Orally active thrombin inhibitors. Part 2: Optimization of the P2-moiety

Article abstract of DOI:10.1016/j.bmcl.2006.01.046

Synthesis and SAR of orally active thrombin inhibitors of the d-Phe-Pro-Arg type with focus on the P2-moiety are described. The unexpected increase in in vitro potency, oral bioavailability, and in vivo activity of inhibitors with dehydroproline as P2-isostere is discussed. Over a period of 24 h the antithrombin activity of the most active inhibitors with IC₅₀s in the nanomolar range was determined in dogs demonstrating high thrombin inhibitory activity in plasma and an appropriate duration of action after oral administration.

Full text of DOI:10.1016/j.bmcl.2006.01.046

Bioorganic & Medicinal Chemistry Letters 16 (2006) 2648–2653
Orally active thrombin inhibitors. Part 2: Optimization
of the P2-moiety
Udo E. W. Lange,a,* Dorit Baucke,b Wilfried Hornberger,a Helmut Mack,a,*
Werner Seitzb and H. Wolfgang Ho¨ffkenb
aAbbott GmbH & Co. KG, D-67061 Ludwigshafen, Germany
bBASF AG, D-67056 Ludwigshafen, Germany
Received 6 December 2005; revised 9 January 2006‘; accepted 9 January 2006
Available online 3 February 2006
Abstract—Synthesis and SAR of orally active thrombin inhibitors of the D-Phe-Pro-Arg type with focus on the P2-moiety are
described. The unexpected increase in in vitro potency, oral bioavailability, and in vivo activity of inhibitors with dehydroproline
as P2-isostere is discussed. Over a period of 24 h the antithrombin activity of the most active inhibitors with IC50s in the nanomolar
range was determined in dogs demonstrating high thrombin inhibitory activity in plasma and an appropriate duration of action after
oral administration.
ꢀ 2006 Elsevier Ltd. All rights reserved.
Thrombin has long been the main focus of thrombosis
research. Its role in blood coagulation and its inhibition
as a therapeutic opportunity for thromboembolic disor-
ders like deep vein thrombosis (DVT), myocardial
infarction, unstable angina, pulmonary embolism, and
ischemic stroke have been extensively reviewed.1
inhibitors of the D-Phe-Pro-Arg type with an N-(carb-
oxymethyl)-D-cyclohexyl alanine unit as P4–P3-moiety
and a heteroarylamidine as P1-moiety.8 A major part
of these studies was concerned with the detailed explora-
tion of the P pocket of thrombin resulting in modifica-
tions of the P2-moiety that not only led to inhibitors
with improved in vitro potency but also with enhanced
oral bioavailability.
Although several highly potent and selective thrombin
inhibitors have been described, most of them lack suitable
oral bioavailability2 and plasma half-life. The main strat-
egies1b pursued to address these problems have been to in-
crease the lipophilicity of the inhibitor and/or to reduce
the basicity of the Arg mimicking moiety. Approaches
drawing on more lipophilic inhibitors are often hampered
by substantial interspecies variations of the oral bioavail-
ability3 and high plasma protein binding.4 Attempts to re-
duce the basicity of the P1-moiety led to prodrug
approaches like ximelagatran the double prodrug of the
thrombin inhibitor melagatran5 which have both recently
been approved in some European countries for DVT6
despite some liver related safety concerns.7
The following discussion will be based on structures
with the above-mentioned P4–P3-building block, 2-am-
idino-5-aminomethylpyridine as P1-moiety and varia-
tions of the P2-moiety. A detailed discussion of the
SAR obtained by the variation of the P4–P3- and P1-
moieties has in part been reported earlier8i and will be
presented separately.8a,b
One of the major structural differences between throm-
bin and the related serine protease trypsin is its Tyr-
Pro-Pro-Trp insertion loop at position 60. When the
natural substrate fibrinogen is bound, the constrained
P pocket is occupied by Val. Bajusz and co-workers9
introduced Pro as a P2-isostere, a building block that
was subsequently used in numerous thrombin inhibitors.
