Synthesis and structure activity relationships of carbamimidoylcarbamate derivatives as novel vascular adhesion protein-1 inhibitors

Article abstract of DOI:10.1016/j.bmc.2017.09.036

Vascular adhesion protein-1 (VAP-1) is a promising therapeutic target for the treatment of diabetic nephropathy. Here, we conducted structural optimization of the glycine amide derivative 1, which we previously reported as a novel VAP-1 inhibitor, to improve stability in dog and monkey plasma, and aqueous solubility. By chemical modification of the right part in the glycine amide derivative, we identified the carbamimidoylcarbamate derivative 20c, which showed stability in dog and monkey plasma while maintaining VAP-1 inhibitory activity. We also found that conversion of the pyrimidine ring in 20c into saturated rings was effective for improving aqueous solubility. This led to the identification of 28a and 35 as moderate VAP-1 inhibitors with excellent aqueous solubility. Further optimization led to the identification of 2-fluoro-3-{3-[(6-methylpyridin-3-yl)oxy]azetidin-1-yl}benzyl carbamimidoylcarbamate (40b), which showed similar human VAP-1 inhibitory activity to 1 with improved aqueous solubility. 40b showed more potent ex vivo efficacy than 1, with rat plasma VAP-1 inhibitory activity of 92% at 1 h after oral administration at 0.3 mg/kg. In our pharmacokinetic study, 40b showed good oral bioavailability in rats, dogs, and monkeys, which may be due to its improved stability in dog and monkey plasma.

Full text of DOI:10.1016/j.bmc.2017.09.036

Accepted Manuscript
Synthesis and structure activity relationships of carbamimidoylcarbamate de-
rivatives as novel vascular adhesion protein-1 inhibitors
Susumu Yamaki, Hiroyoshi Yamada, Akira Nagashima, Mitsuhiro Kondo,
Yoshiaki Shimada, Keitaro Kadono, Kosei Yoshihara
PII:
S0968-0896(17)31686-3
https://doi.org/10.1016/j.bmc.2017.09.036
BMC 13996
DOI:
Reference:
To appear in:
Bioorganic & Medicinal Chemistry
Received Date:
Revised Date:
Accepted Date:
24 August 2017
21 September 2017
22 September 2017
Please cite this article as: Yamaki, S., Yamada, H., Nagashima, A., Kondo, M., Shimada, Y., Kadono, K., Yoshihara,
K., Synthesis and structure activity relationships of carbamimidoylcarbamate derivatives as novel vascular adhesion
protein-1 inhibitors, Bioorganic & Medicinal Chemistry (2017), doi: https://doi.org/10.1016/j.bmc.2017.09.036
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 structure activity relationships of carbamimidoylcarbamate
derivatives as novel vascular adhesion protein-1 inhibitors
*
Susumu Yamaki, Hiroyoshi Yamada, Akira Nagashima, Mitsuhiro Kondo, Yoshiaki
†
Shimada, Keitaro Kadono, Kosei Yoshihara
Drug Discovery Research, Astellas Pharma Inc., 21 Miyukigaoka, Tsukuba, Ibaraki
305-8585, Japan
*
Corresponding author:
Susumu Yamaki
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 structural optimization of the glycine
amide derivative 1, which we previously reported as a novel VAP-1 inhibitor, to
improve stability in dog and monkey plasma, and aqueous solubility. By chemical
modification of the right part in the glycine amide derivative, we identified the
carbamimidoylcarbamate derivative 20c, which showed stability in dog and monkey
plasma while maintaining VAP-1 inhibitory activity. We also found that conversion of
the pyrimidine ring in 20c into saturated rings was effective for improving aqueous
solubility. This led to the identification of 28a and 35 as moderate VAP-1 inhibitors
with excellent aqueous solubility. Further optimization led to the identification of
2
-fluoro-3-{3-[(6-methylpyridin-3-yl)oxy]azetidin-1-yl}benzyl
carbamimidoylcarbamate (40b), which showed similar human VAP-1 inhibitory activity
to 1 with improved aqueous solubility. 40b showed more potent ex vivo efficacy than 1,
with rat plasma VAP-1 inhibitory activity of 92% at 1 h after oral administration at 0.3
mg/kg. In our pharmacokinetic study, 40b showed good oral bioavailability in rats, dogs,
and monkeys, which may be due to its improved stability in dog and monkey plasma.
2

Keywords
Vascular adhesion protein-1, Diabetic nephropathy, Carbamimidoylcarbamate, Stability
in plasma, Aqueous solubility.
3

