Noncovalent tripeptidic thrombin inhibitors incorporating amidrazone, amine and amidine functions at P1

Article abstract of DOI:10.1016/S0960-894X(02)00093-8

A series of noncovalent tripeptidic thrombin inhibitors incorporating amidrazone, amine and amidine functions at P1 was investigated. While the amidrazone and the amine series displayed limited oral absorption, the amidine series demonstrated generally good oral absorption and strong antithrombotic activity; the single-digit picomolar K_i achieved from this series is among the best yet reported. The present work highlights the benzamidine compound 11f (LB30812) that exhibits excellent overall profiles of potency, oral absorption and antithrombotic efficacy.

Full text of DOI:10.1016/S0960-894X(02)00093-8

Bioorganic & Medicinal Chemistry Letters 12 (2002) 1017–1022
Noncovalent Tripeptidic Thrombin Inhibitors Incorporating
Amidrazone, Amine and Amidine Functions at P1
Koo Lee,* Won-Hyuk Jung, Cheol Won Park, Hee Dong Park,
Sun Hwa Lee and O Hwan Kwon
Life Science R&D, LGCI, PO Box 61 Yu-Sung, Science Town, Taejon 305-380, Republic of Korea
Received 3 December 2001; accepted 1 February 2002
Abstract—A series of noncovalent tripeptidic thrombin inhibitors incorporating amidrazone, amine and amidine functions at P1
was investigated. While the amidrazone and the amine series displayed limited oral absorption, the amidine series demonstrated
generally good oral absorption and strong antithrombotic activity; the single-digit picomolar Ki achieved fromthis series is among
the best yet reported. The present work highlights the benzamidine compound 11f (LB30812) that exhibits excellent overall profiles
of potency, oral absorption and antithrombotic efficacy. # 2002 Elsevier Science Ltd. All rights reserved.
Thrombin, a crucial enzyme in the blood coagulation,
has been a target for antithrombotic therapy. Orally
active thrombin inhibitors would provide effective and
safe prophylaxis for venous and arterial thrombosis.
While several classes of small molecules have been
developed in this research field, recent attention has
been devoted to d-Phe-Pro-Agmatine-type tripeptides
that lack an electrophilic serine trap.1 It is well under-
stood that the noncovalent inhibitors, owing to their
fast-binding nature, are more efficacious and thus have
a better therapeutic index than covalent inhibitors.1b,2
Since thrombin inhibitors with strongly basic P1 tend to
have poor oral absorption albeit melagatran (1)3 is not
the case, P1 elements with appropriate basicity (pKa
6.5–9) have been of research interest. In this context,
several publications have appeared where amino-
pyridine,4 acylguanidine5 and imidazole6,7 were explored
at the P1 position of the noncovalent tripeptide scaffold.
Our previous research efforts for thrombin inhibitors
had been focused on a class of arylsulfonyl-phenyl-
alanine amides. LB30057 (2) is a prototype compound
of a benzamidrazone class (pKa ꢀ8.9), which demon-
strated good oral availability in animals and humans.8
The benzylamine series derived from this template (e.g.,
4) is also orally bioavailable,9 but the amidine analogues
(e.g., 3) were not orally absorbed despite good anti-
thrombotic activity.10 In our previous study, a rigid and
mildly basic imidazolylethynyl P1 was incorporated into
the tripeptide scaffold in combination with diphenyl-
alanine at P3, leading to potent, selective and orally
available thrombin inhibitors (i.e., 5).7 As an extension
of this work, we describe SAR of the amidrazone, amine
and amidine functions at the P1 position of the d-diphenyl-
Ala-Pro template and identification of a number of
orally available thrombin inhibitors with excellent
potency.
*Corresponding author. Tel.: +82-42-866-2258; fax: +82-42-861-2566; e-mail: kleec@lgci.co.kr
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(02)00093-8