Based on these results and our own findings8 we initiat-
ed the following synthesis program directed toward the
optimization of the P2-moiety.
SAR studies and optimization of our initial leads finally
resulted in the development of orally active thrombin
Keywords: Thrombin inhibitor; Anticoagulant; Factor IIa inhibitor;
Serine protease inhibitor; Amidine.
*
Corresponding authors. Tel.: +49 621 589 3404; fax: +49 621 589
63466 (U.E.W.L.); fax +49 621 589 63419 (H.M.); e-mail addresses:
udo.lange@abbott.com; helmut.mack@abbott.com
The thrombin inhibitors were prepared by a convergent
synthesis. A representative example is depicted in
0960-894X/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.01.046

U. E. W. Lange et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2648–2653
2649
Scheme 1.10 The dehydroproline synthesis is suitable for
scale-up.8h,11 The P3–P4-building block 4 was easily ob-
tained from Cha-OBn.8b,h The two amide bonds of the
inhibitor were established with 1-propanephosphonic
acid cyclic anhydride (PPA)12 as coupling reagent and
the extremely mild conversion of nitrile 5 to amidine 6
did neither affect the stereocenters nor the double bond
of the pyrrolidine ring.13
Table 1. Effect of flexible P2 moieties on the in vitro potency and
in vivo activity of the thrombin inhibitors
NH
NH2
H
N
N
P2
HO
N
H
O
O
Compound P2
Chromogenic
ECT (s), ECT(s),
thrombin assay po (rat)b iv (rat)c
a
IC50 (nM)
Thrombin inhibitors of the type described above with
flexible amino acid building blocks in the P2 position
exhibit good to moderate in vitro potency and weak to
moderate in vivo activity in rat (Table 1). The alanine
derivative 7 had about the same in vitro potency and
in vivo activity as melagatran14 (cf. Table 2).15,16 The
in vitro potency of the epimer 8 was substantially lower.
The N-methyl alanine derivative 9 was slightly more po-
tent in vitro than 7. The X-ray structure of 9 in throm-
bin revealed that the two methyl groups of the P2 unit
are engaged in favorable hydrophobic interactions in
the P pocket (cf. Fig. 1).17
CH3
7
76.8
261
n.d.
153
83
N
H
O
CH3
8
9
53600
n.d.
75
N
H
O
CH3
20.2
N
CH3
O
H3C CH3
10
11
12
581
92
n.d.
80
59
N
H
O
The a-methyl alanine 10 and the inhibitors 11 and 13
containing
a 1-amino-1-carboxy-cycloalkyl building
4090
n.d.
46
N
H
O
block had only low to moderate activities. The disubsti-
tuted a-position either leads to an inhibitor conforma-
tion unfavorable for binding to the enzyme or the
additional a-substituent sterically interferes with the
van der Waals surface of the P pocket. The latter is con-
sistent with the low activity observed for D-alanine 8. An
exception is the 1-amino-1-carboxy-cyclopentyl deriva-
tive 12 that showed good potency in the chromogenic
thrombin assay, although the cyclopentyl ring came in
close contact with Cd-2 of Leu99. Trp60D had to be
47.3
N
H
O
13
103
43
35
N
H
O
n.d., not determined.
a See Ref. 15.
˚
pushed away by 0.86 A to accommodate this P2-moie-
b Measured 60 min after 21.5 mg · kgÀ1 po, see Ref. 16.
ty.18 The energy required for this rearrangement is
somewhat compensated by the large additional hydro-
phobic contact area between the cyclopentyl ring and
the 60 loop of thrombin (cf. Fig. 1).
c Measured 60 min after 1.0 mg · kgÀ1 iv, see Ref. 16.
bin inhibitor 14,8d which also had good oral activity
in rat (Table 2). Introduction of dehydroproline (Dhp,
compound 6) in place of Pro slightly increased the in vitro
potency and the high in vivo activity was maintained.