1. Introduction
Diabetic nephropathy develops in up to 40% of patients with type 1 and type 2
diabetes, and along with diabetic retinopathy and diabetic neuropathy, is a major
1
diabetic microvascular complication. As symptoms progress, dialysis is required,
which markedly lowers the quality of life (QOL) of patients. Further, 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
2
million by 2040. The increase in diabetes cases is paralleled by an increase in people
with diabetic nephropathy. Furthermore, diabetic nephropathy has become the primary
1
reason for starting dialysis in recent years. Therefore, prevention of the development
and progression of diabetic nephropathy is an important issue. Although the exact cause
of diabetic nephropathy is not clear, its development and progression have been
hypothesized to be linked with hyperglycemia and resulting factors, such as acceleration
of advanced glycation end-product (AGE) formation, oxidative stress, and (glomerular)
3
hypertension, along with genetic susceptibility. Current therapeutic strategies for
diabetic nephropathy are glycemic control, correction of (glomerular) hypertension, and
1
dietary restriction of protein. Recently, glycemic control has been achieved to some
extent by the approval of novel class of drugs with various mechanisms of action, such
4

as dipeptidyl peptidase 4 (DPP4) inhibitors and sodium-glucose cotransporter 2
4
SGLT2) inhibitors. Inhibition of the renin-angiotensin system (RAS) also has a
(
1
beneficial effect on the progression of diabetic nephropathy. However, as many
patients still progress to end-stage renal disease and start dialysis, these current
therapeutic strategies are not sufficient. Therefore, development of novel therapeutic
strategies for diabetic nephropathy are needed.
Vascular adhesion protein-1 (VAP-1) was recently linked to diabetic nephropathy;
5
for example, it was reported to induce oxidative stress and accelerate AGE formation.
VAP-1 is
a
member of the family of copper-containing amine
oxidases/semicarbazide-sensitive amine oxidase (AOC/SSAO), which is encoded by the
AOC3 gene. VAP-1 has been reported to function as both an adhesion molecule and
6
amine oxidase. As an adhesion molecule, VAP-1 is involved in leukocyte rolling,
adhesion, and transmigration, which are key steps in leukocyte extravasation to sites of
7
inflammation. As an amine oxidase, VAP-1 catalyzes the conversion of biogenic
primary amines into their corresponding aldehydes, while releasing ammonia and
8
hydrogen peroxide by the reaction with topaquinone (TPQ) in its active site. Increased
VAP-1 activity is found in patients with diabetes mellitus, and even higher activities
have been observed in patients with diabetic microvascular complications, such as
5

5
diabetic retinopathy and diabetic nephropathy. Elevated VAP-1 activity causes the
production of large amounts of degradated products, such as aldehydes and hydrogen
peroxide, which are (cyto)toxic products known to damage the renovasculature directly
5
or indirectly for example by inducing oxidative stress and accelerating AGE formation.
Therefore, VAP-1 is a possible contributing factor for further deterioration of
pathologies, suggesting that it is a promising novel therapeutic target for the treatment
of diabetic nephropathy.
We previously reported compound 1 (Figure 1), a glycine amide derivative, as a novel
VAP-1 inhibitor that shows potent human VAP-1 inhibitory activity with an IC50 value
of 0.025 µM, and inhibition of urinary protein excretion in streptozotocin
9
STZ)-induced diabetic rats after oral administration at 0.3 and 1 mg/kg. While 1 was
(
stable in rat and human plasma, it was unstable in dog and monkey plasma, with less
than 1% detected after incubation with dog and monkey plasma for 6 h (Table 1). Since
1
is unstable in dog and monkey plasma, its pharmacokinetic properties in these species
is likely to be poor, suggesting that safety assessment of 1 in these species would be
difficult. Since the International Conference on Harmonization Guideline ICH M3 states
that some nonclinical safety studies should be carried out in both rodent and non-rodent
1
species to support the safe conduct of clinical trials, we proposed that structural
0
6

optimization of 1 was required to improve its stability in plasma and therefore
pharmacokinetic properties in non-rodent species, such as dogs and monkeys.
Since structurally related compounds that have the same right part with 1 (Figure 1)
were also unstable in dog and monkey plasma (data not shown), we suspected that
chemical groups in this part of the compound may be responsible for its instability in
2 2
dog and monkey plasma. As this right part contains a CH -NH group, it may be a
suitable substrate for SSAO. It was previously reported that mammals other than
5
c, 11
humans have much higher plasma SSAO activity than humans and rats.
For
example, the Vmax (mU/L) of plasma SSAO in dogs, humans, and rats are 9000 ± 1000,
5
52 ± 102, and 27, respectively. Therefore, we hypothesized that 1, a potential
c
3
substrate for SSAO, might be oxidized rapidly by high SSAO activity, making it
unstable in dog and monkey plasma. We therefore focused our efforts on identifying
2 2
novel VAP-1 inhibitors lacking CH -NH group in the right part. Inoue et al reported
2 2
the guanidine analog 2 as a VAP-1 inhibitor without the CH -NH group with moderate
human VAP-1 inhibitory activity (IC50 = 0.23 µM), and rat plasma VAP-1 inhibitory
1
activity after subcutaneous administration. Since guanidine derivatives can not be
2
2 2
substrates for SSAO due to the absence of CH -NH group, we expected that they are
stable in dog and monkey plasma. Therefore, we designed acylguanidine,
7

