Synthesis and pharmacological evaluation of glycine amide derivatives as novel vascular adhesion protein-1 inhibitors without CYP3A4 and CYP2C19 inhibition

Article abstract of DOI:10.1016/j.bmc.2017.05.059

Vascular adhesion protein-1 (VAP-1) is a promising therapeutic target for the treatment of diabetic nephropathy. Here, we conducted optimization studies of our lead compound 1, which we previously reported as a novel VAP-1 inhibitor, to enhance the inhibition of human VAP-1 and to reduce CYP3A4 and CYP2C19 inhibition. As a result, we identified 3-chloro-4-{4-[5-(3-{[glycyl(methyl)amino]methyl}phenyl)pyrimidin-2-yl]piperazin-1-yl}benzoic acid (17h) as a novel orally active VAP-1 inhibitor, with 14-fold increased human VAP-1 inhibitory activity compared to 1, without CYP3A4 and CYP2C19 inhibition. Oral administration of 17h significantly inhibited the progression of proteinuria in streptozotocin (STZ) induced diabetic rats at 0.3 and 1?mg/kg, suggesting that this compound has potential to be a therapeutic agent for the treatment of diabetic nephropathy.

Full text of DOI:10.1016/j.bmc.2017.05.059

Accepted Manuscript
Synthesis and pharmacological evaluation of glycine amide derivatives as novel
Vascular Adhesion Protein-1 inhibitors without CYP3A4 and CYP2C19 inhib-
ition
Susumu Yamaki, Yuji Koga, Akira Nagashima, Mitsuhiro Kondo, Yoshiaki
Shimada, Keitaro Kadono, Ayako Moritomo, Kosei Yoshihara
PII:
S0968-0896(17)30433-9
DOI:
http://dx.doi.org/10.1016/j.bmc.2017.05.059
BMC 13778
Reference:
To appear in:
Bioorganic & Medicinal Chemistry
Received Date:
Revised Date:
Accepted Date:
2 March 2017
8 May 2017
30 May 2017
Please cite this article as: Yamaki, S., Koga, Y., Nagashima, A., Kondo, M., Shimada, Y., Kadono, K., Moritomo,
A., Yoshihara, K., Synthesis and pharmacological evaluation of glycine amide derivatives as novel Vascular
Adhesion Protein-1 inhibitors without CYP3A4 and CYP2C19 inhibition, Bioorganic & Medicinal Chemistry
(
2017), doi: http://dx.doi.org/10.1016/j.bmc.2017.05.059
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Synthesis and pharmacological evaluation of glycine amide derivatives as novel
Vascular Adhesion Protein-1 inhibitors without CYP3A4 and CYP2C19 inhibition
*
Susumu Yamaki, Yuji Koga, Akira Nagashima, Mitsuhiro Kondo, Yoshiaki Shimada,
†
Keitaro Kadono, Ayako Moritomo, and Kosei Yoshihara
Drug Discovery Research, Astellas Pharma Inc., 21 Miyukigaoka, Tsukuba, Ibaraki
305-8585, Japan
*
Corresponding author:
Tel: +81-29-863-6550
Fax: +81-29-852-5387
E-mail: susumu.yamaki@astellas.com
†
Present address:
Haupt Pharma Toride Co., Ltd, 5662 Omonmai, Toride, Ibaraki 302-0001, Japan
1

Abstract
Vascular adhesion protein-1 (VAP-1) is a promising therapeutic target for the treatment
of diabetic nephropathy. Here, we conducted optimization studies of our lead compound
1, which we previously reported as a novel VAP-1 inhibitor, to enhance the inhibition
of human VAP-1 and to reduce CYP3A4 and CYP2C19 inhibition. As a result, we
identified
3
-chloro-4-{4-[5-(3-{[glycyl(methyl)amino]methyl}phenyl)pyrimidin-2-yl]piperazin-1-
yl}benzoic acid (17h) as a novel orally active VAP-1 inhibitor, with 14-fold increased
human VAP-1 inhibitory activity compared to 1, without CYP3A4 and CYP2C19
inhibition. Oral administration of 17h significantly inhibited the progression of
proteinuria in streptozotocin (STZ) induced diabetic rats at 0.3 and 1 mg/kg, suggesting
that this compound has potential to be a therapeutic agent for the treatment of diabetic
nephropathy.
2

Keywords
Vascular adhesion protein-1, Diabetic nephropathy, Glycine amide, CYP2C19,
CYP3A4
3

