generally consistent. Importantly, the
oral AUCn was high in all species, so 28 was tested fur-
ther in our primary pharmacodynamic assay.
X-ray crystal structures of the more active diastereomers
20 and 28 bound to the active site of DPP-4 were ob-
tained (Fig. 3).9a,b These structures established unam-
biguously the 1,4-trans relationship of the cyclohexyl
A lean mouse oral glucose tolerance test (OGTT) was
performed using 28 in order to determine the effect of
the inhibitor on blood glucose excursion following a

5810
D. E. Kaelin et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5806–5811
Table 5. PK data for 28
Acknowledgments
Species
Selected pharmacokinetic parameters (1/2 mpk iv/po)
We are grateful to D. Hora, J. Fenyk-Melody, I. Capod-
anno, P. Cunningham, M. Donnelly, J. Hausamann, C.
Nunes, X. Shen, J. Strauss, and K. Vakerich for in vivo
pharmacokinetic studies.
AUCn
(lM h)
Clp
(mL/min/kg)
Cmax
(lM)
t1/2
F
(h)
(%)
Rat
Doga
Rhesus
4.4
46
5.3
3.3
16
4.1
3.9
4.1
63
0.94
5.6
100
97
6.5
2.8
a po dose—0.5 mpk.
References and notes
1. For recent GLP-1 references, see: (a) Holst, J. J. Curr. Opin.
Endocrin. Diabetes 2005, 12, 56; (b) Knudsen, L. B. J. Med.
Chem. 2004, 47, 4128; (c) Vahl, T. P.; D’Alessio, D. A.
Expert Opin. Invest. Drugs 2004, 13, 177, and references
therein.
2. For reviews on DPP-4 inhibitors, see: (a) Nielson, L. L.
Drug Discovery Today 2005, 10, 703; (b) Augustyns, K.;
Van der Veken, P.; Haemers, A. Expert Opin. Ther. Patents
2005, 15, 1387; (c) Mentlein, R. Expert Opin. Invest. Drugs
2005, 14, 57; (d) Weber, A. E. J. Med. Chem. 2004, 47,
4135; (e) Augustyns, K.; Van der Veken, P.; Senten, K.;
Haemers, A. Expert Opin. Ther. Patents 2003, 13, 499, and
references therein.
400
300
200
100
vehicle
X:
0.01 mg/kg
0.03 mg/kg
0.1 mg/kg
0.3 mg/kg
control
0
pre
0
20 40 60
120
time (min)
Figure 2. Effects of 28 on glucose levels after an oral glucose tolerance
test in lean mice.
3. Kim, D.; Wang, L.; Beconi, M.; Eiermann, G. J.; Fisher,
M. H.; He, H.; Hickey, G. J.; Kowalchick, J. E.; Leiting, B.;
Lyons, K.; Marsillio, F.; McCann, M. E.; Patel, R. A.;
Petrov, A.; Scapin, G.; Patel, S. B.; Roy, R. S.; Wu, J. K.;
Wyvratt, M. J.; Zhang, B. B.; Zhu, L.; Thornberry, N. A.;
Weber, A. E. J. Med. Chem. 2005, 48, 141.
4. (a) Xu, J.; Wei, L.; Mathvink, R.; He, J.; Park, Y.-J.; He,
H.; Leiting, B.; Lyons, K. A.; Marsilio, F.; Patel, R. A.;
Wu, J. K.; Thornberry, N. A.; Weber, A. E. Bioorg. Med.
Chem. Lett. 2005, 15, 2533; (b) Edmondson, S. D.;
Mastracchio, A.; Duffy, J. L.; Eiermann, G. J.; He, H.;
Ita, I.; Leiting, B.; Leone, J. F.; Lyons, K. A.; Makarewicz,
A. M.; Patel, R. A.; Petrov, A.; Wu, J. K.; Thornberry, N.
A.; Weber, A. E. Bioorg. Med. Chem. Lett. 2005, 15, 3048;
(c) Edmondson, S. D.; Mastracchio, A.; Mathvink, R. J.;
He, J.; Harper, B.; Park, Y.-J.; Beconi, M.; Di Salvo, J.;
Eiermann, G. J.; He, H.; Leiting, B.; Leone, J. F.; Levorse,
D. A.; Lyons, K.; Patel, R. A.; Patel, S. B.; Petrov, A.;
Scapin, G.; Shang, J.; Sinha Roy, R.; Smith, A.; Wu, J. K.;
Xu, S.; Zhu, B.; Thornberry, N. A.; Weber, A. E. J. Med.