carbamimidoylcarbamate, and carbamimidoylurea analogs as bioisosteres of the right
2 2
part of 1 without CH -NH group.
1
also exhibited poor aqueous solubility (2.3 µM) in the Japanese Pharmacopoeia 2nd
fluid for disintegration test (JP2; pH = 6.8) (Table 1). Therefore, improvement of the
aqueous solubility of 1 is also required since compounds with low aqueous solubility
1
3
tend to show poor bioavailability and dose proportionality.
Here, we describe the synthesis of a novel series of carbamimidoylcarbamate
derivatives, and the successful development of an orally active VAP-1 inhibitor with
improved aqueous solubility and good pharmacokinetic properties in rats, dogs, and
monkeys.
Central phenyl ring
Right part
1
2
Right part
Figure 1. Structures of compound 1 and 2
Table 1
Profile of compound 1
b
Comp Human VAP-1 Residual ratios (% of initial concentration)
Solubility
c
a
ound
IC50 (µM)
Human
99.2
Rat
104
Dog
<1
Monkey
<1
(µM)
d
1
0.025
2.3
8

a
IC50 values are shown as the mean of independent experiments (n = 2).
b
Residual ratios (% of initial concentration) of compound 1 after 6 h incubation with
plasma from each species.
c
Aqueous solubility in the Japanese Pharmacopoeia 2nd fluid for disintegration test
(
JP2; pH = 6.8).
d
Dihydrochloride salt.
9

2. Chemistry
The synthesis of target compounds is outlined in Schemes 1–7. Lithium-halogen
exchange of 3 and subsequent formylation gave aldehyde 4. The Wittig reaction
between aldehyde
4
and (3-nitrobenzyl)(triphenyl)phosphonium bromide or
(
4-nitrobenzyl)(triphenyl)phosphonium bromide followed by hydrogenation in the
presence
of
Pd/C
gave
5a–b.
The
reaction
followed
with
by
N,N´-bis(tert-butoxycarbonyl)-1H-pyrazole-1-carboximidamide
deprotection of tert-butoxycarbonyl (Boc) group yielded 6a–b. Boc protection of 7
followed by benzylic bromination and phosphonium salt formation gave 8. The Wittig
reaction between aldehyde 4 and 8 and subsequent hydrogenation yielded 9. Removal of
Boc group of 9 gave phenethylamine intermediate, which was converted to 10 in a
manner similar to that for 6a–b.
4
3
4
3
a
b, c
d, e
5
5
a: 3-NH2
6a: 3-guanidino
6b: 4-guanidino
3
4
b: 4-NH2
f, g, h
i, c
e, d, e
-
+
7
8
9
10
Scheme 1. Reagents and conditions: (a) n-BuLi, THF, -78 °C, then DMF, -78 °C to
°C; (b) (3-nitrobenzyl)(triphenyl)phosphonium bromide or
4-nitrobenzyl)(triphenyl)phosphonium bromide, t-BuOK, THF, 0 °C to room
temperature; (c) Pd/C, EtOH/THF; (d)
0
(
2
H ,
1
0

N,N´-bis(tert-butoxycarbonyl)-1H-pyrazole-1-carboximidamide, DIPEA, THF; (e) 4 M
HCl (EtOAc solution), MeOH; (f) Boc O, Et N, THF; (g) NBS, AIBN, CCl , reflux; (h)
PPh , toluene, 100 °C; (i) 4, t-BuOK, THF, 0 °C to room temperature.
2
3
4
3
Scheme 2 shows the synthesis of acylguanidine analog 14. The Wittig reaction
between aldehyde 4 and phosphonium salt 11 followed by acid-catalyzed esterification
gave 12, which was hydrogenated to yield 13. Hydrolysis of 13 using NaOH gave the
carboxylic acid intermediate, which was activated by carbonyldiimidazole (CDI) and
then reacted with guanidinium carbonate to give 14.
a, b
c
d, e
-
+
1
1
12
13
14
Scheme 2. Reagents and conditions: (a) 4, NaH, DMF, -20 °C to room temperature; (b)
conc.H SO , EtOH, 70 °C; (c) H , Pd/C, EtOH; (d) 1M NaOH aq., EtOH; (e) CDI,
DMF, 0 °C to room temperature, then guanidinium carbonate.
2
4
2
2
0a–h and 24 were prepared according to Scheme 3. Miyaura borylation of compound
15a–b gave boronic ester 16a–b. Suzuki-Miyaura coupling between 17 and various
substituted boronic acids or esters, including 16a–b, or between 18 and
(
3-bromophenyl)methanol gave 19a–h. 20a was synthesized from ester 19a in a manner
similar to that for 14. Reductive amination of 19b with methylamine followed by the
reaction with CDI and guanidinium carbonate yielded 20b. Benzyl alcohol derivatives
1
1