1. Introduction
Diabetic nephropathy, along with diabetic retinopathy and diabetic neuropathy, is a
major diabetic microvascular complication that develops in up to 40% of patients with
1
type 1 and type 2 diabetes. Dialysis is required as symptoms progress, which
significantly lowers the quality of life (QOL) of patients. The number of diabetic
patients has continued to increase; over 400 million people suffered from diabetes
worldwide in 2015, and this number is estimated to increase to more than 600 million
2
by 2040. With the increase in diabetes cases follows an increase in patients with
diabetic nephropathy. Furthermore, diabetic nephropathy has become the principal
1
cause of starting dialysis in recent years. Therefore, prevention of the development and
progression of diabetic nephropathy is an important issue. The exact cause of diabetic
nephropathy is unknown but hyperglycemia and resulting factors, such as acceleration
of advanced glycation end-product (AGE) formation, oxidative stress, and (glomerular)
hypertension, along with genetic susceptibility are hypothesized to be involved in its
3
development and progression. The current therapeutic strategies for diabetic
nephropathy are glycemic control, correction of (glomerular) hypertension, and dietary
1
restriction of protein. Glycemic control has been achieved to some extent by the
approval of new drugs with various mechanisms of action, such as dipeptidyl peptidase
4

4
4
(DPP4) inhibitors and sodium-glucose cotransporter 2 (SGLT2) inhibitors. Inhibition
of the renin-angiotensin system (RAS) also has a beneficial effect on the progression of
1
diabetic nephropathy. However, these current therapeutic strategies are not sufficient as
many patients still progress to end-stage renal disease and require dialysis. Therefore,
development of novel therapeutic strategies for diabetic nephropathy are needed.
Vascular adhesion protein-1 (VAP-1) was recently reported to be associated with
diabetic nephropathy, for example by inducing oxidative stress and accelerating AGE
5
6
formation. VAP-1 has two known functions: as an adhesion molecule, VAP-1 is
involved in leukocyte rolling, adhesion, and transmigration, which are central steps in
7
leukocyte extravasation to sites of inflammation; VAP-1 also acts as an amine oxidase
by catalyzing the conversion of biogenic amines into their corresponding aldehydes,
8
while releasing ammonia and hydrogen peroxide. These (cyto)toxic products
(
aldehydes and hydrogen peroxide) are known to damage the renovasculature directly or
5
indirectly for example by inducing oxidative stress and accelerating AGE formation.
Further, increased VAP-1 activity is found in patients with diabetes mellitus, and even
higher increases in patients with diabetic microvascular complications such as diabetic
5
retinopathy and diabetic nephropathy. Elevated VAP-1 expression causes increased
production of degradated products, such as aldehydes and hydrogen peroxide, and is
5

therefore considered to contribute to further deterioration of pathologies. These facts
suggest that VAP-1 is a promising novel therapeutic target for the treatment of diabetic
nephropathy.
Previously, we reported compound 1 (Table 1), a glycine amide derivative, as a novel
VAP-1 inhibitor that inhibits rat plasma VAP-1 activity by 60% after oral
9
administration at 1 mg/kg to normal rats. However, this compound inhibited human
VAP-1 with less potency than rat VAP-1 (human IC50 = 0.34 µM; rat IC50 = 0.036 µM).
To address this issue, we examined human and rat VAP-1 sequence homology. Taken
together with the results of docking analysis of 1 with human VAP-1 model, which we
9
have previously reported, it was found that two amino acid residues (Phe173/Thr,
Leu447/Phe) differed around the ligand binding pocket of 1 (Figure 1). Since these two
amino acid residues were adjacent to the left phenyl ring of 1, we hypothesized that
species differences in the potency of 1 might be reduced by altering this moiety. We
have already reported that introducing a substituent into the 2- or 3-position of the left
phenyl ring results in attenuation of VAP-1 inhibitory activity, and that the 4-position is
9
an optimal position for introducing substituent groups. On the other hand, replacement
with a hetero aromatic ring has not been attempted; thus, in this study, we converted the
left phenyl ring into a hetero aromatic ring.
6