Although the increase in in vitro potency of the dehy-
droproline derivative compared to its proline analog is
Within this set of inhibitors (Table 1) N-methyl alanine 9
displayed the best in vitro potency, but did not exhibit
sufficient oral activity. Replacement of this flexible
moiety by the more rigid Pro gave the nanomolar throm-
HO
CN
a) - b)
(1)
c)
H
O
OH
N
N
N
N
N
H
O
O
O
O
O
O
N
O
1
2
3
NH
NH2
(2)
CN
d)
e) - f)
H
H
N
N
N
N
O
OH
O
N
HO
N
N
N
H
O
O
O
O
O
O
O
O
5
6
O
O
O
O
4
P4
P3
P2
P1
Scheme 1. Convergent synthesis of a D-Phe-Pro-Arg type thrombin inhibitor. Reagents and conditions: (a) 1. 1.1 equiv NEt3, 1.05 equiv MsCl,
1 mol% DMAP, DCM, À10 to 0 ꢁC, 2 h, 98%, 2. 2 equiv aqueous 1 N NaOH, dioxane, 0–5 ꢁC, 2.5 h, 87%; (b) 2.7 equiv MeOCH2CH2OH, 2.5 equiv
NaH, DME, 0 ꢁC to rt, 12 h, 52%; (c) 1. 5-aminomethyl-2-cyanopyridine, 1.3 equiv 50% PPA in EtOAc, 4 equiv DIEA, DCM, 2. HCl, i-PrOH, 80%
(2 steps); (d) 3, 1 equiv 50% PPA in EtOAc, 4 equiv DIEA, DCM, 0 ꢁC to rt, 95%; (e) NH3(g), 1.05 equiv N-acetylcysteine, MeOH, D; (f) HCl, H2O,
60 ꢁC, 3 h, 45% (2 steps).

2650
U. E. W. Lange et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2648–2653
Table 2. Effect of cyclic P2 moieties on the in vitro potency and in vivo
activity of the thrombin inhibitors
Compound P2
Chromogenic
ECT (s), ECT(s),
thrombin assay po (rat)b iv (rat)c
a
IC50 (nM)
14
6
3.2
1.8
800
792
100
100
104
138
54
N
N
O
O
15
16d
14.8
71.2
N
N
O
O
59
F
F
17
290
96
38
N
O
CH3
H3C
18
44.2
31.9
56.3
55
229
74
37
59
47
Figure 1. X-ray structure derived structures showing the orientation of
the thrombin inhibitors 9 (purple) and 12 (green) in the active site of
thrombin. The amino acids Leu99, Tyr60A, Trp60D, and His57
forming the P pocket are highlighted (yellow: complex with 9; red:
complex with 12).
N
O
19
N
O
20d
the coordinates of the backbone Cb and Cd unchanged
(cf. Fig. 2). The resulting conformational changes in
the protein minimize these interactions, but reduce the
binding energy of 14 since the hydrophobic contact area
of Pro is similar to that of Dhp. This effect is even more
pronounced for derivatives 158d and 16 with the larger
pipecolinic and dehydropipecolinic acid moieties.22 The
more flexible piperidine derivative 15 had a 4 times high-
er in vitro potency than the more rigid dehydropiperidine
16, which obviously required substantial conformational
N
N
N
N
O
O
O
O
S
21
58.4
16.0
516
196
141
726
49
77
S
22d
23
O
N
288
64
24d
25d
1.7
201
N
O
HN
3.4
640
247
114
98
N
O
Melagatran
69.2
a See Ref. 15.
b Measured 60 min after 21.5 mg · kgÀ1 po, see Ref. 16.
c Measured 60 min after 1.0 mg · kgÀ1 iv, see Ref. 16.
d See Ref. 21.