19c–e were activated by CDI and then reacted with guanidinium salt to give 20c–e.
4
Reduction of 19f–h using NaBH afforded benzyl alcohol intermediates, which were
converted to 20f–h in a manner similar to that for 20c–e. Substitution of 21 with
morpholine followed by bromination using N-bromosuccinimide (NBS) gave 22, which
was reacted with (2-fluoro-3-formylphenyl)boronic acid under Suzuki-Miyaura reaction
condition to afford 23. 23 was converted into 24 in a manner similar to that for 20f–h.
a
OH
O H
Br
PinB
R2
R2
1
1
5a: R2 = Me
5b: R2 = Cl
16a: R2 = Me
16b: R2 = Cl
d, e (for 20a)
or
f, g (for 20b)
or
e (for 20c-e)
or
h, e (for 20f-h)
R1
R1
H
N
Br
X
NH2
NH
N
N
N
X
R2
O
N
N
b
c
R2
N
N
N
N
O
O
O
1
7
2
2
2
2
2
2
2
2
0a: R1 = R = H, X = bond
1
9a: R1 = R2 = H, X = CH CO Et
2
2
2
0b: R1 = R = H, X = NMe
BPin
19b: R1 = R2 = H, X = CHO
2
N
0c: R1 = R2 = H, X = O
1
1
1
1
1
1
9c: R = R2 = H, X = CH OH
1 2
N
N
0d: R1 = H, R = Me, X = O
9d: R1 = H, R2 = Me, X = CH OH
2
O
2
0e: R1 = H, R = Cl, X = O
9e: R1 = H, R2 = Cl, X = CH OH
2
2
1
8
0f: R1 = F, R2 = H, X = O
9f: R1 = F, R = H, X = CHO
2
0g: R1 = H, R = F, X = O
9g: R1 = H, R2 = F, X = CHO
9h: R1 = R2 = F, X = CHO
2
0h: R1 = R = F, X = O
2
Me
Me
Me
Me
H
N
Br
O
NH2
NH
i, j
N
b
h, e
N
CHO
N
N
F
F
O
N
N
N
N
N
N
Cl
N
O
O
O
2
1
22
23
24
Scheme
3.
Reagents
and
conditions:
(PPh , KOAc, dioxane,
, solvents, 80 °C or Ar-B(OH)
O, 100 °C; (c) (3-bromophenyl)methanol,
O, 80 °C; (d) 1 M NaOH aq., EtOH/THF; (e) CDI, DMF,
–70 °C, then guanidinium carbonate or guanidinium chloride, NaH; (f) methylamine
hydrochloride, NaOAc, NaBH(OAc) , AcOH, 0 °C to room temperature; (g) CDI, DMF,
°C to room temperature, then guanidinium carbonate, DBU, 70 °C (h) NaBH
MeOH/THF, 0 °C to room temperature; (i) morpholine, EtOH, 80 °C; (j) NBS, CH Cl
(a)
4
,4,4',4',5,5,5',5'-octamethyl-2,2'-bi-1,3,2-dioxaborolane, PdCl
0 °C; (b) Ar-B(OH) or Ar-B(Pin), Pd(PPh , Na CO
, SPhos, K PO , toluene/H
, Na CO , DME/H
2
3 2
)
8
2
3
)
4
2
3
2
,
Pd(OAc)
Pd(PPh
2
3
4
2
3
)
4
2
3
2
0
3
0
4
,
,
2
2
1
2

0
°C to room temperature.
The synthesis of 28a–b and 32a–e is shown in Scheme 4 and Scheme 5, respectively.
Compound
26
was
prepared
from
25
by
alkylation
with
1
-(chloromethyl)-4-methoxybenzene. Buchwald-Hartwig cross coupling reaction
between 26 and corresponding amines followed by removal of para-methoxybenzyl
(
PMB) group afforded benzyl alcohol intermediates 27a–b, which were converted to
28a–b in a manner similar to that for 20c–e. Compound 30a was commercially
available
and
30b
was
prepared
from
29
by
silylation
with
tert-butyl(chloro)dimethylsilane.
30a–b
and
various
amines
underwent
Buchwald-Hartwig cross coupling reaction followed by removal of
tert-butyldimethylsilyl (TBS) group gave 31a–e. 31a–e were converted into 32a–e in a
manner similar to that for 20c–e.
a
b, c
d
2
5
26
2
7a: R =
7b: R =
28a: R =
28b: R =
2
Scheme 4. Reagents and conditions: (a) 1-(chloromethyl)-4-methoxybenzene, NaH,
DMF, 0 °C to room temperature; (b) amine, Pd (dba) , BINAP, t-BuONa, toluene,
00 °C; (c) TFA, CH Cl , 0 °C; (d) CDI, DMF, then guanidinium carbonate.
2
3
1
2
2
1
3