Besides species differences, 1 also inhibited CYP3A4 and CYP2C19 (Table 1).
CYP3A4 and CYP2C19 are members of the CYP450 family and play major roles in
drug metabolism; in particular, CYP3A4 is reported to be involved in the metabolism of
1
0
almost 50% of approved drugs. CYP inhibition may mediate drug-drug interactions
(
DDIs) that can lead to serious adverse effects or a reduction in drug efficacy. A drug
with DDI risk should be carefully prescribed, especially to elderly people who often
suffer from multiple diseases and are prescribed multiple drugs. As diabetic
nephropathy is prevalent in aged diabetic patients, an agent without CYP inhibition
would be optimal for the treatment of diabetic nephropathy. Thus, optimization of
compound 1 to reduce inhibitory activity against CYP3A4 and CYP2C19 is required.
We focused on reducing CYP3A4 and CYP2C19 inhibition because inhibition of other
major isoforms (CYP1A2, CYP2C9, and CYP2D6) was not a concern for compound 1.
Here, we describe the synthesis of a novel series of glycine amide derivatives, and the
successful development of an orally active VAP-1 inhibitor without CYP3A4 and
CYP2C19 inhibition.
7

Table 1
Profile of compound 1
CYP inhibition
(Residual
VAP-1
b
Rat ex vivo
a
IC50 (µM)
c
activity)
Structure
Dose
(mg/kg,
po)
Inhibition
ratio
Human
Rat
3A4
57%
2C19
63%
(at 1 h)
0.34
0.036
1
60%
a
IC50 values are shown as the mean of independent experiments (n = 2).
b
Inhibitory effect on plasma VAP-1 activity in rats (n = 4) at 1 h after oral
administration of compound 1.
c
Residual activity of human liver microsomes (HLM) was evaluated using midazolam
(
for CYP3A4) or S-mephenytoin (for CYP2C19) as a probe substrate. Details are
described in Section 5.4.
Central phenyl ring
Morpholine moiety
(solvent side)
2
3
4
1
Right part
Left phenyl ring
(
a)
Figure 1. (a) Structure of compound 1. (b) Molecular modeling results for compound 1
pale pink, ball and stick) with human VAP-1. Phe173 and Leu447, both depicted as a
(b)
(
stick, are the residues that differ between human and rat VAP-1 （Thr173 and Phe447,
respectively, in rat VAP-1）.
8

2. Chemistry
The synthesis of glycine amide derivatives is outlined in Schemes 1–3. In Scheme 1,
1
1
Suzuki-Miyaura coupling between aryl bromide 2 with boronic esters 3a–b followed
by removal of tert-butoxycarbonyl (Boc) group afforded 5a–b. Miyaura borylation of
compound 2 gave boronic ester 6. Suzuki-Miyaura coupling of 6 with 7 followed by
deprotection of Boc group yielded 9.
3
3
a: X = N, Y = CH
b: X = Y = N
b
a
4
4
a: X = N, Y = CH
b: X = Y = N
5a: X = N, Y = CH
5b: X = Y = N
2
d
b
c
6
8
9
7
Scheme 1. Reagents and conditions: (a) 3a or 3b, Pd(PPh3)4 Na CO , DME/H
2
O,
,
2
3
2 2 2 3 2
90 °C; (b) 4 M HCl (EtOAc solution), MeOH; (c) B (Pin) , PdCl (PPh ) , KOAc,
dioxane, 80 °C; (d) 7, Pd(PPh ) Na CO , DME/H O, 90 °C.
2
3
4
,
2
3
The synthesis of 14a–b is shown in Scheme 2. Condensation of 10a–b with
N-Boc-glycine gave 11a–b. Suzuki-Miyaura coupling of 11a–b with 12 followed by
deprotection of Boc group yielded 14a–b.
9