not significant, it was observed as a general trend in most
of the D-Phe-Pro-Arg type thrombin inhibitors studied.19
The X-ray structure of 6 in complex with human throm-
bin20 revealed an optimal fit of the Dhp moiety in the
˚
˚
P pocket with distances of 3.92 A and 3.91 A from Cc
of Dhp to Cd-2 of Leu99, and from Cc of Dhp to Cg-
2 of Trp60D, respectively. Ringflip of the pyrrolidine-
Cc in 14 between the two possible ring conformations
representing the two local minima of the Pro showed in
both cases unfavorable close contacts with either Leu99
or Trp60D when modeled onto the Dhp of 6 leaving
Figure 2. Orientation of the Dhp (Cc and Cb in green) and Pro (Cc
and Cb in blue) moiety of thrombin inhibitors 6 and 14, respectively, in
the P pocket of thrombin based on data from the X-ray structure of the
thrombin–inhibitor complex from Dhp derivative 6 and modeling
results.

U. E. W. Lange et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2648–2653
2651
changes in the P pocket to be accommodated. Their
in vivo activities, however, were comparable.
or oxygen (cf. 17, 21, and 23) increases the electron den-
sity in this position introducing an unfavorable interac-
tion with the p-electrons of Tyr60A leading to a loss in
potency. In the case of the 4-oxaproline 23, this effect
was especially emphasized since in solution the hydro-
philic oxygen can assumed to be hydrated. The bound
water molecule has to be removed before the oxazolidine
ring can be accommodated in the P pocket. The 3-thia-
proline 22 is not involved in such negative interactions.
Due to the larger ring size caused by the sulfur atom its
in vitro potency, however, is only comparable to that of
the slightly less potent pipecolinic acid derivative 15.
Inhibitors with substituents on the proline ring like fluo-
rine (17),23 methyl (18)24 or methylene (19 and 20)25 were
less potent than the parent compounds 14 and 6. The 4-
thiaproline 21 exhibited surprisingly high oral activity
compared to its moderate potency in the chromogenic
thrombin assay. The corresponding 3-thiaproline 2226
did not prolong the ecarin clotting time (ECT) to the ex-
tent as the better in vitro data would suggest. This could
be attributed to its lower metabolic stability. The 4-oxa-
proline 2327 was less potent than its Pro analog 14. 5-aza-
dehydroproline 2428 and 5-azaproline 2529 were more
potent in vitro. To our surprise they also had a compara-
tively high oral activity indicating their good metabolic
stability. The in vivo activity after oral administration of
24did, however, notquitematchthatofdehydroproline 6.
The 5-nitrogen of 5-azadehydroproline 24 and 5-azapr-
oline 25 does neither improve nor hinder the binding
to the enzyme30 because it is exposed to the solvent
and not interacting with the protein. The observed in vi-
tro potencies therefore closely resemble those of Dhp
derivative 6 and Pro derivative 14. It is worth noting
that the 5-azadehydroproline 24 was slightly more po-
tent in vitro and more active in vivo than the corre-
sponding 5-azaproline 25.
Molecular modeling studies helped to explain the ob-
served in vitro potencies. Cb of the Pro moiety in 14 forms
a van der Waals contact with Cd-2 of His57 and Cc of the
Pro is located close to the benzene ring of Tyr60A.
Pro 14 and Dhp 6, two of the most potent inhibitors in
this series, were chosen for preliminary pharmacody-
namic and pharmacokinetic profiling in dog. They dis-
played an oral bioavailability of 25% and 36%,
respectively (Table 3).
Any substituent larger than hydrogen in these positions
therefore has a negative effect on the binding energy (cf.
18–20, respectively). Exchange of the c-methylene group
in the Pro moiety of 14 against difluoromethylene, sulfur
Table 3. Effect of Dhp versus Pro in P2 on the in vitro potency, oral bioavailability and in vivo activity of the thrombin inhibitors
NH
NH2
H
A
N
HO
N
N
H
O
O
O
Compound
A
P2
Chromogenic
thrombin assay
a
IC50 (nM)
AUC0–24 h by
ECTmean SD · 103 (s)b
(n = 2–8) (dog)
Foral (%),c (dog)
14
6
Pro
3.2
1.8
74.3 3.5
25
36
34
35
48
58
24
N
N
Dhp
Pro
185.0 68.6*
100.1 26.7
168.0 38.5*
158.9 70.2
278.9 47.4*
74.6 27.7
26
27
28
29
2.7
Dhp
Pro
1.7
S
S
1.0
Dhp
1.0
Melagatran
69.2
a See Ref. 15.
b After administration of 4.64 mg · kgÀ1, po. Calculated from data depicted in Figure 3. Asterisk refers to significance (P < 0.05, calculated by
unpaired t test with Welch’s correction) in comparison to the corresponding proline analog.
c Calculated from the AUCs0–24 h after iv and po administration, see Ref. 33.