1018
K. Lee et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1017–1022
Scheme 1. (a) CN–Ar–CH2NH2 (12–18), NMM, EDC, HOBT, quant; (b) TFA, CH2Cl2 or AcCl, MeOH; (c) R2-Cl (sulfonylchlorides, sulfa-
moylchlorides, chlorosulfonic acid, chloroformate, chlorophosphate) Et3N, CH2Cl2; for 9g and 11n t-BuO2CCH2Br, DIPEA, CH3CN, rt, 36 h; for
11k Pd(Ph3P)4, CuI, Et3N, K2CO3, PhCH2NEt3Cl, iodobenzene, DMF, 100 ꢁC, 4 h; (d) (i) H2S, py, Et3N; (ii) MeI, CH3CN, reflux; (e)
.
.
NH2NH2 H2O, MeOH, rt; (f) NH4OAc, MeOH, reflux; (g) (i) NH2OH HCl, Na2CO3, EtOH/H2O; (ii) H2, 10% Pd/C, Ac2O, MeOH; (h) H2 (60 psi),
concd HCl, MeOH; (i) CuCN, DMF, reflux; (j) NBS, benzoylperoxide, CCl4, reflux; (k) (Boc)2NH, NaH, THF; (l) BH3/THF, 0 ꢁC; (m) NaBH4,
MeOH, rt; (n) CBr4, PPh3, CH2Cl2.
The target compounds were prepared as outlined in
Scheme 1. The N-protected dipeptides 6 were coupled to
the P1 precursors 12–18 and the N-Boc group was
removed. The free amines 7 were converted to 8 by
reacting with R2-Cl in the presence of base. N-Phenylation
of 7 was accomplished by a Pd-mediated cross coupling
reaction. The intermediates 8 were then subjected to a
Pinner-type reaction which proceeds via a thioimidate
intermediate to produce the amidrazones 9 and the
amidines 11.8,10 The nitrile group could also be con-
verted to the amidine moiety by catalytic hydrogenation
of an amidoxime intermediate under acetylation.11 The
amines 10 were prepared from the same intermediate by
hydrogenation in the presence of HCl.9 The requisite
CN–Ar–CH2NH2, 12–14 and 16, were prepared starting
fromthe corresponding CH 3–Ar–Br in several steps
involving CuCN-mediated cyanation, radical bromination
and a Gabriel-type amination.12 Synthesis of com-
pounds 17 and 18 began with aldehyde reduction of
4-bromothiophene-2-carboxaldehyde and 4-iodothio-
phene-3-carboxaldehyde,13 respectively: following cya-
nation afforded better yields of the hydroxycyano
intermediates than inverse reactions. The hydroxyl
group was subsequently brominated prior to amination.
be related with the vertical N–NH2 of the amidrazono
group which weakly interacts with Asp189 of thrombin
via a lateral interaction (Fig. 1a). It was thus anticipated
that five-membered heterocycles would flip up the NH2
pendant allowing the amidrazone to favorably interact
with Asp189 via a bidentate bridge (Fig. 1b). Indeed,
replacement with thiophene nuclei (9b–d) resulted in a
great enhancement in potency. The expected bidentate
interaction was confirmed by X-ray crystallographic
study on the 9b–thrombin complex where the ami-
drazone group forms H-bonds to the Asp189 carboxylic
oxygens (3.24 and 2.99 A).14 Furthermore, the thio-
phenes 9c and 9d displayed better selectivity for throm-
bin over trypsin. It was, however, disappointing that
none of these amidrazone inhibitors revealed good
absorption behavior when orally administered in rats
(Cmax<0.2 mg/mL, 30 mg/kg). It appears that the ami-
drazone function in this series is not advantageous for
gut membrane permeability. Boc-, acetyl and carboxy-
methyl replacement at the N-terminus (9e–g) led to
improved oral absorption over 9a–d but at the expense
of thrombin potency.
The benzylamine 10a showed potent thrombin inhibi-
tion (Ki=4 nM). This inhibitor was also orally bio-
available to some extent when administered to rats
(Cmax=0.5 mg/mL, 30 mg/kg, n=3). Substitution at the
The benzamidrazone compound 9a exhibited a Ki of 47
nM for thrombin (Table 1). This moderate potency may