a
b
c
1
1
0a: X = N, Y = CH
0b: X = CH, Y = N
11a: X = N, Y = CH
11b: X = CH, Y = N
13a: X = N, Y = CH
13b: X = CH, Y = N
14a: X = N, Y = CH
14b: X = CH, Y = N
1
2
Scheme 2. Reagents and conditions: (a) N-Boc-glycine, WSCD·HCl, HOBt, CH
(
2
Cl
2
b) 12, Pd(PPh ) Na CO , DME/H O, 80 °C; (c) 4 M HCl (EtOAc solution), EtOAc or
3 4, 2 3
2
;
MeOH.
Scheme 3 shows the preparation of compounds 17a–h. Suzuki-Miyaura coupling
between boronic ester 6 with 5-bromo-2-chloropyrimidine gave 15. Substitution of 15
with corresponding amines gave 16a–h. Deprotection of Boc group of 16a–d afforded
17a–d. Hydrolysis of 16e–h by NaOH and subsequent removal of Boc group yielded
carboxylic acid analogs 17e–h.
a
b
c or d, c
6
15
16a: R = (2
R
)-2-(hydroxymethyl)pyrrolidin-1-yl
17a: R = (2 )-2-(hydroxymethyl)pyrrolidin-1-yl
R
1
1
1
6b: R = (2
S
)-2-(hydroxymethyl)pyrrolidin-1-yl
17b: R = (2S)-2-(hydroxymethyl)pyrrolidin-1-yl
6c: R = piperidin-1-yl
6d: R = 4-(2-hydroxyethyl)piperidin-1-yl
17c: R = piperidin-1-yl
17d: R = 4-(2-hydroxyethyl)piperidin-1-yl
1
6e: R =
16f: R =
17e: R =
17f: R =
1
6g: R =
16h: R =
17g: R =
17h: R =
Scheme 3. Reagents and conditions: (a) 5-bromo-2-chloropyrimidine, Pd(PPh3)4,
Na CO , DME/H O, 80 °C; (b) amine, K CO , DMF, room temperature or 50–60 °C;
c) 4 M HCl (EtOAc or dioxane solution), MeOH or dioxane; (d) 1 M NaOH aq.,
solvents.
2
2
3
2
3
(
1
0

3. Results and discussion
The inhibitory activity of the synthesized compounds against human and rat VAP-1
1
4
was measured by a radioenzyme assay using C-benzylamine as an artificial substrate.
Plasma VAP-1 activity in normal rats after oral administration of test compounds was
evaluated using the same radioenzyme assay. Selected compounds were also evaluated
for inhibitory activity against CYP3A4 and CYP2C19 using midazolam and
S-mephenytoin as probe substrates, respectively.
We first examined the effect of converting each phenyl ring of 1 into a pyridine or
pyrimidine ring (Table 2). Replacing the left phenyl ring of 1 with a pyridine ring (5a)
retained rat VAP-1 inhibitory activity (IC50 = 0.037 µM). In addition, 5a showed a
6
-fold increase in inhibition of human VAP-1 compared to 1, with an IC50 value of
.055 µM. The pyrimidine analog 5b also showed increased VAP-1 inhibitory activity
0
both in rat (2-fold, IC50 = 0.015 µM) and in human (11-fold, IC50 = 0.031 µM)
compared to 1. These results suggest that the left aryl ring plays an important role in
human VAP-1 inhibitory activity, and that its conversion into a hetero aromatic ring
reduces the species differences. The enhanced potency of 5b on human VAP-1 may be
explained by electrostatic effects. We calculated the atomic charge distributions of 5b, 1,
and the electrostatic surface of the ligand binding pocket of human VAP-1 (Figure 2).
1
1

For 5b, negative charge (red) was distributed around the nitrogen atom of the
pyrimidine ring (Figure 2a), compared to a positive charge (blue) in the corresponding
region in 1 (Figure 2b). In the ligand binding pocket of human VAP-1, a positive charge
was distributed around the corresponding site (Figure 2c). While the predicted binding
mode of 5b was similar to that of 1, 5b might have higher affinity for human VAP-1
compared to 1 due to these electrostatic effects. This might be responsible for the
enhanced potency of 5b on human VAP-1 compared to that of 1. As for rat VAP-1, it
may have a relatively larger binding pocket around the left phenyl ring of 1 due to a less
bulky Thr173 residue compared to the Phe173 in human VAP-1. This larger binding
pocket in rat VAP-1 compared to human VAP-1 might have led to reduced electrostatic
effects in rat VAP-1, and this led to only a 2-fold increased rat VAP-1 inhibitory
activity of 5b compared to that of 1.
In contrast, replacing the central phenyl ring of 1 with a pyridine ring (14a, 14b)
resulted in reduced inhibition of both human and rat VAP-1. These results suggest that
conversion of the central phenyl ring into a pyridine ring is not well tolerated.
1
2

Table 2
In vitro activity of biaryl derivatives
R
a
VAP-1 IC50 (µM)
Compound
R
Human
Rat
1b
0.34
0.036
5
ac
bc
4ac
4bd
0.055
0.031
1.5
0.037
0.015
0.12
5
1
1
2.0
0.16
a
b
c
d
IC50 value is shown as the mean of independent experiments (n = 2).
Ethanedioate salt.
(2R, 3R)-tartrate salt.
Dihydrochloride salt.
1
3