2652
U. E. W. Lange et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2648–2653
To study the effect of the Dhp P2-moiety on the in vitro
potency, oral bioavailability, and in vivo activity in
more detail, a series of thrombin inhibitors as sets of
Pro and Dhp pairs with different P1 building blocks
was prepared. The benzamidine 2631 and thienylamidine
2832 have also been described by others. For comparison
melagatran was also included in this study. The results
of the pharmacodynamic and pharmacokinetic experi-
ments are summarized in Table 3 and Figure 3.
Acknowledgments
We thank M. Becker, G. Flo¨rchinger, R. Ha¨uselmann,
H.-J. Helfrich, M. Klein, H.-J. Krafczyk, C. Lo¨tterle,
A. Michel, S. Pister, P. Scarano, and K. Studenroth
for supporting chemical synthesis, S. Brodt, A. Freytag,
and K. Rabach for performing the in vitro assays and
the pharmacokinetic experiments, and A. Riebel, P.
Reis, and W. Houy for carrying out the soaking exper-
iments and recording the X-ray structures.
All compounds (6, 14, and 26–29) exhibited higher
in vitro potency and in vivo activity higher than those
of melagatran. The observed ECTs in particular illus-
trate the improved oral bioavailability and antithrombin
properties of these inhibitors. In all cases, the Dhp deriv-
ative matched or surpassed the in vitro potency of the
corresponding Pro compound. For the thrombin inhib-
itors bearing a heterocyclic P1-moiety the Dhp deriva-
tives (6 and 29) showed a substantially increased oral
bioavailability in dogs compared to the Pro containing
compounds (14 and 28). In the case of the benzamidine,
the measured oral bioavailability was about the same for
both derivatives (26 and 27).
References and notes
1. (a) Steinmetzer, T.; Sturzebecher, J. Curr. Med. Chem.
¨
2004, 11, 2297; (b) Steinmetzer, T.; Hauptmann, J.;
Sturzebecher, J. Exp. Opin. Invest. Drugs 2001, 10, 845;
¨
(c) Rewinkel, J. B. M.; Adang, A. E. P. Curr. Pharm. Des.
1999, 5, 1043; (d) Wiley, M. R.; Fisher, M. J. Exp. Opin.
Ther. Patents 1997, 7, 1265.
2. Menear, K. Current Med. Chem. 1998, 5, 457.
3. (a) Brady, S. F.; Stauffer, K. J.; Lumma, W. C.; Smith, G.
M.; Ramjit, H. G.; Lewis, S. D.; Lucas, B. J.; Gardell, S.
J.; Lyle, E. A.; Appleby, S. D.; Cook, J. J.; Holahan, M.
A.; Stranieri, M. T.; Lynch, J. J.; Lin, J. H.; Chen, I.-W.;
Vastag, K.; Naylor-Olsen, A. M.; Vacca, J. P. J. Med.
Chem. 1998, 41, 401; (b) Tucker, T. J.; Lumma, W. C.;
Lewis, S. D.; Gardell, S. J.; Lucas, B. J.; Sisko, J. T.;
Lynch, J. J.; Lyle, E. A.; Baskin, E. P.; Woltmann, R. F.;
Appleby, S. D.; Chen, I.-W.; Dancheck, K. B.; Naylor-
Olsen, A. M.; Krueger, J. A.; Cooper, C. M.; Vacca, J. P.