a
b, c
d
H
N
OTBS
O H
O
NH2
NH
O H
R
Br
R
Br
X
X
N
X
O
N
F
N
N
3
0a: X = H
0b: X = F
2
9
N
N
3
1a: X = H, R =
32a: X = H, R =
3
N
N
N
N
N
N
N
N
3
3
1b: X = H, R =
1c: X = F, R =
32b: X = H, R =
32c: X = F, R =
N
N
N
N
N
N
3
3
1d: X = F, R =
1e: X = F, R =
32d: X = F, R =
32e: X = F, R =
N
N
O
O
N
N
N
N
O
O
Scheme 5. Reagents and conditions: (a) tert-butyl(chloro)dimethylsilane, 1H-imidazole,
THF; (b) amine, Pd (dba) , BINAP, t-BuONa, toluene, 80–100 °C; (c) 1 M HCl aq. or
2
3
TBAF, THF; (d) CDI, DMF, then guanidinium carbonate.
Scheme 6 shows the synthesis of 35. Coupling reaction between 30a and
1
-Boc-piperazine gave 33. Removal of both Boc group and TBS group in 33 using HCl
followed by reductive amination with tetrahydro-4H-pyran-4-one yielded 34. 34 was
converted to 35 in a manner similar to that for 20c–e.
a
b, c
d
3
0a
33
34
35
Scheme 6. Reagents and conditions: (a) 1-Boc-piperazine, Pd
2
(dba)
3
, BINAP, t-BuONa,
toluene, 100 °C; (b) 4 M HCl (EtOAc solution), MeOH; (c) tetrahydro-4H-pyran-4-one,
NaBH(OAc) , AcOH, CH Cl ; (d) CDI, DMF, then guanidinium carbonate.
3
2
2
1
4

4
0a–b were prepared according to Scheme 7. Alkylation of 36 with tert-butyl
3
-iodoazetidine-1-carboxylate followed by removal of Boc group gave 38a. 38b was
prepared
from
37
by
the
Mitsunobu
reaction
with
tert-butyl
3
-hydroxyazetidine-1-carboxylate and subsequent deprotection of Boc group.
Buchwald-Hartwig cross coupling reaction between 38a–b and 30b followed by
deprotection of TBS group afforded 39a–b, which were converted to 40a–b in a manner
similar to that for 20c–e.
a, b
3
6
d, e
f
c, b
3
3
8a: X = Me, Y = H
8b: X = H, Y = Me
39a: X = Me, Y = H
39b: X = H, Y = Me
40a: X = Me, Y = H
40b: X = H, Y = Me
3
7
Scheme 7. Reagents and conditions: (a) tert-butyl 3-iodoazetidine-1-carboxylate,
CO , DMF, 100 °C; (b) TFA, DCE; (c) tert-butyl 3-hydroxyazetidine-1-carboxylate,
DIAD, PPh , THF, 55 °C; (d) 30b, Pd (dba) , BINAP, t-BuONa, toluene, 90 °C; (e)
TBAF, THF; (f) CDI, DMF, then guanidinium carbonate.
K
2
3
3
2
3
1
5

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 the test compounds
was evaluated using the same radioenzyme assay. Selected compounds were also
evaluated for aqueous solubility in JP2 and for inhibitory activity against CYPs.
In converting the right part of 1, we evaluated 2-(morpholin-4-yl)pyrimidin-5-yl
9
derivatives and compared them with 41 , which has equipotent VAP-1 inhibitory
activity to 1 (IC50 = 0.031 µM), because compound 41 exhibits two advantages:
convenience of synthesis and excellent aqueous solubility (≥ 100 µM). We designed
carbamimidoylurea analog 20b, carbamimidoylcarbamate analog 20c, and
2 2
acylguanidine analog 20a as bioisosteres of the right part in 41 without CH -NH group.
While 20b showed more than 10-fold reduced potency compared to 41, 20c and 20a
retained VAP-1 inhibitory activity. Since there was very limited steric tolerance around
1
4
the carbonyl group in the right part of the glycine amide derivatives , bulkiness of the
methyl group adjacent to the carbonyl group in 20b might be responsible for the
reduced potency. Further synthesis of (acyl) guanidine derivatives was carried out to
elucidate the structure activity relationships (SARs). Since compound 2 has the ethylene
1
6