(
a)
(b)
(c)
Figure 2. Atomic charge distributions of compound 5b (a), compound 1 (b), and the
electrostatic surface of the ligand binding pocket of human VAP-1 (c; the docked
1
2
compound is 1). Connolly surfaces are colored red for negative charge and blue for
positive charge.
1
4

Since 5b showed improved inhibition of human VAP-1 and reduced species
differences, we evaluated its CYP inhibition profile (Table 3). 5b exhibited reduced
inhibition of both CYP3A4 and CYP2C19 (residual activity: 91% and 74%,
respectively) compared to 1. Lowering lipophilicity is known to reduce CYP
1
0,13
inhibition,
which may explain the case here (1: CLogP = 1.22; 5b: CLogP = 0.59;
1
4
calculated using ACD LogP prediction software ). However, as 5b still inhibited
CYP2C19, we considered that there was room for further improvement. Thus, we
conducted further optimization of 5b to reduce its inhibition of CYP2C19 while
maintaining VAP-1 inhibitory activity.
Inhibition of CYP2C19 activity can be reduced by introducing polar substituent
1
0,15
groups.
For 5b, we hypothesized that the optimal position for introducing polar
substituent groups is around the morpholine moiety, since this side is predicted to be on
the solvent side (Figure 1), and is well tolerated for introduction of additional
9
substituents as we reported previously. In contrast, it would be difficult to introduce
polar substituents into the central phenyl ring and the right part (Figure 1) as steric
9
tolerance around these sites are limited. Therefore, we next investigated the
introduction of polar substituent groups around the morpholine moiety (Table 4).
1
5

Table 3
Profiles of 1 and 5b
R
CYP inhibition
VAP-1 IC50
(
Residual
a
c
Compound
R
(µM)
CLogP
b
activity)
Human
Rat
3A4
2C19
1d
0.34
0.036
0.015
57%
91%
63%
1.22
0.59
5
be
0.031
74%
a
IC50 values are shown as the mean of independent experiments (n = 2).
b
Residual activities of HLM were evaluated using midazolam (for CYP3A4) or
S-mephenytoin (for CYP2C19) as a probe substrate. Details are described in Section
5
.4.
c
14
CLogP values calculated using ACD LogP prediction software.
d
e
Ethanedioate salt.
(2R, 3R)-tartrate salt.
1
6

To evaluate the effect of introducing polar substituents, we first synthesized an
unsubstituted piperidine analog (17c), which inhibited human VAP-1 with an IC50 value
of 0.25 µM. Next, conversion of the morpholinyl group of 5b into
a
2
-hydroxymethyl-pyrrolidinyl group (17a, 17b) or 4-hydroxy-piperidinyl group (9)
resulted in 4–10-fold less potent inhibition of human VAP-1 compared to 5b. In
contrast, the 4-(2-hydroxyethyl)-piperidine derivative 17d showed equipotent activity
(
IC50 = 0.045 µM) with 5b. Compared to 17c, 9 and 17d exhibited enhanced human
VAP-1 inhibitory activities, suggesting that introduction of a hydroxyl group may
increase human VAP-1 inhibitory activity and that the position of the hydroxyl group
may play an important role in affective inhibition of human VAP-1. Introducing a
carboxyl group produced similar observations. While the isonipecotic acid analog 17e
showed approximately 55-fold decreased human VAP-1 inhibitory activity, the
3
-(piperidin-4-yl)propanoic acid analog 17f exhibited only 5-fold less potent activity
(
IC50 = 0.14 µM), compared to 5b. These results suggest that the position of the
carboxyl group also plays an important role in human VAP-1 inhibitory activity.
Regarding CYP inhibition, while 17d showed slightly reduced inhibition of CYP2C19
(
residual activity: 82%), it inhibited CYP3A4 (residual activity: 46%). In contrast, the
introduction of carboxyl group (17e, 17f) abolished inhibition of both CYP2C19 and
1
7