J. Med. Chem. 1997, 40, 3687.
The activity after oral administration based on the
ecarin clotting time of the Dhp derivatives (6, 27, and
29) was significantly higher than that of the correspond-
ing Pro analogs (14, 26, and 28; cf. Table 3, Fig. 3). The
improvements in ECTs were much higher than the
observed differences in in vitro potency would suggest.
The cause of this remarkable effect is unknown, but
could, e.g., either be due to a higher metabolic stability
of the Dhp containing inhibitors and/or could be caused
by better active resorption of these analogs possibly
favored by the more rigid P2 isostere.
4. Tucker, T. J.; Lumma, W. C.; Lewis, S. D.; Gardell, S. J.;
Lucas, B. J.; Baskin, E. P.; Woltmann, R.; Lynch, J. J.;
Lyle, E. A.; Appleby, S. D.; Chen, I.-W.; Dancheck, K. B.;
Vacca, J. P. J. Med. Chem. 1997, 40, 1565.
5. Gustafsson, D.; Nystro¨m, J. E.; Carlsson, S.; Bredberg,
U.; Eriksson, U.; Gyzander, E.; Elg, M.; Antonsson, T.;
Hoffmann, K.; Ungell, A.; Sorensen, H.; Nagard, S.;
Abrahamsson, A.; Bylund, R. Thromb. Res. 2001, 101,
171.
6. Fiessinger, J.-N.; Huisman, M. V.; Davidson, B. L.;
Bonameaux, H.; Francis, C. W.; Eriksson, H.; Lundstro¨m,
´
T.; Berkowitz, S. D.; Nystro¨m, P.; Thorsen, M.; Ginsberg,
In summary, the synthesis of orally active thrombin
inhibitors of the D-Phe-Pro-Arg type with high
antithrombin potency has been described. The SAR
with respect to the interactions observed for different
P2-moieties in the P pocket of thrombin has been
discussed. The noteworthy positive effect of the
P2-isostere dehydroproline on the in vitro potency and
especially the in vivo activity after oral administration
has been highlighted.
J. S. JAMA 2005, 293, 681.
7. Gurewich, V. JAMA 2005, 293, 736.
8. (a) Modification of the P1-moiety: Baucke, D.; Hornber-
ger, W.; Lange, U. E. W.; Mack, H.; Seitz W. Bioorg.
Med. Chem. Lett. preceding manuscript; (b) Modification
of the P3-moiety and P4-moiety: Baucke, D.; Hornberger,
W.; Lange, U. E. W.; Mack, H.; Seitz W. manuscript in
preparation; (c) Bo¨hm, H.-J.; Koser, S.; Mack, H.;
Pfeiffer, T.; Seitz, W.; Ho¨ffken, H. W.; Hornberger, W.
WO9535309, 1995; Chem. Abstr. 1995, 124, 233168; (d)
Bo¨hm, H.-J.; Ho¨ffken, H. W.; Hornberger, W.; Koser, S.;
Mack, H.; Pfeiffer, T.; Seitz, W.; Zierke, T. WO9625426,
1996; Chem. Abstr. 1996, 125, 301605; (e) Baucke, D.;
Lange, U.; Mack, H.; Seitz, W.; Zierke, T.; Ho¨ffken, H.-
W.; Hornberger, W. WO9806741, 1998; Chem. Abstr.
1998, 128, 192940; (f) Baucke, D.; Lange, U.; Mack, H.;
Seitz, W.; Ho¨ffken, H.-W.; Hornberger, W. WO9937668,
1999; Chem. Abstr. 1999, 131, 116527; (g) Solid-phase
synthesis of thrombin inhibitors: Lange, U. E. W.; Zechel,
C. Bioorg. Med. Chem. Lett. 2002, 12, 1571; (h) Technical
scale synthesis of a new and highly potent thrombin
inhibitor: Bernard, H.; Bulow, G.; Lange, U. E. W.;
¨
Mack, H.; Pfeiffer, T.; Scha¨fer, B.; Seitz, W.; Zierke, T.