Chem. 2006, 49, 3614; (d) Parmee, E. R.; He, J.; Mastrac-
chio, A.; Edmondson, S. D.; Colwell, L.; Eiermann, G.;
Feeney, W. P.; Habulihaz, B.; He, H.; Kilburn, R.; Leiting,
B.; Lyons, K.; Marsilio, F.; Patel, R. A.; Petrov, A.; Di
Salvo, J.; Wu, J. K.; Thornberry, N. A.; Weber, A. E.
Bioorg. Med. Chem. Lett. 2004, 14, 43; (e) Caldwell, C. G.;
Chen, P.; He, J.; Parmee, E. R.; Leiting, B.; Marsilio, F.;
Patel, R. A.; Wu, J. K.; Eiermann, G. J.; Petrov, A.; He, H.;
Lyons, K. A.; Thornberry, N. A.; Weber, A. E. Bioorg.
Med. Chem. Lett. 2004, 14, 1265.
5. Lankas, G. R.; Leiting, B.; Sinha Roy, R.; Eiermann, G. J.;
Biftu, T.; Cahn, C.-C.; Edmondson, S. D.; Feeney, W. P.;
He, H.; Ippolito, D. E.; Kim, D.; Lyons, K. A.; Ok, H. O.;
Patel, R. A.; Petrov, A. N.; Pryor, K. A.; Qian, X.; Reigle,
L.; Woods, A.; Wu, J. K.; Zaller, D.; Zhang, X.; Zhu, L.;
Weber, A. E.; Thornberry, N. A. Diabetes 2005, 54, 2988.
6. Duffy, J. L.; Kirk, B. A.; Wang, L.; Eiermann, G. J.; He,
H.; Leiting, B.; Lyons, K. A.; Patel, R. A.; Patel, S. B.;
Petrov, A.; Scapin, G.; Wu, J. K.; Thornberry, N. A.;
Weber, A. E. Bioorg. Med. Chem. Lett. 2007, 17, 2879.
7. For assay conditions for DPP-4 inhibition, see: (a) 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. J. Biochem. 2003, 371, 525; (b) For assay conditions
for DPP-8 and DPP-9, see Ref. 5.
Figure 3. Inhibitors 20 (cyan) and 28 (yellow) bound to DPP-4.
Interactions of the inhibitors with DPP-4 are shown as dotted red lines.
substituents in these inhibitors. The more potent diaste-
reomer of each pair in Tables 3 and 4 is presumed to be
trans by analogy. The interactions between 20/28 and
DPP-4 are similar to those seen with related com-
pounds.4c It is noteworthy that the heterocycle in 28 is
capable of participating in a hydrogen bonding interac-
tion with Arg358 whereas the heterocycle in 20 cannot.
This interaction appears to have no impact on intrinsic
potency.
In conclusion, the hybridization of DPP-4 inhibitors 3
and 4 gave a new class of compounds that effectively
incorporated the desirable properties of each. The 4-aryl
and heteroarylcyclohexylalanines described are highly
potent and selective over DPP-8 and DPP-9. One mem-
ber of this new class, 28, has improved selectivity over
hERG compared to 1 and an improved pharmacoki-
netic profile relative to 4. Compound 28 showed good
efficacy in a murine OGTT experiment.

D. E. Kaelin et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5806–5811
5811
8. Pharmacokinetic parameters were analyzed with WATSON
software (version 6.4.0.04 Innaphase Corp) by noncom-
partmental methods using formulas described in Gibaldi,
M.; Perrier, D. In Pharmacokinetics; Swarbrick, J., Ed., 2nd
ed.; Marcel Dekker: New York, 1982; p 409.
9. X-ray crystal structures were obtained following the pub-
lished protocol: (a) Kim, D.; Wang, L.; Beconi, M.;
Eiermann, G. J.; Fisher, M. H.; He, H.; Hickey, G. J.;
Kowalchick, J. E.; Leiting, B.; Lyons, K.; Marsilio, F.;
McCann, M. E.; Patel, R. A.; Petrov, A.; Scapin, G.; Patel,
S. B.; Sinha Roy, R.; Wu, J. K.; Wyvratt, M. J.; Zhang, B.
B.; Zhu, L.; Thornberry, N. A.; Weber, A. E. J. Med.
Chem. 2005, 48, 141; (b) The structures of DPP-4 bound to
compound 20 and 28 have been deposited with the RCSB
Protein Data Bank, Accession Nos. 2QT9 and 2QTB,
respectively.