CYP3A4. These results suggest that introduction of a carboxyl group is effective in
reducing CYP inhibition. Accordingly, we focused on optimization of carboxylic acid
derivatives to enhance VAP-1 inhibitory activity while maintaining reduced CYP
inhibition. Since the position of the carboxyl group was important for inhibiting human
VAP-1 (17e versus 17f), we hypothesized that fixing the carboxyl group at a more
optimal position than that in 17f may improve VAP-1 inhibition. Based on this
speculation, we evaluated a benzoic acid analog, which has a more rigid structure than
17f. For convenience of synthesis, a piperazine type was synthesized. Fortunately, the
benzoic acid analog 17g exhibited more than 6-fold potent human VAP-1 inhibitory
activity compared to 17f. According to structure-activity relationship (SAR) studies of
structurally related compounds, introduction of a substituent at the 2-position of the
phenyl group at the left end is well tolerated (data not shown); thus, we evaluated a
2
-substitued benzoic acid analog. The 2-chloro derivative 17h showed equipotent
inhibition of human VAP-1 with 17g. Further, 17g and 17h both exhibited reduced CYP
inhibition, as expected.
1
8

Table 4
Substituent effect on VAP-1 potency and CYP inhibition.
Human
CYP inhibition
b
Compound
R
VAP-1 IC50
(Residual activity)
a
(
µM)
3A4
2C19
5
bc
7cd
7ad
0.031
0.25
0.30
0.24
0.13
0.045
1.7
91%
74%
e
e
1
NT
NT
e
e
1
NT
NT
7bd
NT
e
NT
e
1
c
e
e
9
NT
NT
1
7dc
46%
112%
108%
82%
104%
119%
1
7ed
1
7fd
0.14
1
7gd
0.024
0.025
100%
98%
106%
109%
1
7hd
a
IC50 values are shown as the mean of independent experiments (n = 2).
b
Residual activities of HLM were evaluated using midazolam (for CYP3A4) or
S-mephenytoin (for CYP2C19) as a probe substrate. Details are described in Section
.4.
5
1
9

c
d
e
(
2R, 3R)-tartrate salt.
Dihydrochloride salt.
NT = not tested.
To elucidate the mechanisms governing the improved potency of benzoic acid
derivatives like 17g and 17h to inhibit human VAP-1, we conducted docking analysis of
7h with human VAP-1 (Figure 3). This model revealed several interactions. First, 17h
formed three CH-π interactions with human VAP-1: central phenyl ring with Leu469,
1
1
6
pyrimidine ring with Leu447, and the hydrogen in the piperazine with Tyr394. Second,
17h formed two CH-O interactions with human VAP-1: the oxygen in the glycine
1
7
amide moiety with Leu468, and the hydrogen at the α-position of the carbonyl group
in the glycine amide moiety with the carbonyl oxygen of the Leu468 backbone. Third,
the chloro group formed a halogen-O interaction with the carbonyl oxygen of the
ASP446 backbone. In addition to these interactions, increased Van der Waals
interaction due to higher molecular weight compared to 1, and the electrostatic effects
of the pyrimidine ring described above may also contribute to potent VAP-1 inhibitory
activity of 17h. The enhancement of in vitro potency by the introduction of a carboxyl
group (17c versus 17f, 17g, and 17h) might be explained by the presence of basic
residues like His450 and Lys423 around this functional group. Although the effect of
the phenyl group on the left end and the nitrogen atom in the piperazine ring is not clear,
2
0

one of the reasons for the improved potency of benzoic acid derivatives like 17g and
7h compared to the 3-(piperidin-4-yl)propanoic acid analog 17f may be due to the
1
fixed optimal position of the carboxyl group of benzoic acid derivatives with respect to
these basic resides.
(a)
(b)
Figure 3. (a) Molecular modeling results for 1 (pale pink) and 17h (green, ball and
stick) with human VAP-1. (b) Two-dimensional diagram prepared by the ligand
1
8b
interactions application in MOE . Arrows indicate interactions.
2
1

Based on their enhanced human VAP-1 inhibitory activity without CYP3A4 and
CYP2C19 inhibition, 17g and 17h were selected for further evaluation (Table 5). As for
in vitro potency against rat VAP-1, 17g and 17h showed comparable activities to that of
human VAP-1 with IC50 values of 0.010 µM and 0.015 µM, respectively, resulting in
reduced species differences compared to 1 (1: human IC50 = 0.34 µM; rat IC50 = 0.036
µM). Encouraged by this result, we next evaluated the ex vivo efficacy of these
compounds. Oral administration of 17g and 17h at 1 mg/kg inhibited rat plasma VAP-1
activity by 68% and 96% (after 1 h), respectively. Furthermore, 17h inhibited VAP-1 in
rat plasma at lower doses, with oral administration at 0.3 mg/kg resulting in 64%
inhibition of rat plasma VAP-1. The inhibitory effects of these compounds on plasma
VAP-1 activity were similar at 1 and 6 h after oral administration. The more potent ex
vivo efficacy of 17h compared to 17g may be explained in part by the improved
aqueous solubility of 17h, especially in the presence of sodium taurocholate, compared
to 17g, since compounds with low aqueous solubility are more likely to have poor
1
9
bioavailability. The improved aqueous solubility of 17h might be due to disruption of
2
0
the molecular symmetry by introduction of a chloro group at the ortho position.
The pharmacokinetic profile of 17h is shown in Table 6. Although the bioavailability
of 17h was only 17%, oral administration at 1 mg/kg resulted in a plasma concentration
2
2