Figure 3. Effect of Dhp versus Pro in P2 on the ecarin clotting time in
dogs at a dose of 4.64 mg · kgÀ1, po.

U. E. W. Lange et al. / Bioorg. Med. Chem. Lett. 16 (2006) 2648–2653
2653
Synthesis 2004, 2367; (i) Modification of the P1-moiety
giving unexpected selectivity: Lange, U. E. W.; Baucke,
D.; Hornberger, W.; Mack, H.; Seitz, W.; Ho¨ffken, H. W.
Bioorg. Med. Chem. Lett. 2003, 13, 2029.
23. Compound 17 was prepared by linear synthesis, cf. Ref. 8d
examples 32 and 93. For the synthesis of (S)-4-difluoropr-
oline building block, cf.: (a) Burger, K.; Rudolph, M.;
Fehn, S.; Sewald, N. J. Fluorine Chem. 1994, 66, 87; (b)
O’Neil, I. A.; Thompson, S.; Kalindjian, S. B.; Jenkins, T.
C. Tetrahedron Lett. 2003, 44, 7809.
9. (a) Pozsgay, M.; Szabo, G. C. S.; Bajusz, S.; Simonsson,
´
R. Eur. J. Biochem. 1981, 115, 491; (b) Bajusz, S.; Barabas,
E.; Bragdy, D. In Proc. 17th Eur. Pept. Symp. 1982, Blaja,
K., Malon, P., Eds.; Walter de Gruyter & Co.: Berlin—
New York, 1983, p 643.;
24. Compound 18 was prepared by linear synthesis, cf. Ref. 8d
examples 48 and 93. For the synthesis of (S)-4,4-dim-
ethylproline building block, cf.: (a) Hippeli, C.; Reissig, H.
U. Liebigs Ann. Chem. 1990, 475; (b) Ezquerra, J.;
Pedregal, C.; Rubio, A. J. Org. Chem. 1994, 59, 4327.
25. For the synthesis of (S)-4-methylene- and (S)-3-methy-
lene-pyrrolidine-2-carbonic acid building blocks, respec-
10. All thrombin inhibitors were prepared by this route unless
indicated otherwise. Alternatively the linear route
described in Ref. 8d,i or the solid-phase synthesis
described in Ref. 8g could be used.
´
11. Mack, H.; Pfeiffer, T.; Seitz, W.; Zierke, T.; Balkenhohl,
F.; Lange, U. WO9804523, 1998; Chem. Abstr. 1998, 128,
154380.;
12. Wissmann, H.; Kleiner, H.-J. Angew. Chem., Int. Ed. Engl.
1980, 19, 133.
13. (a) Lange, U. E. W.; Scha¨fer, B.; Baucke, D.; Buschmann,
E.; Mack, H. Tetrahedron Lett. 1999, 40, 7067(b) Less
reactive thienyl- and benzonitrile derivatives were con-
verted to the corresponding amidines by the Thio-Pinner
reaction described in Ref. 8e.;
14. Gustafsson, D.; Antonsson, T.; Bylund, R. Thromb.
Haemost. 1998, 79, 110.
15. (a) For in vitro assay protocols, see Ref. 8f.; (b) The
standard error of the IC50 determinations was less than 1%
with r2 > 0.99 for the curve fit.
16. (a) Nowak, G. Pathophysiol. Haemost. Thromb. 2003, 33,
173(b) Measurement of ecarin clotting time (ECT):
Venous blood is collected from rats or dogs, respectively,
at the desired time after administration of the thrombin
inhibitor. One hundred microliters of citrated blood is
pipetted into a coagulation cuvette and incubated for
2 min at 37 ꢁC in a coagulometer. After addition of 100 lL
of prewarmed (37 ꢁC) ecarin reagent (Pentapharm), the
time (ECT) to formation of a fibrin clot is determined.;
17. X-ray structure analysis of thrombin–inhibitor complex of
9; X-ray crystallographic data have been deposited with the
Brookhaven Protein Data Bank. Deposition code 2A2X.