K. Lee et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1017–1022
Table 1. Thrombin and trypsin inhibitory activities for compounds 9–11
1019
Compd
R1
R2
Ar
Thrombina Ki
(nM)
Trypsinb Ki
(nM)
Factor Xac Ki
(nM)
9a
Ph(Ph)CH
Ph(Ph)CH
Ph(Ph)CH
Ph(Ph)CH
Ph(Ph)CH
Ph(Ph)CH
Ph(Ph)CH
Ph(Ph)CH
Ph(Ph)CH
Ph(Ph)CH
Ph(Ph)CH
Ph(Ph)CH
Ph(Ph)CH
Ph(Ph)CH
Ph(Ph)CH
MeSO2
MeSO2
MeSO2
MeSO2
Boc
47
0.53
1.8
2.2
84
150
12
Inactive
193
9b
9c
540
250
NT
NT
1,100
9.9
Inactive
59,000
9d
9e
Inactive
Inactive
Inactive
Inactive
Inactive
Inactive
Inactive
Inactive
Inactive
100,000
300
9f
Ac
105
31
9g
HO2CCH2
MeSO2
MeSO2
MeSO2
MeSO2
MeSO2
MeSO2
MeSO2
MeSO2
10a
10b
10c
10d
10e
10f
10g
11a
4
12
11
32
110
220
85
6
5.8
2.4
5.3
0.003
68
606
0.7
(continued on next page)

1020
K. Lee et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1017–1022
Table 1 (continued)
Compd
R1
R2
Ar
Thrombina Ki
(nM)
Trypsinb Ki
(nM)
Factor Xac Ki
(nM)
11b
Ph(Ph)CH
MeSO2
0.005
0.5
0.6
9.9
1.3
0.8
—
360
5000
200
120
—
11c
MeSO2
MeSO2
MeSO2
11d
0.02
0.05
1.2
11e
Melagatran (1)
aHuman thrombin.
bBovine trypsin.
cHuman factor Xa.
ortho position of the phenyl ring (10b,c) was detrimental
to the potency. The pyridine heterocycle of this series
(10d) was beneficial to the aqueous solubility, but did
not improve oral absorption. The thiophene analogues
10e–g, while retaining a comparable level of potency for
thrombin, displayed slightly better trypsin-selectivity
and oral absorption (Cmax=0.7–1.8 mg/mL, same con-
ditions). When tested for their ability to inhibit throm-
bus formation in a rat model of venous thrombosis,
compounds 10a and 10d–g exhibited only moderate
antithrombotic activity (30–67% inhibition, iv bolus, 5
mg/kg, n=3).15 The overall in vivo profile made this series
of compounds unsuitable for further evaluation.
Figure 1.
Remarkable activity for thrombin inhibition was
achieved fromthe series of the amidine analogues. The
benzamidine 11a is extremely potent (Ki=3 pM), which
favorably compares with the most potent thrombin
inhibitors reported to date. The thienyl amidine 11b is
substantially equipotent. Analogues possessing cyclo-
hexylGly, cyclohexylAla, and dichlorophenylAla (11c–e)
were significantly less potent as expected, but still
superior to the amidrazone and the amine analogues.
This indicates high contribution of the amidine moiety
to enzyme affinity via a strong salt–bridge interaction
with Asp189 (Fig. 1c). Most notable is that the di-
phenylAla 11a showed a good absorption profile upon
oral dosing in rats (Cmax=2.8 mg/mL, AUC=263 mg h/
mL, 30 mg/kg) whereas the non-diphenylAla derivatives
were significantly less absorbed (Cmax=0.1–0.9 mg/mL).
In addition, the picomolar inhibitor 11a demonstrated
excellent antithrombotic activity when evaluated in the
rat thrombosis model, reflecting its strong in vitro
activity (70% inhibition at 1 mg/kg, iv bolus, n=3).
substituent as shown in Table 2. Compared to the pri-
mary lead 11a, the sulfamide derivative 11f displayed a
similar level of enzyme inhibition and oral absorption,
but superior antithrombotic activity (100 and 80%
inhibition at 1 and 0.5 mg/kg, respectively, iv bolus, rat,
n=3). However, the more hydrophobic sulfonyl and
sulfamoyl substituents (11g,h) were detrimental to oral
absorption while retaining comparable thrombin inhi-
bitory activity. Compounds possessing sulfamic acid
and aniline moieties (11i,k) were also poorly absorbed,
which is probably due to their poor aqueous solubility.
The phosphonate 11j, the acetamide 11l and the carba-
mate 11m showed slightly improved oral absorption but
at the expense of thrombin potency. Carboxymethyl
substitution (11n) resulted in a decrease in both potency
and oral absorption as compared to those of 11a. Also
notable was the unsubstituted amine compound 11o
that was still absorbed despite its highly basic property.
Overall, the sulfamide 11f showed the best in vivo pro-
file in this series of compounds. This inhibitor was sub-
stantially orally bioavailable when evaluated in dogs
and monkeys (Fig. 2) and metabolically stable in S9
fractions from rat, dog, monkey and human species.
Further SAR studies in the benzamidine series was
focused on the variation of the N-terminal sulfonyl