of 111 ng/mL after 1 h, which is approximately 15 times higher than its IC50 value
0.015 µM, rat). Therefore, the ex vivo efficacy of 17h at 1 h was considered reasonable.
(
Further, while the plasma concentration of 17h at 6 h post administration (7.1 ng/mL)
was just below the IC50 value, its inhibitory effect on plasma VAP-1 activity was similar
to that observed after 1 h (1 h: 96% inhibition; 6 h: 93% inhibition). Therefore, we
speculated that once 17h forms a Schiff base intermediate with topaquinone (TPQ) in
the active site of VAP-1, its dissociation may be slowed such as to prolong the duration
of its ex vivo efficacy despite a decrease in plasma concentration.
2
3

Table 5
Profiles of 17g and 17h
CYP inhibition
(Residual
a
c
d
VAP-1 IC50 (µM)
Rat ex vivo
Solubility (µM)
b
activity)
Comp
ound
R
Inhibi Inhibi
Dose
tion
tion
JP2 +
TC
Human
Rat
3A4
2C19
JP2
(
mg/kg, po) ratio
ratio
at 1 h at 6 h
1
7ge
H
0.024
0.025
0.010
0.015
100% 106%
98% 109%
1
1
68%
96%
64%
60%
93%
62%
<1
<1
1
7he Cl
2.3
89.6
0.3
a
IC50 values are shown as the mean of independent experiments (n = 2).
b
Residual activities of HLM were evaluated using midazolam (for CYP3A4) or
S-mephenytoin (for CYP2C19) as a probe substrate. Details are described in Section
5
.4.
c
Inhibitory effect on plasma VAP-1 activity in rats (n = 4) at 1 h or 6 h after oral
administration of test compounds.
d
Aqueous solubility in 2nd fluid for disintegration test in Japanese Pharmacopeia (JP2;
pH = 6.8) or in JP2 + 15 mM of sodium taurocholate (JP2 + TC).
e
Dihydrochloride salt.
2
4

Table 6
Pharmacokinetic profiles of compound 17h in rats
AUC0-24h
C
max
CLtot
t
1/2
V
dss
BA
(%)
Route
(
ng·h/mL) (ng/mL) (mL/min/kg)
(h)
(L/kg)
iv (1 mg/kg)
po (1 mg/kg)
1920
334
8.69
1.39
1.29
0.386
113
17.4
2
5

Encouraged by these results, we next investigated the in vivo efficacy of 17h on the
progression of proteinuria in STZ-induced diabetic rats. As shown in Figure 4, oral
administration of 17h at 0.3 and 1 mg/kg significantly inhibited progression of
proteinuria. 17h also significantly inhibited plasma VAP-1 activity in STZ-induced rats
after 4 weeks of administration at 0.3 and 1 mg/kg. These results suggest that
progression of proteinuria in STZ-induced rats is suppressed by inhibition of VAP-1,
and that 17h may have a beneficial effect for the treatment of diabetic nephropathy.
(a)
(b)
Figure 4. (a) Effect of oral administration of 17h on the progression of proteinuria in
STZ-induced diabetic rats. (b) Inhibitory effect on plasma VAP-1 activity 24 h after the
final dose of a 4-week 17h regimen in STZ-induced diabetic rats. The data represent the
##
mean ± SEM: ***p<0.001 versus sham group (Student’s t-test), p<0.01 versus STZ
control group (Dunnett’s multiple comparison test).
2
6