linker (Figure 1), we introduced an ethylene linker into 20a. While compound 14
retained in vitro potency, removal of the carbonyl group resulted in loss of activity (10).
This suggests that the carbonyl group in the right part is important for VAP-1 inhibition,
1
which is consistent with the SARs of glycine amide derivatives . Phenylguanidine
4
analogs (6a and 6b) also exhibited reduced VAP-1 inhibitory activity. Considering the
1
importance of the thiazole-amide moiety in compound 2 for VAP-1 inhibition, the loss
2
of VAP-1 inhibitory activity in 6a and 6b may suggest that the
2
-(morpholin-4-yl)pyrimidin-5-yl moiety in these compounds is not suitable for the
VAP-1 binding pocket. From the chemical modification of the right part of the glycine
amide derivative, we identified acylguanidine analogs (20a and 14) and the
2 2
carbamimidoylcarbamate analog 20c as novel VAP-1 inhibitors without CH -NH
group. Of these compounds, 20c was evaluated for stability in plasma because
acylguanidine analogs were degraded in JP2 (data not shown), suggesting low chemical
stability.
1
7

Table 2
SARs of 2-(morpholin-4-yl)pyrimidin-5-yl derivatives
Human
Compound
R
VAP-1 IC50
a
(
µM)
b
4
1
0.031
0.53
0.047
0.056
0.054
>10
b
0b
2
b
2
0c
b
2
0a
b
14
c
1
6
6
0
c
c
a
>10
b
5.1
a
b
c
IC50 values are shown as the mean of independent experiments (n = 2).
2R, 3R)-tartrate salt.
Dihydrochloride salt.
(
1
8

As shown in Table 3, 20c was more stable in dog and monkey plasma compared to 1,
as expected, with no apparent decrease observed after incubation with plasma from each
species for 6 h. This suggests that conversion of the right part into the
2 2
carbamimidoylcarbamate (without CH -NH group) from a glycine amide (with
2 2
CH -NH group) is important for stability in dog and monkey plasma, while
maintaining VAP-1 inhibitory activity.
However, 20c showed 2-fold decreased inhibition of VAP-1 compared to 1 and low
aqueous solubility (8.3 µM); therefore, further optimization of 20c was conducted to
increase potency and aqueous solubility. Lowering the lipophilicity is a well-known
1
strategy for increasing aqueous solubility, with Curatolo et al reporting that a LogP
5
1
6
value < 3.25 is desirable for a typical small molecule to achieve a solubility >100 µM.
Interestingly, while 41 and 20c have similar cLogP values below 3.25 (20c = 0.63, 41 =
0.59), their aqueous solubility was quite different (20c = 8.3 µM, 41 ≥ 100 µM). Along
with solubility, lipophilicity is also known to play a crucial role in determining
absorption, distribution, metabolism, elimination, and toxicology (ADMET)
1
properties, with a LogP range of 0-3 being optimal for achieving good oral
5
1
bioavailability. Therefore, lowering the lipophilicity may not be an appropriate
7
strategy for improving the aqueous solubility of 20c because its lipophilicity is already
1
9

1
7
near the lower limit of the recommended range.
The difference in aqueous solubility between 20c and 41 suggests that the
carbamimidoylcarbamate moiety in 20c may be responsible for its poor aqueous
solubility. However, as this moiety is essential for both the inhibition of VAP-1 and the
compound’s stability in dog and monkey plasma, we decided that this moiety should not
be modified. In addition to lowering the lipophilicity, disruption of molecular planarity
1
or symmetry are also known strategies for increasing aqueous solubility. Therefore,
8
we focused on modifying the biaryl moiety of 20c.
2
0

Table 3
Profiles of 1, 20c, and 41.
Human
Residual ratios (% of initial
b
concentration)
Comp
ound
Solubility
c
d
Structure
VAP-1
cLogP
3.38
(µM)
a
IC50 (µM)
Human
99.2
Rat
Dog
<1
Monkey
<1
e
1
0.025
0.047
0.031
104
2.3
8.3
f
2
0c
101
105
98.3
103
0.63
f
g
41
NT
≥100
0.59
a
IC50 values are shown as the mean of independent experiments (n = 2).
b
Residual ratios (% of initial concentration) of test compound after 6 h incubation with
plasma from each species.
c
Aqueous solubility in the Japanese Pharmacopoeia 2nd fluid for disintegration test
(
JP2; pH = 6.8).
d
19
cLogP values calculated using ACD LogP prediction software.
Dihydrochloride salt.
e
f
(2R, 3R)-tartrate salt.
g
NT = not tested.
2
1