18. X-ray structure analysis of thrombin–inhibitor complex of
12; X-ray crystallographic data have been deposited with
the Brookhaven Protein Data Bank. Deposition code
2ANK.
tively, cf.: (a) Manfre, F.; Kern, J.-M.; Biellmann, J.-F.
J. Org. Chem. 1992, 57, 2060; (b) Tandon, M.; Wu, M.;
Begley, T. P.; Myllyharju, J.; Pirskanen, A.; Kivirikko, K.
Bioorg. Med. Chem. Lett. 1998, 8, 1139. Compound 19
was prepared by linear synthesis, cf. Ref. 8d examples 32
and 93.
26. For the synthesis of the rac-thiazolidine-2-carbonic acid
building block, cf.: Johnson, R. L.; Smissman, E. E.
J. Med. Chem. 1978, 21, 165.
27. Compound 23 was prepared by linear synthesis, cf. Ref. 8d
examples 32 and 93. Synthesis of the 4-oxaproline building
block: Conti, S.; Cossu, S.; Giacomelli, G.; Falorni, M.
Tetrahedron 1994, 50, 13493.
28. Compound 24 was prepared by linear synthesis. Com-
pound 4 was coupled successively with the P2-building
block and with 5-aminomethyl-2-cyanopyridine. The
standard amidine synthesis and deprotection gave 24.
Synthesis of the 5-azadehydroproline building block: (a)
Mish, M. R.; Guerra, F. M.; Carreira, E. M. J. Am. Chem.
Soc. 1997, 119, 8379; (b) Guerra, F. M.; Mish, M. R.;
Carreira, E. M. Org. Lett. 2000, 2, 4265.
29. Compound 25 was prepared by linear synthesis. Com-
pound 4 was coupled successively with the P2-building
block and with 5-aminomethyl-2-cyanopyridine. The
standard amidine synthesis and deprotection gave 25.
Synthesis of the 5-azaproline building block: Liu, B.;
Brandt, J. D.; Moeller, K. D. Tetrahedron 2003, 59, 8515.
30. The conclusions were drawn by analogy visualizing
molecule 6 in the binding pocket of thrombin with
QUANTA (Accelry, San Diego, USA 2002).;
31. (a) Antonsson, K. T.; Bylund, R.; Gustafsson, N. D.;
Nilsson, N. O. WO9429336, 1994; Chem. Abstr. 1995, 122,
285553.; (b) Onshima, M.; Iwase, N. Sugiyama, S.
EP669317, 1995; Chem. Abstr. 1996, 124, 56705.;
19. Unpublished results.
20. X-ray structure analysis of thrombin–inhibitor complex of
6; X-ray crystallographic data have been deposited with
the Brookhaven Protein Data Bank. Deposition code
2ANM.
32. Smith, G. F.; Wiley, M. R.; Schacht, A. L.; Shuman, R. T.
WO9523609, 1995; Chem. Abstr. 1996, 124, 87791.;
33. (a) Plasma levels of thrombin inhibitors were determined
in dogs by means of ex vivo measurements of ECT in
plasma (cf. Ref. 16b). The ECT increases linearly with
plasma concentration and therefore provides a suitable
method for preliminary pharmacodynamic/pharmacoki-
netic assessment (cf. Ref. 16a). For the calculation of oral
bioavaolability (F) the AUC0–24 h of the plasma levels were
determined after oral administration of 4.64 mg · kgÀ1
21. The inhibitor was obtained as a mixture of diastereomers
due to a racemic P2-building block, the diastereomers
were separated by HPLC and tested separately. In Table 2,
the biological data of the more active analog are shown.
By comparison with very related inhibitors of known
absolute configuration we are confident to assign the
stereochemistry of the P2-moiety depicted in Table 2.
22. For the synthesis of (S)-1,2,3,6-tetrahydropyridine-2-car-
bonic acid, cf.: Guillena, G.; Najero, C. J. Org. Chem.
2000, 65, 7310.
and intravenous administration of 1.0 mg · kgÀ1
,
respectively.