4. Conclusion
In this study, we conducted structural optimization of our lead compound 1 to identify
a novel orally active VAP-1 inhibitor for the treatment of diabetic nephropathy. Based
on molecular modeling analysis and protein sequence information, we found that
replacement of the left phenyl ring in 1 with a pyrimidine ring lead to over 10-fold
greater inhibition of human VAP-1 compared to 1. Further, reduced inhibition of
CYP3A4 and CYP2C19 while maintaining VAP-1 inhibitory activity was achieved by
the introduction of a carboxyl group at the appropriate position. These findings led to
the identification of benzoic acid analogs like 17g and 17h, which had enhanced human
VAP-1 inhibitory activity without CYP3A4 and CYP2C19 inhibition. Of these
compounds, 17h showed more potent ex vivo efficacy, with rat plasma VAP-1
inhibitory activity of 64% at 1 h after oral administration at 0.3 mg/kg. Oral
administration of 17h at 0.3 and 1 mg/kg significantly inhibited the progression of
proteinuria in STZ-induced diabetic rats, suggesting that this compound may have a
beneficial effect for the treatment of diabetic nephropathy.
2
7

5
. Experimental
.1 Chemistry
5
1
H NMR spectra were recorded on a Varian VNS-400, JEOL JNM-LA400 or JEOL
JNM-AL400 spectrometer. Chemical shifts were expressed in δ values (ppm) using
tetramethylsilane as the internal standard (s = singlet, d = doublet, t = triplet, q = quartet,
m = multiplet, dd = double doublet, dt = double triplet, and br = broad peak). Mass
spectra (MS) were recorded on a JEOL LX-2000, Waters ZQ-2000, Waters LCT
Premier mass spectrometer or Thermo Fisher Exactive Plus Orbitrap. Elemental
analyses were conducted using a Yanaco JM10 (C, H, N), Elementar Vario EL III (C, H,
N), Dionex ICS-3000 (S, halogen), and Dionex DX-5000 (S, halogen) and were within
±0.4% of theoretical values. All reactions were carried out using commercially available
reagents and solvents without further purification. The following abbreviations are used:
Ar, aryl; CHCl
3
, chloroform; CH
2
Cl
2
, dichloromethane; iPr O, diisopropyl ether; DME,
2
1
,2-dimethoxyethane; DMF, N,N-dimethylformamide; DMSO, dimethyl sulfoxide;
EtOAc, ethyl acetate; EtOH, ethanol; WSCD, 1-ethyl-3-(3’-dimethylaminopropyl)
carbodiimide HCl, hydrochloric acid; HOBt, 1-hydroxybenzotriazole; MgSO
4
,
magnesium sulfate; MeOH, methanol; KOAc, potassium acetate; K
2
CO , potassium
3
carbonate; Na
2
CO
3
, sodium carbonate; NaHCO , sodium hydrogen carbonate; NaOH,
3
2
8

sodium hydroxide; Na
2
SO , sodium sulfate; and THF, tetrahydrofuran.
4
5.1.1.
tert-Butyl
[
(
2-(methyl{3-[6-(morpholin-4-yl)pyridin-3-yl]benzyl}amino)-2-oxoethyl]carbamate
4a)
To a mixture of tert-butyl {2-[(3-bromobenzyl)(methyl)amino]-2-oxoethyl}carbamate
2; 226 mg, 0.63 mmol) in DME (2.3 mL)/water (1.1 mL) were added
(
4
-[5-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)pyridin-2-yl]morpholine (3a; 193 mg,
.66 mmol), Na CO (201 mg, 1.90 mmol), and Pd(PPh (22 mg, 0.019 mmol) under a
0
2
3
3 4
)
nitrogen atmosphere. The mixture was stirred at 90 °C for 24 h. After being cooled to
room temperature, the mixture was concentrated in vacuo. The residue was diluted with
3 4
water and extracted with CHCl . The organic layer was dried over MgSO and
concentrated in vacuo. The residue was purified by column chromatography on silica
gel (CHCl
3
2
/MeOH = 98:2), and washed with iPr O to give the product (230 mg, 83%)
1
as a colorless solid. H NMR (CDCl
3
): this compound exists as a pair of rotamers at
room temperature. δ 1.44 (minor rotamer, 9H, s), 1.46 (major rotamer, 9H, s), 2.92
major rotamer, 3H, s), 3.02 (minor rotamer, 3H, s), 3.51–3.64 (4H, m), 3.81–3.93 (4H,
m), 4.02 (major rotamer, 2H, d, J = 3.9 Hz), 4.06 (minor rotamer, 2H, d, J = 3.9 Hz),
(
2
9