To improve aqueous solubility by disruption of planarity, Kasuga et al introduced
2
substituents into the ortho position of the biaryl structure . Therefore, we also
0
investigated the effect of introducing substituents into the ortho position of the biaryl
1
moiety in 20c (Table 4). While introduction of a fluorine atom into the R position (20f)
resulted in 2-fold decreased inhibition of VAP-1 compared to 20c, aqueous solubility
was improved (68 µM). This suggests that disruption of planarity by introduction of a
substituent into the ortho position of the biaryl moiety is also effective for increasing
2
aqueous solubility in the present case. In contrast, 20g (R = F) showed 5-fold enhanced
potency with an IC50 value of 0.0092 µM, but its aqueous solubility was not improved
2
compared to that of 20c. Since the introduction of a substituent into the R position
enhanced potency, further SAR studies were conducted around 20g to increase VAP-1
inhibitory activity and aqueous solubility. Replacement of the fluorine atom in 20g with
more bulky substituents (20d and 20e) resulted in a significant reduction in activity.
2
This suggests that the steric tolerance at the R position is limited and that a fluorine
atom is an optimal substituent for VAP-1 inhibition. We next evaluated the effect of
1 3
additional substituents at the R position (20h) or the R position (24) of 20g. 20h
showed an 8-fold reduction in VAP-1 inhibitory activity with slightly improved aqueous
solubility compared to 20g. While 24 exhibited more than 10-fold improved aqueous
2
2

solubility (90 µM) compared to that of 20g, its potency was reduced over 40-fold. These
results (Table 4) suggest that the introduction of a fluorine atom into the R position
2
enhances VAP-1 inhibitory activity, while introduction of substituents into the ortho
position of the biaryl moiety can improve aqueous solubility in some cases, such as in
20f and 24. However, it is difficult to improve aqueous solubility while maintaining
VAP-1 inhibition by only introducing substituents into the ortho position. Thus, we
attempted an alternative approach to improve aqueous solubility.
2
3

Table 4
SARs of carbamimidoylcarbamate derivatives
Human
Solubility
Compound
R
1
R
2
R
3
VAP-1 IC50
a
b
(
µM)
8.3
68
(
µM)
c
2
0c
H
F
H
H
H
H
0.047
0.099
0.0092
0.41
d
2
0f
e
c
2
0g
H
H
H
F
F
H
7.5
34
2
0d
Me
Cl
F
H
d
c
2
0e
H
0.60
8.2
17
2
0h
H
0.077
0.40
c
2
4
H
F
Me
90
a
IC50 values are shown as the mean of independent experiments (n = 2).
Aqueous solubility in the Japanese Pharmacopoeia 2nd fluid for disintegration test
JP2; pH = 6.8).
b
(
c
(2R, 3R)-tartrate salt.
d
e
Dihydrochloride salt.
Hydrochloride salt.
2
4

Replacing one of the aryl rings in the biaryl structure with a saturated ring is also a
2
known strategy for increasing aqueous solubility by disruption of planarity. To
1
attempt this strategy, we converted the pyrimidine ring in 20c into a saturated ring
(
Table 5), since conversion of the central phenyl ring may lead to loss of VAP-1
inhibition due to the altered position of the right part that interacts with the active site of
VAP-1. As expected, aqueous solubility of piperidine analog 28a and piperazine analog
35 was markedly improved (≥100 µM), but VAP-1 inhibition was reduced by 2–3 fold
(
IC50 = 0.16 and 0.11 µM, respectively) compared to 20c. These results suggest that
conversion of the pyrimidine ring in 20c into a saturated ring, such as a piperidine ring
or piperazine ring, is effective for improving aqueous solubility. Encouraged by the
remarkable improvement in solubility with only a 2-fold reduction in VAP-1 inhibition,
we conducted further optimization of 35 to increase VAP-1 inhibitory activity.
9
, 14
Based on SAR studies of glycine amide derivatives,
we speculated that introduction
of polar substituents around the tetrahydropyran ring in 35 may increase VAP-1
inhibition. We therefore evaluated the effects of replacing the tetrahydropyran ring in 35
with a pyridine ring. VAP-1 inhibitory activity was 4-pyridyl > 3-pyridyl > 2-pyridyl
(
28b > 32a > 32b). While 28b and 32a showed more potent inhibition of VAP-1
compared to 20c, and excellent aqueous solubility, these compounds also inhibited
2
5