K. Lee et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1017–1022
1021
Table 2. Enzyme inhibitory activities and pharmacokinetic parameters for compounds 11a–o
Compd
R2
Thrombina Ki
(nM)
Trypsinb Ki
(nM)
Plasma concentrationc
Cmax
(mg/mL)
AUC
(mg h/mL)
11a
11f
11g
11h
11i
11j
11k
11l
MeSO2
NH2S(O) 2
PhCH2S(O) 2
Cyclohexyl-NHS(O) 2
HOS(O) 2
(MeO)2P(O) 2
Ph
0.003
0.003
0.005
0.003
0.10
0.044
0.045
0.16
0.7
0.3
0.58
0.25
4.4
2.9
1.5
NT
0.31
3.0
2.7
2.8
0.9
<0.3
0.55
4.0
0.3
3.0
263
272
68
—
35
271
53
244
594
120
107
Ac
11m
11n
11o
MeOC(O)
HO2CCH2
H
0.036
0.015
0.036
3.9
1.3
1.4
2.5
aHuman thrombin.
bBovine trypsin.
cConcentration after oral dosing at 30 mg/kg as a water solution or as a 20–30% co-solvent (PEG/EtOH/Twin20=85/10/5) solution in water.
References and Notes
1. For recent reviews, see: (a) Hauptmann, L.; Sturzebecher,
J. Thromb. Res. 1999, 93, 203. (b) Sanderson, P. E. J. Med.
Res. Rev. 1999, 19, 179.
2. Stone, S. R.; Tapparelli, C. J. Enzyme Inhib. 1995, 9, 3.
3. (a) Elg, M.; Gustafson, D.; Deinum, J. Thromb. Haemost.
1997, 78, 1286. (b) Eriksson, B. I.; Carlsson, S.; Halvarsson,
M.; Risberg, B.; Mattsson, C. Thromb. Haemost. 1997, 78,
1404.
4. Feng, D.-M.; Gardell, S. J.; Lewis, S. D.; Bock, M. G.;
Chen, Z.; Freidinger, R. M.; Naylor-Olsen, A. M.; Ramjit,
H. G.; Woltmann, R.; Baskin, E. P.; Lynch, J. J.; Lucas, R.;
Shafer, J. A.; Chen, I.-W.; Dancheck, K. B.; Mao, S.-S.;
Krueger, J. A.; Hare, T. R.; Mulichak, A. M.; Vacca, J. P. J.
Med. Chem. 1997, 40, 3726.
5. Adang, A. E. P.; Lucas, H.; de Man, A. P. A.; Engh, R. A.;
Grootenhuis, P. D. J. Bioorg. Med. Chem. Lett. 1999, 9, 3603.
6. Wiley, M. R.; Weir, L. C.; Bliggs, S. L.; Chirgadze, N. Y.;
Clawson, D.; Gifford-Moor, D. S.; Smith, G. F.; Vasudevan,
V.; Zornes, L. L.; Klimkowski, V. J. Bioorg. Med. Chem. Lett.
1999, 9, 2767.
7. Lee, K.; Jung, W.-H.; Kang, M.; Lee, S.-H. Bioorg. Med.
Chem. Lett. 2000, 10, 2775.
8. (a) Oh, Y. S.; Yun, M.; Hwang, S. Y.; Hong, S.; Shin, Y.;
Figure 2. Plasma concentrations of compound 11f after oral adminis-
tration in rats, dogs and monkeys.
In conclusion, we have investigated a series of non-