CYPs, with 28b in particular showing potent CYP inhibition against CYP2C8 and
CYP2D6. Since CYP inhibition may mediate drug-drug interactions (DDIs) that can
2
2
lead to serious adverse effects or a reduction in drug efficacy , we sought to reduce
CYP inhibition. Since 32a exhibited an appropriate balance in VAP-1 inhibition,
aqueous solubility, and CYP inhibition, we applied further modifications to 32a.
2
Previous SAR showed that the introduction of a fluorine atom to the R position
increased VAP-1 inhibitory activity by 5-fold (20c versus 20g; Table 4); therefore, we
synthesized 32c by introducing a fluorine atom into the central phenyl ring of 32a.
While 32c exhibited lower aqueous solubility, it showed enhanced VAP-1 inhibitory
activity compared to 32a. To improve the aqueous solubility of 32c by increasing the
molecular flexibility, we designed a compound with an oxygen linker (32d). 32d
showed excellent aqueous solubility but exhibited decreased VAP-1 inhibition and more
potent CYP inhibition against CYP2D6 and CYP3A4 compared to 32c. The potent CYP
inhibition of 32d may be due to the nitrogen atom in the pyridine ring, which is more
2
3
likely to be recognized by CYP enzyme. Therefore, we modified the position of the
nitrogen atom in the pyridine ring by replacing the piperidine ring with an azetizine ring.
The azetidine derivative 32e exhibited increased VAP-1 inhibitory activity with reduced
inhibition of CYP3A4 compared to 32d; however, inhibitory activity against CYP2D6
2
6

was still potent. To further improve CYP inhibition profiles of 32e, a methyl group was
introduced adjacent to the nitrogen atom in the pyridine ring. While the
2
-methyl-pyridyl analog 40a inhibited CYP2C8 with reduced VAP-1 inhibition
compared to 32e, the 6-methyl-pyridyl analog 40b retained in vitro potency (IC50
=
0
.025 µM) and exhibited relatively low CYP inhibition against the six isoforms in Table
. The improved CYP inhibition profiles of 40a and 40b may be due to the disruption of
5
the binding of the pyridyl nitrogen atom to the heme portion of the CYP enzyme by the
23
bulkiness of the methyl group next to the pyridine nitrogen atom .
2
7

Table 5
SARs of carbamimidoylcarbamate derivatives
Human
c
CYP inhibition (residual activity, %)
Comp
ound
VAP-1 Solubilit
b
IC50
R
1
X
y (µM)
1
A2
2C8
83
2C9 2C19 2D6
3A4
109
a
(
µM)
d
0c
2
H
H
H
H
H
H
F
0.047
0.16
8.3
96
92
102
93
e
g
g
2
8a
≥100
≥100
≥100
≥100
≥100
31
NT
NT
d
3
5
0.11
e
8b
2
0.013
0.031
0.14
86
55
25
92
80
91
99
13
68
80
64
e
32a
101
e
2b
g
3
NT
d
d
3
2c
0.017
0.045
0.024
0.040
0.025
94
93
84
89
99
99
61
78
46
91
105
87
96
96
66
91
99
96
11
16
77
95
61
32
56
96
85
3
2d
F
≥100
≥100
≥100
≥100
d
3
2e
F
97
d
40a
F
89
d
0b
4
F
92
a
IC50 values are shown as the mean of independent experiments (n = 2).
b
Aqueous solubility in the Japanese Pharmacopoeia 2nd fluid for disintegration test
JP2; pH = 6.8).
(
2
8

c
Residual activitie of human liver microsomes (HLM) was evaluated. Details are
described in Section 5.5
d
(2R, 3R)-tartrate salt.
e
Trihydrochloride salt.
NT = not tested.
g
.
Based on its similar human VAP-1 inhibitory activity with 1 and improved aqueous
solubility without CYP inhibition, 40b was selected for further evaluation (Table 6).
40b inhibited rat VAP-1 with an IC50 value of 0.0052 µM. Oral administration of 40b at
0.3 mg/kg inhibited rat plasma VAP-1 activity by 92% (after 1 h) and 74% (after 6 h).
9
This ex vivo efficacy of 40b was more potent than that of 1 , which may be due to its
enhanced rat VAP-1 inhibitory activity and its improved pharmacokinetic profile in rats
9
(Cmax (ng/mL): 1 = 113 , 40b = 401; bioavailability (%): 1 = 17 , 40b = 73).
9
Furthermore, as shown in Table 7, 40b also exhibited favorable pharmacokinetic
profiles in dogs and monkeys, with low total clearance (CLtot (mL/min/kg): 8.9 and 0.7,
respectively) and good oral bioavailability (78% and 85%, respectively). While
pharmacokinetic evaluation of 1 in dogs and monkeys could not be performed due to
instability in dog and monkey plasma, the pharmacokinetic profile of 40b in these
species was desirable. This may be attributed to the improved stability of 40b in dog
and monkey plasma by conversion of the right part into the carbamimidoylcarbamate
(
2 2 2 2
without CH -NH group) from the glycine amide (with CH -NH group) of 1.
2
9