covalent tripeptidic thrombin inhibitors that incorp-
orate amidrazone, amine and amidine functions at P1.
The amidrazone and the amine series displayed limited
oral absorption despite their mildly basic nature. In
contrast, the amidine series demonstrated generally
good oral absorption and strong antithrombotic activ-
ity; the single-digit picomolar potency achieved from
this series is among the highest yet reported. The pre-
sent work highlights the benzamidine compound 11f
(LB30812) that exhibits excellent overall profiles of
potency, oral absorption and antithrombotic efficacy.
Further SAR studies of this amidine series are in active
progress and the results will be published in due
course.
Lee, K.; Yoon, K. H.; Yoo, Y. J.; Kim, D. S.; Lee, S. H.; Lee,
Y. H.; Park, H. D.; Lee, C. H.; Lee, S. K.; Kim, S. Bioorg.
Med. Chem. Lett. 1998, 8, 631. (b) Lee, K.; Hwang, S. Y.;
Hong, S. W.; Hong, C. Y.; Lee, C.-S.; Shin, Y.; Kim, S.; Yun,
M.; Yoo, Y. J.; Kang, M.; Oh, Y. S. Bioorg. Med. Chem. 1998,
6, 869.
9. Lee, K.; Jung, W.-H.; Park, C. W.; Hong, C. Y.; Kim, I.;
Kim, S.; Oh, Y. S.; Kwon, O. H.; Lee, S.-H.; Park, H. D.;
Kim, S. W.; Lee, Y. H.; Yoo, Y. J. Bioorg. Med. Chem. Lett.
1998, 8, 2563.
10. Lee, K.; Hwang, S. Y.; Yun, M.; Kim, D. S. Bioorg. Med.
Chem. Lett. 1998, 8, 1683.
11. Judkins, B. D.; Allen, D. G.; Cook, T. A.; Evans, B.;
Sardharwala, T. A. Synth. Commun. 1996, 26, 4351.

1022
K. Lee et al. / Bioorg. Med. Chem. Lett. 12 (2002) 1017–1022
12. Schacht, A. L.; Smith, G. F.; Wiley, M. F. US 5710130.
13. Ewing, W. R.; Becker, M. R.; Manetta, V. A.; Davis,
R. S.; Pauls, H. W.; Mason, H.; Choi-Sledeski, Y. M.; Green,
D.; Cha, D.; Spada, A. P.; Cheney, D. L.; Mason, J. S.;
Maignan, S.; Guilloteau, J.-P.; Brown, K.; Colussi, D.; Bent-
ley, R.; Bostwick, J.; Kasiewski, C. J.; Morgan, S. R.; Leadley,
R. J.; Dunwiddie, C. T.; Perrone, M. H.; Chu, V. J. Med.
Chem. 1999, 42, 3557.
14. Details of this data will be reported elsewhere.
15. Millet, J.; Theveniaux, J.; Brown, N. L. Thromb. Haemost.
1992, 76, 176. In this model, compound 2 is 50% effective
upon iv bolus administration at a dose of 1 mg/kg.